Developing countries' business participation in the AI economy

Lippold , Douglas
Wo king Pape
De eloping coun ies' business pa icipa ion in he AI
economy
CIGI Pape s, No. 323
P o ided in Coope a ion wi h:
Cen e o In e na ional Go e nance Inno a ion (CIGI), Wa e loo, On a io
Sugges ed Ci a ion: Lippold , Douglas (2025) : De eloping coun ies' business pa icipa ion in he AI
economy, CIGI Pape s, No. 323, Cen e o In e na ional Go e nance Inno a ion (CIGI), Wa e loo
(On a io)
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/322457
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
CIGI Pape No. 323 — June 2025
De eloping Coun ies’ Business
Pa icipa ion in he AI Economy
Douglas Lippold
CIGI Pape No. 323 — June 2025
De eloping Coun ies’ Business
Pa icipa ion in he AI Economy
Douglas Lippold
Abou CIGI
The Cen e o In e na ional Go e nance Inno a ion (CIGI) is an independen ,
non-pa isan hink ank whose pee - e iewed esea ch and us ed analysis
in luence policy make s o inno a e. Ou global ne wo k o mul idisciplina y
esea che s and s a egic pa ne ships p o ide policy solu ions o he digi al
e a wi h one goal: o imp o e people’s li es e e ywhe e. Headqua e ed
in Wa e loo, Canada, CIGI has ecei ed suppo om he Go e nmen o
Canada, he Go e nmen o On a io and ounde Jim Balsillie.
À p opos du CIGI
Le Cen e pou l’inno a ion dans la gou e nance in e na ionale (CIGI) es un
g oupe de é lexion indépendan e non pa isan don les eche ches é aluées
pa des pai s e les analyses iables inci en les décideu s à inno e . G âce
à son éseau mondial de che cheu s plu idisciplinai es e de pa ena ia s
s a égiques, le CIGI o e des solu ions poli iques adap ées à l’è e numé ique
dans le seul bu d’amélio e la ie des gens du monde en ie . Le CIGI, don le
siège se ou e à Wa e loo, au Canada, béné icie du sou ien du gou e nemen
du Canada, du gou e nemen de l’On a io e de son onda eu , Jim Balsillie.
C edi s
Resea ch Di ec o , Digi al Economy S. Yash Kalash
Di ec o , P og am Managemen Dianna English
P og am Manage G ace W igh
Manage , Publica ions Jenni e Goyde
Publica ions Edi o Ch is ine Robe son
Publica ions Edi o Susan Bubak
G aphic Designe Sami Chouhda y
Copy igh © 2025 by he Cen e o In e na ional Go e nance Inno a ion
The opinions exp essed in his publica ion a e hose o he au ho and do no
necessa ily e lec he iews o he Cen e o In e na ional Go e nance Inno a ion
o i s Boa d o Di ec o s.
Fo publica ions enqui ies, please con ac [email p o ec ed].
The ex o his wo k is licensed unde CC BY 4.0. To iew a copy o his licence,
isi h p://c ea i ecommons.o g/licenses/by/4.0/.
Fo euse o dis ibu ion, please include his copy igh no ice. This wo k may
con ain con en (including bu no limi ed o g aphics, cha s and pho og aphs)
used o ep oduced unde licence o wi h pe mission om hi d pa ies.
Pe mission o ep oduce his con en mus be ob ained om hi d pa ies di ec ly.
Cen e o In e na ional Go e nance Inno a ion and CIGI a e egis e ed
adema ks.
67 E b S ee Wes
Wa e loo, ON, Canada N2L 6C2
www.cigionline.o g

Table o Con en s
i Abou he Au ho
i Ac onyms and Abb e ia ions
1 Execu i e Summa y
1 In oduc ion
5 Li e a u e Re iew
10 AI S a -Ups and he Case S udy Coun ies
32 Policy Assessmen
34 Conclusions
38 Appendix 1: Key Sou ces
40 Appendix 2: Da a
46 Appendix 3: OECD Recommenda ion o he
Council on A i icial In elligence
47 Wo ks Ci ed
i CIGI Pape No. 323 — June 2025 • Douglas Lippold
Abou he Au ho
Douglas (Doug) Lippold is a CIGI senio ellow
and an in e na ional ade economis based in
Cla emon , Cali o nia. He se ed as chie ade
economis a HSBC Global Resea ch om 2014
o 2020. P e iously, he se ed in a ious oles
as a senio economis a he O ganisa ion o
Economic Co-ope a ion and De elopmen in Pa is,
F ance, du ing a enu e o 22 yea s. Doug’s ea ly
ca ee included se en yea s as an in e na ional
economis wi h he US Depa men o Labo .
He has published ex ensi ely on ade opics
as well as on ela ed aspec s o economic
de elopmen , labou ma ke adjus men ,
inno a ion and in ellec ual p ope y. Doug holds
a Ph.D. in economics om he Ins i u d’é udes
poli iques de Pa is (Sciences Po), an M.A. in
in e na ional s udies om he Uni e si y o Den e
and a B.A. in in e na ional s udies om Washing on
College in Ma yland. He was a Fulb igh Schola
a he Uni e si y o Cologne, Ge many, and a
Peace Co ps olun ee in Bu kina Faso. Du ing
2015–2019, he ep esen ed HSBC as he depu y
delega e on he Business 20 T ade and In es men
Task Fo ce. He is cu en ly a con ibu ing
au ho on a Think20 ade ask o ce eam.
Ac onyms and
Abb e ia ions
AI a i icial in elligence
ASEAN Associa ion o Sou heas Asian Na ions
BRI Bel and Road Ini ia i e
DEFA Digi al Economy F amewo k Ag eemen
DEPA Digi al Economy Pa ne ship Ag eemen
DPA Digi al Policy Ale
ECLAC Economic Commission o La in
Ame ica and he Ca ibbean
G7 G oup o Se en
GenAI gene a i e AI
IoT In e ne o Things
IP in ellec ual p ope y
IPO ini ial public o e ing
LLM la ge language model
MENA Middle Eas and No h A ica
ML machine lea ning
NCAIR Na ional Cen e o AI and Robo ics
NLP na u al language p ocessing
OECD O ganisa ion o Economic
Co-ope a ion and De elopmen
PPP pu chasing powe pa i y
R&D esea ch and de elopmen
SMEs small and medium-sized en e p ises
STEM science, echnology, enginee ing
and ma hema ics
UAE Uni ed A ab Emi a es
VC en u e capi al
WEF Wo ld Economic Fo um
1De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Execu i e Summa y
How a e s a -up businesses in middle-income
de eloping coun ies pa icipa ing in he apidly
expanding a i icial in elligence (AI) economy?
A e some de eloping coun y AI s a -ups
able o ap in o he AI ecosys ems o leading
economies in G oup o Se en (G7) coun ies?1
Wha is he domes ic policy con ex , and how
can condi ions be imp o ed? A e he e signs
o Sou h-Sou h AI ecosys em de elopmen ?
This case s udy akes s ock o he de elopmen
o AI s a -ups engaged wi h G7 in es o s and
in es o s om ac oss 10 middle-income de eloping
coun ies. I conside s he eme ging in es men AI
ecosys ems in hese coun ies, which a e loca ed
ac oss A ica, Sou heas Asia and La in Ame ica.
The signi ican concen a ion o ad anced AI
inno a ion in de eloped economies heigh ens
he impo ance o engagemen wi h coun e pa s
in G7 coun ies who can complemen local
esou ces wi h access o addi ional inancial,
echnological and manage ial suppo .
The esea ch employs a compa a i e case s udy
app oach analyzing i m-le el da a om 2,537 AI
s a -ups ac oss B azil, Colombia, Egyp , Indonesia,
Kenya, Nige ia, Sou h A ica, Thailand, Tunisia
and Vie nam. Pa icula a en ion is gi en o he
ole o G7 in es o s in acili a ing echnology
ans e , wi h an examina ion o 239 AI s a -ups
ha ha e a ac ed such in es men . The analysis
is complemen ed by an assessmen o AI policy
and egula o y amewo ks in each s udy coun y.
F om na ional s a egies o implemen a ion and
en o cemen , a ious dimensions a e e iewed.
Key indings indica e ha :
→ Despi e modes scale, iable AI ecosys ems
a e de eloping in all he case s udy coun ies;
Sou h-Sou h in es men ne wo ks a e eme ging
alongside dominan No h-Sou h ela ionships;
and policy and egula o y de elopmen is
b oadly suppo i e.
→ Among he 10 coun ies, B azil has achie ed
some scale; domes ic in es o s he e supply a
1 The G7 coun ies a e Canada, F ance, Ge many, I aly, Japan, he Uni ed
Kingdom and he Uni ed S a es.
majo i y o in es men engagemen s2 wi h local
AI s a -ups, which is unique o hese ma ke s.
B azil a ac s by a he la ges in low o US
in es o engagemen s, while in mos o he o he
ma ke s, US in es o s play a lead ole.
→ The e is some geog aphic clus e ing o hese AI
s a -ups, poin ing o he possible eme gence
o inno a ion hubs in and a ound ci ies such as
Jaka a, Lagos and São Paulo.
→ All o he hos coun ies a e de eloping na ional
s a egies. Implemen a ion p og ess a ies
widely. In all o he case s udy coun ies,
challenges emain in ela ion o issues such as
in as uc u e cons ain s, skills de elopmen ,
he u ban- u al digi al di ide, egula ing AI
sa e y while acili a ing inno a ion, and capaci y
limi a ions o he public adminis a ion, among
o he issues.
The analysis unde sco es ha while hese
de eloping coun ies a e pu suing di e se
s a egies o AI de elopmen , hey a e also acing
challenges in s i ing o egula o y con e gence
wi h majo AI powe s, which a e hemsel es
di e ging on some issues such as isk managemen .
To succeed, he case s udy coun ies mus
balance in e na ional alignmen wi h domes ic
inno a ion suppo , add ess in as uc u e
and alen sho ages, and de elop go e nance
amewo ks ha p o ec ci izens while enabling
compe i i eness in global AI ma ke s. The s udy
concludes wi h co esponding ecommenda ions.
In oduc ion
The AI economy is ad ancing apidly
echnologically and a ac ing subs an ial
in es men in lows. This is being uelled, in pa ,
by p i a e sec o esea ch and de elopmen
(R&D) expendi u e in a eas anging om chip-
making equipmen o so wa e de elopmen and
comme cial applica ions (Lippold 2024). The e is
signi ican geog aphic concen a ion in la ge-scale
AI inno a i e ac i i y, which is cen ed in oughly
a dozen mos ly de eloped coun ies, oge he wi h
2 An engagemen is de ined he e as a le el o in es men su icien o place
an in es o among he op i e in a i m.
2CIGI Pape No. 323 — June 2025 • Douglas Lippold
China and India.3 Gi en he assessed po en ial o
AI o con ibu e subs an ially o u u e economic
g ow h, his concen a ion in AI inno a ion
p esen s a challenge o de eloping coun ies
o gain access o and sha e in he bene i s
o his inno a ion while also managing he
isks. This geog aphic concen a ion could also
pose a challenge o i ms in o he pa s o he
wo ld looking o ap in o he la es gene a ion
o AI echnology. A ques ion emains as o
whe he he e is po en ial o imp o ed policy
alignmen o o he adjus men o acili a e
u he de elopmen and echnology ans e
in he sec o , wi h app op ia e sa egua ds.
In his pape , using a compa a i e case s udy
app oach, he au ho examines how s a -ups
in a sample o 10 middle-income de eloping
coun ies engage wi h he global AI economy
and conside s domes ic policy condi ions ha
acili a e hei pa icipa ion. These issues a e o
impo ance no only o he di ec s akeholde s
in such businesses bu also o hose conce ned
wi h economic de elopmen mo e gene ally.
In e na ional di usion and applica ion o he
eme ging AI-d i en echnologies om leading
AI economies, along wi h complemen a y local
inno a ion in o he economies, may oge he
con ibu e o imp o ed wel a e in de eloping
coun ies, p o ided isks a e managed. Such isks
may come om in en ional o acciden al ha ms
di ec ly associa ed wi h AI ( o example, due o
biases), as well as o he challenges o cons ain s
such as egula o y impedimen s, adjus men cos s
o a ec ed indi iduals and social esis ance o
echnological change, among o he conside a ions.4
Whe e o Plug In?
De elopmen o AI ounda ion models equi es
signi ican scale and a subs an ial esou ce base
o amass he necessa y compu ing powe and
3 Lippold (2024, 14, 18) ound he leading AI-in ensi e, inno a ion-d i en
i ms (in e ms o R&D expendi u e) o be concen a ed geog aphically.
As o 2021, he co po a e headqua e s we e loca ed in jus 11 coun ies:
Canada, China, Finland, Ge many, India, I eland, Is ael, Japan, Sou h
Ko ea, he Uni ed Kingdom and he Uni ed S a es. These i ms we e o en
clus e ed in pa icula egions, such as Beijing, London o Silicon Valley.
4 Ian B emme and Mus a a Suleyman (2023), o example, highligh key
isks.
da a.5 Few i ms o ins i u ions in he wo ld ha e
su icien capaci y o ackle such a challenge. Fo
example, acco ding o S an o d Uni e si y’s 2024
A i icial In elligence Index Repo (Maslej e al.
2024, 5), OpenAI’s GPT-4 equi ed an es ima ed
US$78 million wo h o compu a ional powe o
ain in 2023. E en hough some less expensi e
s a egies o aining appea o be eme ging
ia challenge s such as DeepSeek, he up on
cos s emain subs an ial.6 In 2023, China, he
Eu opean Union and he Uni ed S a es oge he
p oduced less han 100 no able AI models,
acco ding o he 2024 AI Index Repo (ibid.).
Bu , o una ely, his is no he ull s o y. The e is
oom in he AI economy o smalle de elope s
o build downs eam applica ions ha employ
he la ge models as a base.7 Businesses may
also de elop “small language models” ha
ha e a compac design and possess a ewe
pa ame e s han a la ge language model (LLM)
— pe haps jus one-six h o less o he numbe
o pa ame e s. Ne e heless, hey can excel a
deli e ing AI se ices o speci ic asks, such as
analyzing cus ome eedback o gene a ing p oduc
desc ip ions. And he e a e oppo uni ies o be
ound in many simila niches in he deli e y o
AI se ices o cus ome s in de eloping coun ies,
om building ou ha dwa e and so wa e
capaci y o human capi al de elopmen (in
o he wo ds, skills) and public in as uc u e
o connec i i y, among o he possibili ies.
Policy make s in de eloping coun ies ha e a
subs an ial ole o play in shaping he domes ic
condi ions o business. Thei choices may
con ibu e o a na ion’s abili y o ap in o he
bene i s o he AI economy while sa egua ding
agains isks. O cou se, AI is no a s and-alone
solu ion o he ull ange o cons ain s on
5 In a web pos i led “Re lec ions on Founda ion Models,” Rishi Bommasani
and Pe cy Liang (2021) o S an o d Uni e si y’s Ins i u e o Human-
Cen e ed A i icial In elligence de ine ounda ion models as “models
ained on b oad da a (gene ally using sel -supe ision a scale) ha can
be adap ed o a wide ange o downs eam asks.” They “see ounda ion
models as he subjec o a g owing pa adigm shi , whe e many AI
sys ems ac oss domains will di ec ly build upon o hea ily in eg a e
ounda ion models.”
6 Some ounda ion model de elope s ha e de ised s a egies o educe
aining cos s, in pa , by ading o some e inemen in he esul s.
Chinese s a -up DeepSeek amously ained i s debu model o jus
US$6 million, hough ha numbe allegedly excludes subs an ial up on
cos s incu ed p io o he launch o aining. See Me z (2025).
7 Fo example, see he discussion in a web pos om Amazon Web
Se ices (h ps://aws.amazon.com/wha -is/ ounda ion-models/) and om
he Wo ld Economic Fo um (Whi ing 2025).
9De eloping Coun ies’ Business Pa icipa ion in he AI Economy
egula ion ha suppo s bo h expo -o ien ed
and domes ically o ien ed indus ializa ion.
Ezekiel T. Mu asa, Chi a Dhiwwale and Sunda an
Saga an A. Gopal (2024) examine AI applica ions,
p ospec s and challenges in de eloping economies
ac oss key sec o s, including manu ac u ing,
ag icul u e, e ail, inancial se ices, heal h ca e
and mining. They ake a me a app oach, d awing on
empi ical s udies ha mos ly da e om he pe iod
p io o he eme gence o publicly accessible GenAI.
The au ho s iden i y signi ican bene i s o i m-
le el AI adop ion, such as inc eased p oduc i i y
and inno a ion. The ne gains a e de i ed ia a
a ie y o channels such as imp o ed decision
making, quali y con ol, p edic i e main enance,
ools o op imiza ion o esou ce use and o
educe was age, be e a ge ing o unde se ed
popula ions and imp o ed heal h-ca e sc eening.
They also highligh subs an ial challenges o
success ul i m-le el AI adop ion, including
da a p i acy conce ns, high implemen a ion
cos s, inadequa e digi al in as uc u e, skill
sho ages and esis ance o change.
Mu asa, Dhiwwale and Gopal (2024) highligh policy
ecommenda ions o middle-income coun ies:
→ encou age AI adop ion while add essing e hical
and sociocul u al conside a ions;
→ in es in c i ical in as uc u e, especially in
u al a eas;
→ employ change managemen s a egies o
add ess esis ance o adop ion;
→ suppo aining p og ams o b idge he AI alen
gap;
→ use public-p i a e pa ne ships o mi iga e high
cos s; and
→ add ess c oss-sec o al challenges such as poo
da a quali y and in eg a ion.
A he i m le el, hey ad ise ha
managemen should:
→ look o sec o -speci ic oppo uni ies o apply AI;
→ build on exis ing in as uc u e ( o example, in
mobile echnology);
→ ocus on p ac ical, high-impac solu ions;
→ ailo solu ions o local con ex s ( o example,
unde se ed domes ic popula ions); and
→ seek collabo a ion be ween go e nmen , p i a e
sec o and o he local s akeholde s o o e come
adop ion ba ie s and scale AI solu ions.
Xueyuan Gao and Hua Feng (2023) conside he
ollou o AI in China du ing he pe iod p io o
he ecen leap o wa d in GenAI capabili ies. They
examine how AI adop ion a ec s manu ac u ing
i ms’ p oduc i i y in China. Using mic o-
le el da a om 2010 o 2021, hey ound ha
each pe cen age inc ease in AI pene a ion
was associa ed wi h la ge gains in o al ac o
p oduc i i y. The au ho s iden i ied h ee
mechanisms o his: alue-added enhancemen
(imp o ing p oduc quali y and p oduc ion
p ocesses); skill-biased enhancemen (shi ing
owa d highe -skilled wo ke s); and echnology
upg ading. Implemen a ion o AI was ound o
s imula e inno a ion, including wi h espec o
u he AI inno a ions. Rega ding AI policy, he
au ho s made he ollowing ecommenda ions:
→ a ge any AI subsidies a capi al- and
echnology-in ensi e indus ies, which bene i
mo e han labou -in ensi e ones;
→ conside ma ke s uc u e ( i ms in indus ies
wi h high concen a ion expe ienced g ea e
p oduc i i y gains om AI, pe haps due o scale
e ec s);
→ ake i m owne ship in o accoun , as p i a e
en e p ises ealized signi ican p oduc i i y
imp o emen s while s a e-owned en e p ises
did no , sugges ing ins i u ional e o ms may be
needed alongside echnology adop ion; and
→ accompany AI implemen a ion wi h
complemen a y in es men s in human capi al
(AI d i es demand o highly skilled wo ke s).
Da id Helle and Dominik Asam (2024) examine
he impac o GenAI on s a -up p oduc i i y,
compa ing i ms in he so wa e sec o o o he
sec o s. Using Gi Hub Copilo ’s elease as a
quasi-na u al expe imen , he au ho s employed
C unchbase da a co e ing 21,834 s a -ups ha
secu ed ini ial unding be ween Q1 2020 and
Q3 2023. They ound ha so wa e-de eloping
s a -ups expe ienced a 20 pe cen educ ion in
ime- o- unding (an ea ly indica o o p oduc i i y)
ela i e o o he s a -ups p io o Copilo ’s
elease. The e ec s we e mos p onounced o

10 CIGI Pape No. 323 — June 2025 • Douglas Lippold
s a -ups wi h ounde s possessing echnological
o manage ial expe ience, sugges ing ha
GenAI se es as a compe i i e ad an age when
combined wi h complemen a y human capi al.
Thei indings indica e ha Gi Hub Copilo may
subs i u e o adi ional esou ces such as junio
p og amme s while enhancing he p oduc i i y
o expe ienced en ep eneu s. This is o pa icula
impo ance du ing ea ly s a -up s ages when
esou ce cons ain s a e mos binding.
Nicholas O is e al. (2024) conduc ed a ield
expe imen wi h 640 Kenyan en ep eneu s o
e alua e he impac o a GenAI business assis an
(GPT-4-powe ed) deli e ed ia Wha sApp. While
he s udy ound no signi ican a e age ea men
e ec on business pe o mance (measu ed as
p o i s and e enues), i e ealed subs an ial
he e ogenei y based on p e ea men pe o mance
le els. High-pe o ming en ep eneu s ended o
bene i o e all, wi h an app oxima ely 15 pe cen
imp o emen in pe o mance. Low pe o me s
ended o expe ience nega i e pe o mance, wi h
an o e all decline o eigh pe cen . The di e gen
e ec s s emmed om he manne in which he
en ep eneu s selec ed and implemen ed he AI’s
sugges ions. High pe o me s ended o be e
iden i y oppo uni ies o speci ic imp o emen s,
while low pe o me s disp opo iona ely
implemen ed gene ic ad ice ( o example, p ice
discoun s) ha ha med hei businesses.
Li e a u e Re iew: A Summing Up
This e iew o he li e a u e spans economic
impac s, go e nance challenges and
implemen a ion hu dles. I p o ides a basis
o conside he ans o ma i e oppo uni ies
opened by AI o middle-income coun ies ( o
example, imp o ed p oduc i i y, inno a ion and
sec o -speci ic solu ions), as well as he isks ( o
example, widening inequali y, job displacemen
and egula o y challenges). Tension be ween AI
na ionalism and in e na ional coope a ion is
highligh ed, wi h schola s such as Aa onson no ing
how di e gen egula o y app oaches may c ea e
ma ke access ba ie s o de eloping na ions.
Fi m-le el assessmen s e eal ha success ul AI
adop ion depends, in pa , on con ex ual ac o s
including skills de elopmen , in as uc u e
eadiness and app op ia e egula o y amewo ks.
Se e al s udies emphasize ha while AI has he
po en ial o accele a e economic de elopmen ,
his ou come equi es delibe a e policy design
balancing inno a ion wi h e hical go e nance
(especially ega ding da a go e nance challenges).
The link be ween AI s a -up de elopmen
and na ional economic wel a e can be seen
ia esea ch showing ha i ms wi h local
knowledge can c ea e AI solu ions ailo ed
o local con ex s. This may po en ially c ea e
compe i i e ad an ages o such i ms while
also add essing na ional de elopmen p io i ies
such as heal h ca e, ag icul u e and inancial
inclusion ( o example, Maye 2021; Mu asa,
Dhiwwale and Gopal 2024; Mannu u e al. 2023).
AI S a -Ups and he Case
S udy Coun ies
This sec ion p esen s he i m-le el analysis o he
case s udy. I begins wi h he p incipal exe cise:
a e iew o AI s a -up i ms wi h G7 in es o
engagemen in he case s udy coun ies. This
is ollowed by a b ie assessmen o AI unico n
i ms, wi h conside a ion o hese success ul i ms
in ela ion o he case s udy o AI ecosys ems.
Case S udy Coun ies
The cen al ocus he e is on he p ospec s o
de eloping coun ies o connec and in eg a e in o
he apidly expanding AI economy, aking in o
accoun he ole o AI s a -ups. These a e young,
ac i e, p i a ely held businesses. They cons i u e
a ca ego y ha has demons a ed a pa icula
dynamism in he ield o AI. They a e also likely o
play an impo an ole in he in eg a ion o AI in o
de eloping economies. Fo example, his migh
a ise ia he exploi a ion o niche oppo uni ies o
de elop ailo ed AI applica ions o small language
models o add ess local condi ions in a de eloping
coun y o egion.21 Gi en ha many high-income
coun ies a e al eady leading in AI de elopmen s
o posi ioning o wide adop ion o AI, his case
s udy a ge s he nex ie s o coun ies by income.
In o de o ensu e di e si y o co e age, an
illus a i e sample o 10 middle-income coun ies
was selec ed om ac oss Asia, A ica (( he Middle
Eas and No h A ica [MENA] and Sub-Saha an
21 The alue o exploi ing his ype o oppo uni y is suppo ed by some
o he empi ical wo k ci ed in he li e a u e e iew abo e ( o example,
Maye 2021; Mu asa, Dhiwwale and Gopal 2024).
11De eloping Coun ies’ Business Pa icipa ion in he AI Economy
A ica) and La in Ame ica. The sample includes i e
lowe -middle-income coun ies and i e uppe -
middle-income coun ies, as classi ied by he Wo ld
Bank (Table 1). Box 2 p esen s an o e iew o he
coun ies no selec ed o inclusion in he sample.
The au ho ’s selec ion o speci ic coun ies
was u he guided by da a a ailabili y wi h
espec o s a -up i ms in he AI sec o and
he a ailabili y o s anda dized AI policy and
egula o y in o ma ion. Da a o he i m-le el
analysis was d awn om he C unchbase da a se ,22
which co e s s a -ups globally and has de ailed
desc ip i e in o ma ion abou s a -up i ms and
hei in es o coun e pa s. In o de o a ge
ma ke s wi h a leas nascen AI ecosys ems, he
au ho selec ed sample coun ies wi h mo e han
50 s a -up i ms lis ed as ha ing he key wo ds
“a i icial in elligence” in hei mission o pu pose
desc ip ions. A complemen a y s a -up analysis
was de eloped, d awing on he global unico n
da a se p oduced by CB Insigh s.23 This co e s
22 Fo a b ie o e iew o C unchbase da a s eng hs and weaknesses, see
Appendix 1.
23 See www.cbinsigh s.com/ esea ch-unico n-companies.
s a -ups ha a e s ill p i a e and ha ha e
achie ed a ma ke alua ion o US$1 billion o
mo e. Finally, wi h espec o policy analysis, he
esea ch u ned p ima ily o he OECD AI Policy
Obse a o y, which has b oad coun y co e age,
and s anda dized and eadily accessible policy
and egula o y da a.24 This was supplemen ed and
upda ed using he Digi al Policy Ale da abase25,
as well as egional and na ional sou ces.
On his basis, he coun y sample was es ablished
o include B azil and Colombia; Egyp , Kenya,
Nige ia, Sou h A ica and Tunisia; and Indonesia,
Thailand and Vie nam (Figu e 1). As can be seen
in Table 1, hese na ions ep esen a b oad ange
in e ms o economic scale and popula ion. The
sample is illus a i e o a di e se g oup o coun ies
ha ha e — o a ying deg ees — achie ed some
24 Fo an o e iew o he con en and coun y co e age o he OECD AI
Policy Obse a o y da abase, see Appendix 1 and
h ps://oecd.ai/en/dashboa ds/o e iew.
25 Fo an o e iew and a link o he da abase websi e, see Appendix 1.
26 The economic da a ci ed in his pa ag aph is d awn om he au ho ’s
abula ions and he Wo ld Bank’s Wo ld De elopmen Indica o s,
a ailable a h ps://da a.wo ldbank.o g/indica o . The AI s a -up da a is
om C unchbase and is a ailable a www.c unchbase.com/.
27 The numbe s in pa en heses ep esen he numbe o AI s a -ups in each
o hese coun ies as o Q1 2025.
Box 2: Economies No Selec ed o he AI S a -Up Sample
The 91 middle-income de eloping coun ies no co e ed di ec ly in his case s udy ep esen
a slice o he global economy ha is oughly i e imes g ea e han he collec i e sha e o he
case s udy economies (US$30.6 illion e sus US$6.1 illion in 2023, in cu en US dolla s).27
The middle-income coun ies no co e ed we e somewha less well o , on a e age, han he
case s udy coun ies by some indica o s. Fo example, hose no co e ed had a median GDP
(pu chasing powe pa i y [PPP]) pe capi a o US$11,245 in 2023 e sus US$15,304 o he sample
coun ies. A po ion o his di e en ial may e lec a selec ion bias in ha he au ho se ou
o ind middle-income coun ies ha had al eady demons a ed some engagemen in he AI
economy, wi h a leas 50 o mo e AI s a -ups. This choice was a na u al consequence o he
au ho ’s esea ch design o examine he nascen AI sec o ac i i y in an illus a i e sample o
middle-income coun ies. While some o hose no selec ed we e compe ing in he AI sec o
( o example, Pakis an wi h 262 AI s a -ups, Ghana wi h 53 o Pe u wi h 70), many coun ies
no selec ed ha e each aced a unique combina ion o challenges ha p e en ed he eme gence
o an adequa e numbe o AI s a -ups. Examples o middle-income de eloping coun ies
wi h less han he au ho ’s minimum selec ion h eshold o 50 AI s a -ups and wi h GDP
(PPP) pe capi a below he au ho ’s AI sample median include Hondu as ( wo), Laos (ze o),
Libya (one) and Tanzania (10).28 Some challenges o AI s a -up de elopmen — o example
wi h espec o ma ke openness o human capi al de elopmen — may also ha e economic
causes o consequences and ha may be e lec ed in he lowe median pe capi a incomes.
12 CIGI Pape No. 323 — June 2025 • Douglas Lippold
ac ion in he AI economy. Thei si ua ion may
o e some use ul insigh s o u he de elopmen
o he AI economy in o he de eloping na ions.
Con ex o AI Implemen a ion in De eloping
Coun ies
The e is b oad awa eness o AI ac oss a subs an ial
sha e o he popula ion in de eloping coun ies.
This is e idenced, o example, by ela i ely
equen use o Cha GPT-4 (Maslej e al. 2024,
chap e 9, 449). As can be seen in Table 2, o
coun ies co e ed by he Ipsos su ey da a,28
a majo i y o adul s eel posi i ely abou AI
bene i s and us wo hiness. This may be
associa ed wi h he ea ly s age o he sec o ’s
de elopmen in hese coun ies, bu i ce ainly
con as s s ikingly wi h he iew o much o
he public in he mo e ad anced economies.
28 See Appendix 1 o de ails o he Ipsos su ey.
In ou sample coun ies, some o he op imism
may be associa ed wi h he you h ulness o he
popula ion. Fo mos o hese coun ies, he
median age is a ound he global median o younge
(Thailand is an excep ion in his ega d).29 Age
appea s o be one ac o in luencing a i udes o
AI, a poin no ed in S an o d Uni e si y’s 2024 AI
Index Repo (ibid., 438). Indeed, i may be ha
such public awa eness and posi i e a i udes
con ibu es an impulse owa d in e es in AI
en ep eneu ship in he sample coun ies.
These posi i e a i udes seem o ha e ca ied
o e o many cu en business leade s. A su ey
o business execu i es by he Wo ld Economic
Fo um (WEF) in 2024 ound ha esponden s in
eigh o he sample coun ies did no ank “ isk
o ad e se ou comes om AI echnologies” as a
29 This is based on da a a ailable a www.cia.go / he-wo ld- ac book/ ield/
median-age/coun y-compa ison/.
Figu e 1: Case S udy Coun ies
Colombia B azil Egyp Indonesia Kenya Nige ia
Sou h A ica Thailand Tunisia Vie nam
Sou ce: Mic oso Excel map; au ho ’s abula ions.
13De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table 1: Case S udy Coun ies, O e iew
Economy
Income S a us GDP Popula ion AI S a -Up Fi ms,
as Reco ded in he
C unchbase Da abase
FY2025, Based
on 2023 pe
Capi a GDP
2023, Cu en
US$, Billions
Pe cen
o Wo ld
To al
2023,
Millions
Pe cen
o Wo ld
To al
Numbe
Ac i e, as o
1Q2025
Pe cen
o Wo ld
To al
B azil Uppe middle
income
2,173.7 2.0 211.1 2.6 973 1.1
Colombia Uppe middle
income
363.5 0.3 52.3 0.6 348 0.4
Egyp Lowe middle
income
396.0 0.4 114.5 1.4 98 0.1
Indonesia Uppe middle
income
1,371.2 1.3 281.2 3.5 199 0.2
Kenya Lowe middle
income
108.0 0.1 55.3 0.7 79 0.1
Nige ia Lowe middle
income
363.8 0.3 227.9 2.8 230 0.3
Sou h A ica Uppe middle
income
380.7 0.4 63.2 0.8 294 0.3
Thailand Uppe middle
income
515.0 0.5 71.7 0.9 80 0.1
Tunisia Lowe middle
income
48.5 0.0 12.2 0.2 50 0.1
Vie nam Lowe middle
income
429.7 0.4 100.4 1.2 186 0.2
Wo ld
o al (all
economies)
106,170.0 100.0 8,061.9 100.0 86,235 100.0
Sou ce: www.c unchbase.com/disco e /o ganiza ion.companies; Me eau, Young and Eapen (2024);
h ps://da a.wo ldbank.o g/indica o /NY.GDP.MKTP.CD2024; h ps://da a.wo ldbank.o g/indica o /SP.POP.TOTL.
No e: Fo he iscal yea beginning July 1 2024, he Wo ld Bank de ines lowe -middle-income coun ies as
ha ing g oss na ional income pe capi a anging be ween US$1,146 and US$4,515; uppe -middle-income
coun ies a e de ined as ha ing g oss na ional income pe capi a anging be ween US$4,516 and US$14,005.
Tunisia’s sha e in wo ld o al GDP is 0.05. C unchbase elies on a a ie y o sou ces and is in pa c owdsou ced,
so he e may be a ia ion in he quali y o he da a om coun y o coun y (see Appendix 1).
14 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Table 2: Popula Pe cep ions o AI
Public eelings abou AI
P oduc s and se ices using AI ha e
mo e bene i s han d awbacks (%
ag ee “ e y” o “somewha ”)
Public us in AI
I us companies ha use AI as
much as I us o he companies (%
ag ee “ e y” o “somewha ”)
Execu i es' opinions on " isk o ad e se
ou comes o AI echnologies": Is his a op- i e
isk o e he nex wo yea s in you coun y?
I “yes,” hen ank is gi en (1 = mos ci ed
o 5 = i h mos ci ed); i “no,” hen “no”
Global coun y
(simple) a e age 54 52 Yes, #1
Indonesia 78 69 Yes, #1
Thailand 74 73 No
Mexico 73 66 No
Malaysia 69 70 No
Pe u 67 60 No
Tü kiye 67 65 No
Sou h Ko ea 66 55 No
Colombia 65 56 No
India 65 67 No
B azil 64 60 No
Singapo e 64 57 Yes, #5
Romania 61 62 No
Sou h A ica 59 55 No
Chile 59 51 No
A gen ina 57 52 No
I aly 55 53 No
Japan 52 44 No
Spain 50 49 No
Hunga y 48 46 No
Poland 47 50 No
G ea B i ain 46 45 Yes, #4
New Zealand 44 43 No
The Ne he lands 43 44 No
Ge many 42 45 No
I eland 40 39 No
Aus alia 40 42 No
Belgium 39 39 No
Sweden 39 42 n/a
Canada 38 39 Yes, #5
F ance 37 37 No
Uni ed S a es 37 36 Yes, #3
Egyp n/a n/a No
Kenya n/a n/a No
Nige ia n/a n/a No
Tunisia n/a n/a No
Vie nam n/a n/a Yes, #1
Sou ce: Public eelings and public us : Ipsos (2023, 9 and 15); Execu i es’ opinion o AI isk: Elsne e al. (2025, Appendix
C, 81–91).
No e: See Appendix 1 o de ails o he Ipsos su ey and he WEF su ey.

15De eloping Coun ies’ Business Pa icipa ion in he AI Economy
op isk o e he nex wo yea s (Table 2).30 The
excep ions we e Indonesia and Vie nam, whe e
execu i es did a e AI as he op isk. This may be
associa ed wi h he de elopmen and ad ancing
implemen a ion o AI go e nance in he pa icula
cul u al con ex o hose coun ies which may
aise awa eness o isks, as well as sec o -speci ic
isk issues, among o he possible explana ions.
Obs acles o Business
S a -ups a e young companies and ulne able
o a a ie y o challenges, such as exhaus ing
hei liquidi y, misjudging ma ke demand o a
p oduc and un a ou able egula o y changes,
among many o he s. S a -ups o en ail. E en in
coun ies such as he Uni ed Kingdom and he
Uni ed S a es, many s a -ups — and, in some
yea s, mos — do no su i e pas hei i h o
six h anni e sa ies.31 In he s udy coun ies, i
is unlikely ha pe o mance is much be e .
To p o ide a glimpse in o business pe cep ions
o obs acles in he case s udy coun ies, he
pape u ns o he Wo ld Bank En e p ise
Su eys. These ha e been conduc ed in dozens
o coun ies, including each o he 10 case s udy
coun ies and h ee o he G7 economies. China is
included he e o he sake o compa ison. Table
3 p esen s a abula ion o i ms’ pe cep ions in
each coun y o hei bigges obs acles. Businesses
a e su eyed ac oss a b oad ange o sec o s
and i m sizes and do no necessa ily ha e a
ocus on AI (see he able no es o de ails).
Fo mos o he coun ies, he obs acles shown
accoun o a majo i y o he op conce ns ci ed
by businesses. The able also e eals a s iking
con as be ween he case s udy coun ies and
he G7 coun ies. Fo mos o hese na ions,
excep B azil, “poli ical ins abili y,” “access o
elec ici y” and “access o inance” a e among
he leading ca ego ies o obs acles among hose
30 Responden s could selec isks om among 34 op ions ac oss i e
ca ego ies including economic, en i onmen al, geopoli ical, socie al and
echnological isks. See he Key Sou ces Annex o de ails o he WEF
su ey.
31 Fo example, among all US p i a e businesses es ablished in in he
yea ending Ma ch 2018, only 51.9 pe cen su i ed i e yea s on, and
by Ma ch 2024, only 47.5 pe cen su i ed. See www.bls.go /bdm/
us_age_naics_00_ able7. x , accessed Oc obe 30, 2024. Acco ding
o he Wo ld Bank G oup “P ospe i y Da a360” online da abase, he
i e-yea su i al a e o UK businesses in he “ o al indus y, cons uc ion
and ma ke se ices excep holding companies” sec o was jus 31
pe cen as o 2018 (in o he wo ds, 69 pe cen had ailed). See h ps://
p ospe i yda a360.wo ldbank.o g/en/indica o /OECD+BDI+YS5_R.
shown in he able. One o mo e o hese basic
ope a ional conside a ions was ci ed by mo e
han 15 pe cen o businesses su eyed in each
o he case s udy coun ies (excluding B azil).
In Egyp , Kenya and Vie nam, in o mal sec o
compe i ion also weighed as a op conce n, being
ci ed by mo e han 15 pe cen o businesses in
hose h ee na ions. Chinese businesses also ci ed
access o inance and in o mal sec o compe i ion
as op conce ns. In addi ion, in h ee o hese
economies, 10 pe cen o mo e o businesses anked
“co up ion” as he op obs acle. B azil was an
ou lie among he de eloping na ions, wi h a p o ile
mo e closely esembling he G7 na ions shown.
Fo he G7 economies co e ed in he able,
hose issues we e much less equen ly ci ed by
esponden s. Ins ead, many G7 i ms poin ed o
sho alls in he a ailabili y o adequa ely ained
wo k o ce pa icipan s as being a op conce n,
by a . I is no able, howe e , ha mo e han 10
pe cen o businesses in B azil, China and Vie nam
also ci ed his as a op conce n. This la e poin
is an o -ci ed conce n o i ms de eloping o
implemen ing AI sys ems a ound he wo ld.32
AI s a -ups may be pa icula ly ulne able o
cons ain s such as high capi al needs, pa chy
b oadband and limi ed a ailabili y o human
esou ces, which can di ec ly hampe AI solu ion
deploymen o scaling. Fo example, a ecen
B ookings Ins i u ion s udy on le e aging AI o
suppo A ica’s economic de elopmen (Signé
2025) poin s o he ollowing cons ain s.
→ limi ed digi ized da a a ailabili y wi h espec
o A ica ( o example, only 0.02 pe cen o o al
in e ne con en is in A ican languages);
→ limi ed a ailabili y o digi al skills and ele an
human capi al (uni e si ies in A ica a e
in oducing AI cou ses, bu o en he e is a lack
oppo uni ies o hands-on lea ning); and
→ limi ed R&D expendi u e lows o A ica, wi h
A ican use-case de elopmen o en neglec ed.
Mo eo e , such challenges a e magni ied
by ope a ional cons ain s ela ed o poo
in as uc u e and high cos s o add ess he
bo lenecks ( o example, o de elop cloud
compu ing capaci y). S a -ups seeking o add ess
32 Fo mo e on AI sec o skills demand, see Maslej e al. (2024, chap e 4,
sec ion 2, on jobs; chap e 6 on educa ion)
16 CIGI Pape No. 323 — June 2025 • Douglas Lippold
hese challenges may be cons ained by he
e y condi ions in which hey a e ope a ing,
which con ibu e o isk a e sion on he pa o
some in es o s and inancial ins i u ions, and
sho alls in he a ailabili y o needed capi al.
S a ing wi h a Few De ini ions
Be o e p oceeding o he de ailed analysis, he
pape i s p esen s a ew de ini ions. Fo he
pu poses o his analysis, he pape de ines “AI
s a -up” as a i m ha has “AI” in eg a ed as
pa o i s mission s a emen o pu pose. The
pape did no impose a s ic age limi , bu in
he s udy popula ion, he median AI s a -up
age by coun y is young, anging om i e yea s
in Colombia o nine yea s in Sou h A ica. The
pape only conside ed i ms ha a e cu en ly
ac i e. Gene ally, he AI s a -up i ms in he
assessmen we e no a he mo e ma u e s ages
o comme cial de elopmen , such as p epa ing
o an ini ial public o e ing (IPO) (see Box 3).
Unico ns a e conside ed in a subsequen sec ion
and end o be mo e ma u e, in he mid- o-
la e s ages o de elopmen o a s a -up.
Table 3: Business Pe cep ions o Thei Bigges Obs acles, Selec ed Conce ns, Mos Recen Yea
A ailable (% o Fi ms Responding)
Coun y Access o
Elec ici y
Co up ion Access o
Finance
Inadequa ely
Educa ed Wo k o ce
In o mal
Sec o
Poli ical
Ins abili y
Sub o als (ou o
100%, by Coun y)
B azil 0.3 3.3 7.5 12.6 12.4 3.0 39.1
Colombia 4.6 8.7 7.5 5.2 8.2 39.0 73.2
Egyp 2.9 6.2 10.7 2.2 15.5 25.7 63.2
Indonesia 1.6 10.1 28.8 3.4 7.3 11.3 62.5
Kenya 3.0 7.7 18.3 1.6 22.9 17.0 70.5
Nige ia 27.2 12.7 30.2 0.4 4.3 4.4 79.2
Sou h A ica 54.6 5.8 16.1 0.0 0.8 13.3 90.6
Thailand 19.8 2.3 4.6 2.6 5.0 20.3 54.6
Tunisia 0.6 15.0 39.4 5.4 8.3 11.5 80.2
Vie nam 4.2 0.3 21.2 11.7 22.1 3.6 63.1
De ’g Coun y A g. 11.9 7.2 18.4 4.5 10.7 14.9 55.7
China 4.8 1.2 22.4 13.0 19.6 0.8 61.8
F ance 3.3 1.5 2.8 23.8 11.1 4.3 46.8
Ge many 0.2 0.4 5.0 53.5 4.1 6.2 69.4
I aly 4.5 1.1 5.4 18.8 3.8 5.5 39.1
Sou ce: Sou ce: Wo ld Bank (2025), “En e p ise Su eys,” Global Indica o s Depa men , Da a Visualiza ion, h ps://www.
en e p isesu eys.o g/en/g aphing- ool.
No es: 1) In he Table, “De ’g Coun y A g” e e s o he a e age sco es ac oss he 10 case s udy coun ies. “Sub o als”
e e s o he ally o pe cen ages o he six ca ego ies shown o each coun y. Fo mos coun ies, he able cap u es a
majo i y o he op obs acles iden i ied by esponden s (NB, each esponden could only selec one op obs acle). 2) The
su eys co e egis e ed i ms wi h i e o mo e employees and one pe cen o mo e o p i a e owne ship. The su ey
co e s mos o he p i a e sec o . See Appendix 1 o de ails. 3) The su ey esponden s could selec among 15 ca ego ies
o bigges obs acle. The six p esen ed he e we e selec ed o hei gene ally high equency in he case s udy coun ies
and hei ele ance o AI i ms (e.g., as opposed o i ems such as “access o land” o “c ime, he and diso de ”). See he
Da a Visualiza ion page, linked abo e, o he ull lis o obs acles o which da a is a ailable. By sub ac ing he sub o al
numbe om 100 pe cen o each coun y, he eade can ob ain he o al alue o he omi ed ca ego ies. 4) Su ey
yea s a e as ollows: B azil (2009), China (2012), Colombia (2023), Egyp (2020), F ance (2021), Ge many (2021), Indonesia
(2023), I aly (2024), Kenya (2018), Nige ia (2014), Sou h A ica (2020), Thailand (2016), Tunisia (2020), Vie nam (2023).
17De eloping Coun ies’ Business Pa icipa ion in he AI Economy
In looking a he comme cial linkages o he
AI s a -ups, he pape ocuses on “in es o
engagemen ,” de ined as a case whe e an in es o
has su icien ly suppo ed an AI s a -up o each
he le el o being a op- i e in es o in he i m.
This may in ol e one o mo e “deals,” which
e e o ansac ions whe e an in es o p o ides
capi al in exchange o an equi y s ake in he i m
o con e ible deb . A e a gene al in oduc ion
o capi al lows in he sec o (including deals),
he pape hen u ns o use in es o engagemen
o he assessmen o he AI s a -ups. The au ho
chose in es o engagemen a he han deal
coun s in o de o ocus on he mo e subs an ial
inancial ela ionships, which can en ail addi ional
comme cial suppo o an AI s a -up.
Figu e 2: To al Qua e ly AI Funding Deal Coun s, Q1 2020–Q3 2024, Sha es by Region (%)
0
10
20
30
40
50
60
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
2020 2021 2022 2023 2024
Asia
Eu ope
Canada
Uni ed S a es
La in Ame ica Canada All O he RegionsA ica
Eu ope
AsiaUni ed S a es
Sou ces: See www.cbinsigh s.com/ epo s/CB-Insigh s_AI-Repo -2024.xlsx; au ho ’s abula ions.
No e: The A ica se ies is de i ed using p opo ions om he deal da a o 2020 o 2024 in CB Insigh s Q3
2024 applied o he upda ed da a se in www.cbinsigh s.com/ epo s/CB-Insigh s_AI-Repo -2024.xlsx.
18 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Si ua ing Ou Case S udy in
he Global VC Con ex
In o de o si ua e he assessmen o AI s a -ups
in he 10 case s udy coun ies, i is use ul o s a
wi h a global pe spec i e. Figu e 2 p esen s an
assessmen o he global sha es in AI s a -up deal
unding by egion. As can be seen, he Uni ed S a es
has a disp opo iona e sha e o he o al, ollowed
by Asia and Eu ope. Canada is nex , ollowed
by La in Ame ica, all o he egions (including
Oceania) and A ica. The exac anking depends on
de elopmen s in each qua e . Bu i can be seen
om he da a ha wo o he h ee egions home
o he s udy coun ies accoun o jus a small
sha e o he global VC lowing o AI s a -ups. (A
sepa a e b eakdown om Asia o he hi d egion,
Sou heas Asia, is no a ailable in his da a se .)
The compa a i ely small numbe s do no mean
ha he case s udy coun ies a e unimpo an o
AI, howe e . In he i s place, hey collec i ely
ha e some ma ke scale, accoun ing o 1.2billion
people and US$6.2 illion o GDP (in o he wo ds,
jus unde six pe cen o global GDP) (see Table
1). Mo eo e , as no ed abo e, mos o hese
coun ies ha e you h ul popula ion age p o iles
ha all a ound o below he global median age,
meaning ha hese ma ke s may be posi ioned
o be qui e dynamic going o wa d. (Thailand is
he excep ion o his me ic, wi h a median age
10 yea s abo e he global a e age.) In addi ion,
he e is some no able AI de elopmen unde way
al eady, including AI en ep eneu ship and
business linkages o ad anced economy pa ne s,
as well as na ional-le el AI policy ac ions.
To gi e a sense o he scale o AI s a -up ac i i y,
Table 4 p o ides an o e iew o AI s a -ups and
VC deals by egion o wo ecen qua e s (Q3
and Q4 2024).33 This analysis e eals ha he e is
a ce ain amoun o ac i i y in La in Ame ica and
A ica hin ing a po en ial in e linkages o he
inancial mains eam. Running ou o cash and
ailu e o aise new capi al a e he op easons
s a -ups ail, accoun ing o nea ly wo- i hs o
all ailu es (CB Insigh s 2021, 4). These unding
deals play an impo an ole o s a -ups seeking
o a oid such a a e. Bu hey can also help an
AI s a -up o plug in o he mains eam. They
can deli e non- inancial bene i s in he o m
o ad ice ela ed o echnology, manage ial
skills, ope a ions and ne wo k connec ions
( o example, pa ne s in new ma ke s).
A Look a Case S udy
AI S a -Up P o iles
The assessmen o case s udy coun y s a -ups
p oceeds in wo pa s. The i s sec ion — he
33 I should be no ed ha he e is a lo o qua e - o-qua e ola ili y in VC
ma ke s. These wo ecen qua e s a e p esen ed he e o illus a i e
pu poses and no as an indica ion o la ge ends.
Table 4: AI-Rela ed VC Funding and Deals, By Region, Q3 and Q4 2024
Q3 2024 Q4 2024
Deals (Coun s) Funding (US$
millions)
Sha es o To al
Funding (%)
Deals (Coun s) Funding
(US$ millions)
Sha es o To al
Funding (%)
Uni ed S a es 566 11,442 68.0 548 38,028 86.9
Eu ope 279 2,779 16.5 291 2,519 5.8
Asia 316 2,112 12.5 270 2,045 4.7
La in Ame ica 29 62 0.4 17 92 0.2
Canada 27 184 1.1 24 922 2.1
A ica 7 9 0.1 1 0 0.0
All o he egions
(including Oceania)
21 245 1.5 14 160 0.4
To als 1,245 16,832 100.0 1,165 43,766 100.0
Sou ce: www.cbinsigh s.com/ epo s/CB-Insigh s_AI-Repo -2024.xlsx; au ho ’s abula ions..
25De eloping Coun ies’ Business Pa icipa ion in he AI Economy
s udy coun ies. Al oge he , hese local in es o s
con ibu ed 26 in es o engagemen s in o he case
s udy coun ies. While he numbe s a e modes , i
is in e es ing o see he eme gence o some Sou h-
Sou h in es men lows in suppo o AI s a -ups.
The e a e he beginnings o egional ecosys ems
ope a ing, especially among he i e A ican case
s udy coun ies. The e a e also A ican ies o La in
Ame ica wi h some in es o engagemen in bo h
di ec ions. This ype o ac i i y is lacking among
he h ee Sou heas Asian case s udy coun ies.
To ge a be e handle on he o e all lows, we
now u n o examine he op sending coun ies
o in es men engagemen s ecei ed by each o
he case s udy coun ies (see Table 9).38 The Uni ed
S a es domina es he ankings o lead in es o
in se en coun ies. Bu h ee o he case s udy
38 Appendix 2 p esen s de ailed p o iles o he op G7 in es o s (Table A.1);
he op domes ic in es o in each case s udy coun y, excep Thailand
whose co e ed i ms only had o eign in es o s (Table A.2); and he op
in es o s ac oss all he sample coun ies (Table A.3).
coun ies — B azil, Indonesia and Tunisia — a e
hei own lead in es o , wi h he Uni ed S a es
placing second. The s ong local pe o mance in
hese coun ies may p o ide an indica ion o a
dynamic local ecosys em eme ging o AI s a -ups.
By a wide ma gin, he s ong Uni ed S a es- o-
B azil channel accoun s o he la ges single
in e na ional low o in es o engagemen s
shown in he able. US in es o s also showed
ela i ely s ong in e es in Nige ia and
Sou h A ica. Al oge he , i e G7 coun ies a e
ep esen ed ac oss he able. Japan, in pa icula ,
is ep esen ed on in es men lows o all h ee
case s udy coun ies in Sou heas Asia, indica ing
a po en ially impo an egional co ido . (Some
h ee qua e s o Indonesia’s in es o engagemen
o igina e in Asia.) The e a e u he indica ions
o Sou h-Sou h egional hub lows ia in es o
engagemen s such as Saudi A abian engagemen
in Egyp , Mau i ian engagemen in Sou h A ica
and Singapo ean engagemen in Thailand.
Table 8: Domes ic In es o s Based in he Case S udy Coun ies and In es ing in G7 A ilia ed AI
S a -Ups a Home o Ab oad in O he Case S udy Coun ies, 1Q2025
Home
Coun y
Case S udy Coun y
Domes ic In es o
En i ies (Coun )
Case S udy
Domes ic
In es o s In es ing
Domes ically
(Coun )
Domes ic
In es men
Engagemen s
(Coun )
Ou bound Fo eign
In es men
Engagemen s by
Domes ic Fi ms
o O he S udy
Coun ies (Coun )
Des ina ion Coun ies
B azil 132 130 201 3 Kenya (1), Sou h A ica
(1), Tunisia (1)
Colombia 3 3 3 0
Egyp 12 12 17 0
Indonesia 19 19 24 0
Kenya 7 4 4 4 Egyp (1), Sou h A ica
(1), Nige ia (2)
Nige ia 15 13 16 5 B azil (1), Colombia (2),
Egyp (1), Kenya (1)
Sou h A ica 24 21 23 8 Kenya (6), Nige ia (2)
Thailand 0 0 0 0
Tunisia 6 4 5 6 Egyp (5), Sou h A ica (1)
Vie nam 4 4 5 0
To al 222 210 298 26
Sou ce: C unchbase da abase; au ho ’s abula ions and da abase cons uc ion; Claude.ai (assis ed in he compila ion o
his able).
No e: “G7 a ilia ed” means he e is a leas one in es o om a G7 coun y in es ing in he Al s a -up. The in es o
da abase assembled by he au ho o co e en i ies in es ing in Al s a -ups in he 10 case s udy coun ies includes 526
unique in es o s. The e we e 210 domes ic in es o s in es ing domes ically, and he e we e 316 domes ic and o eign
in es o s in es ing in he 10 case s udy coun ies ac oss in e na ional bo de s.

26 CIGI Pape No. 323 — June 2025 • Douglas Lippold
In es men and Non-Financial Assis ance
The in es o p o ile da a e eals ha in es o
engagemen s o en in ol e assis ance beyond
pu ely inancial in es men . Many o hese
in es o s ha e gained expe ience in e na ionally,
which hey a e b inging o bea in suppo o
hei chosen AI s a -ups. While some in es o s
men ion p o ision o specialis suppo o speci ic
issues such as go e nance, many o e ings all
in o wo b oad clus e s o issues: echnology
ans e and ecosys em de elopmen .
Technology ans e akes a a ie y o o ms. When
la ge echnology companies such as Google o
Mic oso in es , hey may p o ide access o hei
echnology s acks and pla o ms.39 Only abou h ee
pe cen o he in es o s in ou sample explici ly
men ion echnology o skills ans e as a ocus o
hei se ice. Mos o en his akes place h ough
knowledge sha ing, men o ship, and ne wo k
39 As no ed abo e, AI s a -ups co e ed in ou s udy ha e gene ally no
decla ed use o pa en p o ec ion. In compa ison, he in e na ional
co po a e in es o s a e mo e likely o employ o mal o ms o in ellec ual
p ope y o p o ec hei own in e es s. This means ha he p o ision o
adequa e le els o in ellec ual p ope y p o ec ion may be a use ul pa
o he hos coun y policy mix in o de o acili a e access o domes ic AI
s a -ups o echnological inpu s om ab oad, along wi h he know-how
on deploymen (Pa k and Lippold 2005, 2014).
connec ions a he han o mal echnology ans e
p og ams. Fo he G7 in es o s, in pa icula ,
managemen expe ise ans e is o en a majo
componen o hei in es men p ocess. Among
o he elemen s, his may include s a egic guidance
and me hodological suppo o ope a ional
scaling in ela ion o echnological upg ades.
Ecosys em de elopmen is ano he
cha ac e iza ion ha some in es o s gi e o hei
suppo . This means building mu ually suppo i e
ela ionships wi h he a ious coun e pa s ha
an AI s a -up migh need o engage in success ully
pu suing i s ope a ions. Ecosys em de elopmen
is an o e ing explici ly men ioned by 24 in es o s.
O he dimensions o his ype o assis ance include
ne wo k access (ci ed by 43 in es o s, his may
include connec ions o po en ial cus ome s,
pa ne s and ollow-on in es o s), accele a ion
se ices (o e ed by 10 in es o s, which may include
op imizing he implemen a ion o AI echnology),
and incuba ion p og ams (o e ing sui es o se ices
using a holis ic app oach, as ci ed by i e in es o s).
The echnology ans e and ecosys em
de elopmen suppo ci ed by hese in es o s
in ou case s udy i ms aligns wi h he li e a u e
ci ed abo e on i m-le el AI adop ion in
de eloping coun ies. As Mu asa e al. (2024)
Table 9: Leading In es men Co ido s (Coun s o In es o Engagemen s in G7 A ilia ed AI S a -
Ups)
Coun y Top In es o 2nd In es o 3 d In es o Top 3 Sha e (%)
B azil B azil (200) Uni ed S a es (153) Uni ed Kingdom (8) 90.5
Colombia Uni ed S a es (10) Colombia (3) Spain (2) 75.0
Egyp Uni ed S a es (14) Egyp (13) Saudi A abia (6) 60.0
Indonesia Indonesia (25) Uni ed S a es (12) Japan (9) 78.0
Kenya Uni ed S a es (17) Sou h A ica (6) Kenya (4) 58.7
Nige ia Uni ed S a es (31) Nige ia (16) Canada (4) 76.1
Sou h A ica Uni ed S a es (27) Sou h A ica (22) Mau i ius (4) 74.6
Thailand Uni ed S a es (2) Singapo e (1) Japan (1) 100.0
Tunisia Tunisia (5) Uni ed S a es (4) F ance (2) 64.7
Vie nam Uni ed S a es (12) Japan (7) Vie nam (5) 75.0
Sou ces: C unchbase, online da abase, www.c unchbase.com/disco e /o ganiza ion.companies (Q1 2025); au ho ’s
abula ions and in es o da abase cons uc ion; Claude.ai p o ided inal able compila ion assis ance.
No e: Numbe s in pa en hesis e e o in es o engagemen s. “G7 a ilia ed” means he e is a leas one in es o om a G7
coun y in es ing in he Al s a -up.
27De eloping Coun ies’ Business Pa icipa ion in he AI Economy
highligh , e ec i e AI implemen a ion equi es
no jus unding bu also knowledge sha ing,
s a egic guidance and ne wo k connec ions,
among o he elemen s. Ce ain o he in es o
o e ings co espond o inpu s ha Fan and Qiang
(2024) ci e as necessa y o p omo ing posi i e
ou comes in he AI economy, including local
ecosys em de elopmen and skills de elopmen .
In addi ion, he eme ging Sou h-Sou h in es men
obse ed in ou case s udy coun ies also lends
suppo o Maye ’s (2021) obse a ion ha i ms
wi h local knowledge can de elop a measu e
o compe i i e ad an age — and, po en ially,
imp o ed esilience — by ailo ing AI solu ions
o ma ke condi ions in de eloping coun ies. In
sum, his complemen a y ela ionship be ween
in es o s’ inancial and non- inancial inpu s
add esses a a ie y o needs a AI s a -ups and
po en ially mo e gene ally in hei home coun ies.
Geog aphy Ma e s
Economis Paul K ugman and o he s ha e no ed
agglome a ion e ec s, whe eby de elopmen
in a c owded echnology cen e may o e i ms
some ad an ages in e ms o de elopmen and
di usion o inno a i e p oduc s (K ugman 1995).
This likely also ope a es in he AI economy, as well.
Fo example, hick labou ma ke s may de elop
and o e la ge pools o sec o - ele an alen .
Imp o ed communica ion a ound inno a ion
could eme ge due o he p oximi y o s akeholde s,
he eby con e ing u he in o ma ion ad an ages.
The geog aphic concen a ion may be suppo ed
by he a ailabili y o VC unding in some o hese
a eas. The combina ion o ample in es men
unding and ad iso y suppo may p o ide a
d aw o speci ic geog aphies. O he d aws migh
include he a ailabili y o access o complemen a y
academic esea ch. Wi h espec o he case s udy
sample o G7 in es o -suppo ed AI s a -ups (see
Table 10: Geog aphic Concen a ion o AI S a -Ups
Coun y To al S a -Ups wi h G7 In es men Ci y Numbe o S a -Ups, by Ci y Ma ke Sha es (%)
B azil 119
São Paulo 75 63.0
Cu i iba 9 7.6
Rio De Janei o 4 3.4
O he ci ies 31 26.1
Colombia 9
Medellín 5 55.6
Bogo á 4 44.4
Egyp 15
Cai o 9 60.0
Alexand ia 3 20.0
Gîza 2 13.3
Zamalek 1 6.7
Indonesia 18
Jaka a 9 50.0
Jaka a Pusa 2 11.1
O he ci ies 7 38.9
Kenya 14
Nai obi 14 100.0
28 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Table 10), one can obse e some clus e ing. In
eigh o he 10coun ies, one ci y accoun ed o a
majo i y o he ac i i y. This was mos p onounced
in São Paulo, B azil, wi h 75 AI s a -ups as o
1Q2025. In Sou h A ica and Vie nam, he e was
a mo e balanced spli be ween wo economic
cen es wi h each accoun ing o subs an ial
sha es o he AI s a -ups (Johannesbu g/
Cape Town and Hanoi/Ho Chi Minh Ci y).
AI S a -Ups Tha Closed
The ocus in his assessmen is on con inuing AI
s a -ups ha a e pa icipa ing in he AI economy in
ou sample o middle-income de eloping coun ies.
Bu i is wo hwhile o pause and conside he
AI s a -ups no included in ou sample because
hey ha e closed (see Table 11). O he 133 closed AI
s a -ups in he 10 case s udy coun ies, C unchbase
epo s ha only 11 had a G7 in es o , while ano he
26 a e lis ed as ha ing exclusi ely domes ic o
hi d coun y in es o s. Fo he o he 96 closed
AI s a -ups, i appea s ha boo s apping was a
p ima y means o unding he i m ( hough we
canno ule ou a gap in he epo ing on in es o
pa icipa ion). O he 110 closed i ms ha epo ed
employmen , h ee qua e s had employmen o
10 o less. Only 30 o he 133 closed i ms p o ided
epo ing on closing o exi da es. Among he
closed i ms, “exi ” is likely he p e e ed way ou ,
meaning he i m was acqui ed o ans o med ia
an ini ial public o e ing (IPO). This ansla es in o
a known “success a e” among he closed i ms
o abou 10 pe cen ( ha is, 13 i ms epo ed an
exi , hough he e may be addi ional i ms in his
ca ego y no epo ed). In addi ion o AI, many o
he “success ul” closed i ms a e lis ed as ha ing
an indus y ocus in one o mo e a eas such as
con en c ea ion, digi al ma ke ing, in e ne ,
business in elligence o ma ke ing au oma ion.
In e es ingly, none o he i ms wi h G7 in es o s
epo ed ha ing an exi , meaning ha hei manne
o closu e canno be assessed. Among all he i ms
epo ing closu e da es (excluding exi s), he age
ended o be ela i ely sho (in o he wo ds,
be ween one and h ee yea s). Sou h A ica was an
excep ion, due o a couple o ou lie obse a ions.
Coun y To al S a -Ups wi h G7 In es men Ci y Numbe o S a -Ups, by Ci y Ma ke Sha es (%)
Nige ia 22
Lagos 14 63.6
Yaba 2 9.1
Abuja 2 9.1
O he ci ies 4 18.2
Sou h A ica 21
Cape Town 9 42.9
Johannesbu g 8 38.1
O he ci ies 4 19.0
Thailand 2
Bangkok 2 100.0
Tunisia 7
Tunis 6 85.7
S ax 1 14.3
Vie nam 12
Hanoi 6 50
Ho Chi Minh Ci y 6 50
To al 239
Sou ce: www.c unchbase.com/disco e /o ganiza ion.companies; au ho ’s abula ions and da abase cons uc ion;
Claude.ai ( able compila ion assis ance).
Table 10: Geog aphic Concen a ion o AI S a -Ups (Con inued)
29De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table 11: Assessmen o AI S a -Ups ha Closed Down, Coun s (Excep “Yea s” Whe e Age Is Repo ed)
Home
Coun y
AI S a -
Ups Tha
Ha e
Closed
Fi ms wi h
Ex e nal
In es o s
Lis ed
Fi ms wi h
Lis ed G7
In es o s
Employmen
(1–10)
Employmen
(11–50)
Employmen
(51–100)
Employmen
(101–250)
Employmen
(251–500)
Employmen
( i ms ailing
o epo )
Fi ms Repo ing
Closing Da e
(excluding
hose exi ing)
A e age
Age A
Closu e,
Yea s
Exi
Repo ed:
Acqui ed
o IPO
A e age Age
A Exi , Yea s
Closed Fi ms
Failing o
Repo Closu e
o Exi Da e
B azil 80 26 6 49 7 3 3 1 17 92.9 11 5.9 60
Colombia 5 1 1 2 2 1 0 0 0 2 1.7 0 --- 3
Egyp 5 2 1 4 1 0 0 0 0 1 1.3 0 --- 4
Indonesia 4 1 1 2 2 0 0 0 0 2 1.2 1 1.9 1
Kenya 8 1 1 7 1 0 0 0 0 0 --- 0 --- 8
Nige ia 13 2 0 11 1 0 0 0 1 1 1.2 0 --- 12
Sou h
A ica
7 3 1 4 1 1 0 0 1 2 7.5 15.0 4
Thailand 1 0 0 1 0 0 0 0 0 0 -- 0 --- 1
Tunisia 3 0 0 3 0 0 0 0 0 0 -- 0 --- 3
Vie nam 7 1 0 1 1 0 0 1 4 0 -- 0 --- 7
To als 133 37 11 84 16 5 3 2 23 17 13 103
Sou ces: C unchbase, online da abase, www.c unchbase.com/disco e /o ganiza ion.companies ( iewed a end o Q1 2025); au ho ’s abula ions
and da abase cons uc ion.
No e: IPO = ini ial public o e ing.
30 CIGI Pape No. 323 — June 2025 • Douglas Lippold
In B azil ( he only coun y wi h mo e han a
couple o obse a ions), he age o success ul AI
s a -ups ha exi ed ended o be olde han o
he a e age AI s a -up ha closed. I may be ha
en ep eneu s and in es o s he e a e ela i ely
quick o pull he plug on a ailing i m, e en as
hey ecognize ha i akes longe o p epa e an
exi o a success ul s a -up. Un o una ely, due
o gaps in he da a se , we a e no able o con i m
his o he o e all popula ion o closed s a -ups.
AI Unico ns
To u he assess he AI s a -up-in es o uni e se
in ou case s udy coun ies, we now u n o look
o unico ns. Such i ms a e in e es ing e e ence
cases o o he s a -ups in ha hey ha e
demons a ed possible pa hways o success. We
d aw on he da abase o unico ns de eloped by CB
Insigh s.40 This da abase p o ides in o ma ion on
he companies, hei alua ions, da e o ecogni ion
as a unico n, coun y and ci y o headqua e s,
applicable indus y, and selec ed in es o s. As o
Janua y 7, 2025, he os e included 1,257 i ms.41
The cu en os e o unico ns includes 37 i ms
based in ou case s udy coun ies.42 Only one o
hese unico ns is also in ou C unchbase da abase
o AI s a -ups. (Tha is he B azilian s a -up
Cloudwalk, an AI-powe ed inancial se ices
p o ide wi h a alua ion o US$2.15 billion.) Bu
he e a e o he AI i ms and in es o s o in e es
in he unico n os e . In o de o iden i y he op
AI-in ensi e i ms on he lis , ou echnique is o
ap in o indus y expe opinion conce ning which
AI i ms excel and hen c oss e e ence his agains
he os e o unico ns.43 The indus y publica ion
eWeek eleases such a lis o leading AI i ms, wi h
he la es edi ion published on Oc obe 1, 2024
and co e ing 150 i ms (Hi e 2024). On his basis,
16 AI-in ensi e i ms can be iden i ied (Table 12).
O hese, 13 a e loca ed in he Uni ed S a es, wo
in he Uni ed Kingdom and one in Canada.
40 See www.cbinsigh s.com/ esea ch-unico n-companies. Also no e
Appendix 1.
41 See www.cbinsigh s.com/ esea ch-unico n-companies
42 The CB Insigh s os e o unico ns includes 37 i ms om eigh o he 10
case s udy coun ies: B azil (18), Colombia (3), Egyp (1), Indonesia (7),
Nige ia (2), Sou h A ica (1), Thailand (3) and Vie nam (2). The e a e no
Kenya- o Tunisia-based i ms on he os e .
43 See Lippold (2024, 10). Ano he op ion is o c oss e e ence he unico n
os e agains ele an indus y associa ion membe ship lis s. Bu such
an app oach may p o e less han imely, as de elopmen s a e e ol ing
apidly.
This unico n analysis is o in e es o ou cu en
case s udy o AI s a -ups o a a ie y o easons.
Fi s o all, he ac ha mos o ou case s udy
coun ies a e ep esen ed on he os e o unico ns
indica es ha i is possible unde local condi ions
o g ow a s a -up o unico n scale. Secondly, gi en
he numbe o ea ly s age AI s a -ups ha we ha e
iden i ied ac oss he 10case s udy coun ies, i may
p o e ele an o keep in mind ha AI s a -ups
can become unico ns and end o do so ea lie
in hei li e cycle han non-AI i ms. Acco ding
o CB Insigh s,44 among unico ns eme ging in
2024, nea ly wo hi ds o AI unico ns a ained
he s a us p io o scaling up and becoming ully
es ablished. Mos o hese AI unico ns we e s ill
in he phases o alida ing hei business models
wi h he ini ial deploymen o esou ces. Fo mo e
han ou in i e non-AI unico ns, hei s a us was
a ained in la e phases o comme cial ma u i y,
as hey scaled and es ablished hemsel es.
Gi en high ma ke expec a ions o AI echnology
s a -up businesses, he e may be le e age
ha can be exploi ed in seeking in es o s.45
Thi dly, he e is a s a -up connec ion o he
success ul AI unico ns ia hei pool o in es o s.
Some o hese in es men i ms ha e accumula ed
signi ican expe ience ia hei collabo a ion
wi h success ul AI s a -ups o e a pe iod o
yea s. As no ed abo e, a success ul in e na ional
engagemen wi h such an in es o can p o ide
a channel o echnology ans e , manage ial
skill de elopmen and ope a ional bene i s such
as access o ne wo ks o ap in o subjec a ea
expe ise, ma ke knowledge and supplie s.
I u ns ou ha a numbe o AI unico n in es o s
ha e di ec ies o he AI s a -ups in he case
s udy coun ies. So his channel o G7 in es o
engagemen is al eady ope a ing o some ex en :
→ And eessen Ho owi z: In es s in AI unico ns
Da ab icks, Andu il and Shield AI. The i m also
ope a es in B azil wi h in es men s in wo AI
s a -ups.
→ Google and Google Ven u es: In es in AI
unico ns An h opic and Syn hesia. Google,
Google o S a ups and Google.o g in es
in B azil (26 s a -ups), Egyp (one s a -up),
44 See www.cbinsigh s.com/ epo s/CB-Insigh s_AI-Repo -2024.xlsx.
45 CB Insigh s uses a “Comme cial Ma u i y” sco ing o s a ups wi h
he ollowing phases: (1) Eme ging; (2) Valida ing; (3) Deploying; (4)
Scaling; (5) Es ablished.

31De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table 12: Leading AI Unico ns (s a -ups alued a US$1bn o mo e), 1Q2025
Company Valua ion
(US$ Billions)
Da e
Designa ed
as Unico n
Coun y Ci y Indus y Selec In es o s
OpenAI $157.00 2019-07-22 Uni ed S a es San F ancisco En e p ise Tech Khosla Ven u es, Th i e
Capi al, Sequoia Capi al
Da ab icks $62.00 2019-02-05 Uni ed S a es San F ancisco En e p ise Tech And eessen Ho owi z,
New En e p ise Associa es,
Ba e y Ven u es
An h opic $16.05 2023-02-03 Uni ed S a es San F ancisco En e p ise Tech Google
Andu il $14.00 2019-09-11 Uni ed S a es I ine Indus ials And eessen Ho owi z,
Founde s Fund,
Re olu ion Ven u es
Glean $4.60 2022-05-18 Uni ed S a es Palo Al o En e p ise Tech Gene al Ca alys , Kleine
Pe kins Cau ield & Bye s,
Ligh speed Ven u e Pa ne s
Hugging
Face
$4.50 2022-05-09 Uni ed S a es New Yo k En e p ise Tech Be awo ks Ven u es,
Addi ion, Lux Capi al
In lec ion
AI
$4.00 2022-05-13 Uni ed S a es Palo Al o En e p ise Tech Ga es F on ie , G eylock
Pa ne s, Ho izons Ven u es
Da aiku $3.70 2019-12-04 Uni ed S a es New Yo k En e p ise Tech Al en Capi al, Fi s Ma k
Capi al, capi alG
Shield AI $2.80 2021-08-24 Uni ed S a es San Diego Indus ials And eessen Ho owi z,
Homeb ew, Poin 72 Ven u es
Mo ewo ks $2.10 2021-06-30 Uni ed S a es Moun ain View En e p ise Tech Ligh speed Ven u e Pa ne s,
Sapphi e Ven u es, Kleine
Pe kins Cau ield & Bye s
Syn hesia $2.10 2023-06-13 Uni ed
Kingdom
London En e p ise Tech Google Ven u es, Kleine
Pe kins Cau ield & Bye s,
Fi s Ma k Capi al
Cohe e $2.00 2023-05-02 Canada To on o En e p ise Tech Index Ven u es, Sales o ce
Ven u es, Sec ion 32
Jaspe $1.50 2022-10-17 Uni ed S a es Aus in En e p ise Tech Founda ion Capi al,
Ins i u ional Ven u e
Pa ne s, Founde s Capi al
Runway $1.50 2023-05-04 Uni ed S a es New Yo k Media &
En e ainmen
Lux Capi al, Compound,
Ampli y Pa ne s
Adep $1.00 2023-03-14 Uni ed S a es San F ancisco En e p ise Tech G eylock Pa ne s,
Addi ion, M12
S abili y AI $1.00 2022-10-05 Uni ed
Kingdom
London En e p ise Tech Ligh speed Ven u e Pa ne s,
Coa ue Managemen
Memo i em
CloudWalk $2.15 2021-09-08 B azil Sao Paulo Financial Se ices Plug & Play Ven u es, Valo
Capi al G oup, DST Global
(also Coa ue, BTG Pac ual)
Sou ce: CB Insigh s (2025a); Hi e (2024); au ho ’s abula ions; Claude.ai o compila ion. Addi ional in es o s o CloudWalk (in
b acke s) we e iden i ied om he C unchbase da abase.
No e: CloudWalk sel iden i ies as “ he in e plane a y paymen ne wo k.”
32 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Indonesia (one s a up); Kenya (5 s a -ups);
Nige ia (5 s a -ups); and Sou h A ica (one s a -
up).
→ Sequoia Capi al: In es s in OpenAI. I is also
ope a ing in B azil (one s a -up).
→ Khosla Ven u es: In es s in OpenAI. I is also
in es ing in Nige ia (one s a -up).
→ Sales o ce/Sales o ce Ven u es: In es s in Cohe e.
The i m is also ope a ing in Tunisia (one s a -
up).
→ Coa ue: In es s in S abili y AI. I is also in es ing
in B azil (CloudWalk).
The B azilian s a -up ecosys em appea s o be
ai ly well in eg a ed wi h ies o ou o hese
in es o g oups. Nige ia has connec ions o wo o
he in es o s and Kenya has had success wi h he
Google g oup. Bu connec ions o he o he s a e
mo e limi ed. Colombia, Thailand and Vie nam a e
no ye engaged ia his channel, so his emains a
possible oppo uni y o hese coun ies o explo e.
Policy Assessmen
In ligh o he o egoing assessmen o s a -ups
wi h G7 in es men , we now u n o he policy
en i onmen in which hey ope a e. We conduc ed
a e iew o AI policy amewo ks in each o he
case s udy coun ies. Da a om he OECD.AI
Policy Obse a o y was employed as a baseline
o he e iew. We supplemen ed his e iew
by d awing on da a om he ac i i y acke a
he Digi al Policy Ale (DPA) o ganiza ion, as
well as in o ma ion om se e al DPA analy ical
epo s.46 O he upda es we e d awn om pee -
e iewed academic and egional sou ces.47
The 10 case s udy coun ies all ha e launched
de elopmen o AI policy amewo ks o ad ance
hei engagemen in he AI economy (see Box
5). They a y in hei p og ess (de ails a e gi en
46 Fo u he in o ma ion on he DPA as a da a sou ce, see Appendix 1.
47 The ull lis o sou ces can be ound a he bo om o Table A.4 in
Appendix 2. No e: A summa y able iew o he OECD.AI da a as o
2021 can be ound in Lippold (2024, able 5, 21–23). The able co e s
he ad anced economies as well as he p esen case s udy coun ies
B azil, Indonesia, Sou h A ica and Vie nam.
in Table A.4 in Appendix 2). I is in e es ing o
see ha Colombia and B azil — wi h he mos
de eloped AI policy amewo ks — also ha e
he la ges numbe s o AI s a -ups among ou
case s udy coun ies (see Table 5). Bo h coun ies
ha e aken s eps owa d he de elopmen
o ad anced egula o y mechanisms such as
sandbox app oaches and isk-based legisla ion.48
Sou h A ica, Thailand, Tunisia and Vie nam
occupy an in e media e posi ion in AI policy
de elopmen wi h de ined na ional s a egies,
designa ed p io i ies and eme ging egula o y
amewo ks. Howe e , hese ou coun ies each
ace implemen a ion challenges. Kenya and
Nige ia show p omising momen um wi h ecen
legisla i e de elopmen s, despi e in as uc u e
limi a ions. Egyp and Indonesia ep esen ea lie
s ages o policy o ma ion, ocusing p ima ily on
capaci y building and in as uc u e de elopmen
while s ill es ablishing basic egula o y
amewo ks. Ac oss all o he case s udy coun ies,
he e is a common ension be ween os e ing
inno a ion and ensu ing e hical go e nance.
Coun ies a e esponding o his ension wi h
engagemen in in e na ional coope a ion and
a ious adap a ions o he di e gen egula o y
app oaches o hei in e na ional pa ne s.
Some a e leaning owa d Wes e n-aligned
amewo ks (in o he wo ds, he EU model)
and o he s o al e na i e models in luenced by
pa ne ships wi h coun ies such as China.
Policy Re iew Findings
D awing on he summa ies abo e and
he de ailed policy da a in Table A.4 in
Appendix 2, se e al hemes eme ge.
Fi s , he 10 case s udy coun ies ha e all begun
o ackle AI policy and egula ion using s a egic
app oaches. Each o he sample coun ies has
ecognized ha he scope o he challenge and
oppo uni y o AI equi es a na ional-le el esponse.
Acco dingly, hey ha e launched ini ia i es o ake
s ock and de elop na ional s a egies o plans
and ac ion agendas. P og ess in hese ini ia i es is
mixed, as can be seen in Table A.4 in Appendix 2.
Ye i is no able ha he sample coun ies a e each
48 Wi h espec o AI s a -ups, i may be ha he consul a i e app oaches
being employed in hese wo na ions a e p o iding s akeholde s wi h
g ea e ce ain y and p edic abili y in he comme cial en i onmen . This is
alued by en ep eneu s and may help o p omo e s a -up de elopmen .
Fo example, see he discussion on ce ain y and p edic abili y in
egula o y p ocesses in Ge adin (2017, sec ion C).
33De eloping Coun ies’ Business Pa icipa ion in he AI Economy
s i ing o capi alize on AI in a s a egic manne ,
aking in o accoun hei na ional condi ions.
B azil and Colombia a e u hes in de elopmen
o hei AI amewo ks and ha e achie ed some
scale in de elopmen o hei AI ecosys ems,
which hey a e bo h seeking o le e age as
eme ging egional playe s. Egyp , Kenya and
Tunisia ha e sough o posi ion hemsel es as
aspi ing egional AI hubs as well, albei on a
smalle scale. Se e al o he s — such as Indonesia,
Sou h A ica and Thailand — e e o de elopmen
o AI as pa o s a egic engagemen in he so-
called Fou h Indus ial Re olu ion, which e e s
o posi i e exploi a ion o echnologies ha can
le e age inpu s om he physical, digi al and
biological sphe es ( o example, wi h espec
o au oma ion o cybe -physical sys ems).
Second, mos o he coun ies a e acing
implemen a ion challenges in ealizing hei
na ional plans. This is, in pa , associa ed
wi h de eloping adequa e capaci y in public
adminis a ion and among o he s akeholde s
Box 5. Case S udy Coun ies’ AI Policy F amewo k S a us in a Nu shell, 1Q2025
B azil: Fai ly comp ehensi e amewo k o obus egula o y de elopmen . The sys em is o ien ed
owa ds EU- ype app oaches o AI go e nance including a ecen ly passed Sena e AI bill ocused on
isk managemen (lowe house ac ion is pending) and a p oposed Na ional Cen e o Algo i hmic
T anspa ency and T us wo hy AI.
Colombia: Ex ensi e policy ini ia i es, including some wi h egula o y sandboxes h ough he
Supe in endencia de Indus ia y Come cio’s inno a ion lab; balancing inno a ion wi h esponsible
use h ough a comp ehensi e pilla -based s a egy.
Egyp : S a egy ocused on becoming a egional AI hub h ough capaci y building and esea ch,
wi h g owing Chinese pa ne ships po en ially c ea ing Wes e n alignmen challenges.
Indonesia: Ea ly-s age amewo k p io i izing in as uc u e de elopmen ac oss i e p io i y
a eas, wi h ecen ly implemen ed da a p o ec ion legisla ion and de eloping e hics guidelines.
Kenya: Eme ging s a egy emphasizing key economic sec o s wi h s ong mobile in as uc u e
and ac i e in e na ional go e nance engagemen bu acing esou ce cons ain s. Regula o y
sys em is s ill in ea ly s ages.
Nige ia: E ol ing amewo k wi h emphasis on e hics and inclusi i y; signi ican legisla i e
p og ess wi h new egula o y commission app o al, despi e pe sis en in as uc u e challenges.
Sou h A ica: Balanced app oach ocusing on inclusi e g ow h and alen de elopmen , wi h he
mos de eloped en u e capi al ecosys em in A ica, bu hampe ed by ch onic elec ici y supply
challenges.
Thailand: Well-de eloped isk-based s a egy wi h de ailed a ge s and implemen a ion plan;
ac i ely pa icipa ing in ASEAN go e nance amewo ks while de eloping sec o -speci ic
applica ions.
Tunisia: S a egy ocused on le e aging enginee ing educa ion o become a egional AI hub, wi h
an EU-aligned da a p o ec ion amewo k, bu acing a challenge om implemen a ion gaps and
limi ed en o cemen .
Vie nam: Fo wa d-looking app oach wi h egula o y inno a ion including machine- eadable AI
labels and sandboxes; de eloping subs an ial specialis wo k o ce amid g owing pa ne ships wi h
Asian echnology leade s.
Sou ces: See Table A.4.
34 CIGI Pape No. 323 — June 2025 • Douglas Lippold
ha need o be engaged. Some ha e sough o
mee his challenge ia public-p i a e pa ne ship
a angemen s.49 In de eloping he egula o y
egimes associa ed wi h implemen a ion o
he na ional plans, some o he case s udy
coun ies ha e also unde sco ed he need o
anspa ency and consul a ion, which can ha e
posi i e e ec s on ou comes. Fo example, B azil,
Colombia, Kenya, Sou h A ica and Thailand all
ha e made explici e e ences o hese issues
wi h espec o speci ic aspec s o egula ion.
Thi d, all 10 o he coun ies a e s ill s uggling
wi h he u ban- u al digi al di ide, a key elemen
o inclusi eness in he AI economy. Closing
he gap will equi e — among o he elemen s
— imp o ed pe o mance in educa ion and
in as uc u e de elopmen ( o example, elec ici y,
elephony, in o ma ion echnology capaci y).
Fou h, nea ly all o he case s udy coun ies
ha e es ablished a ge s o AI human capi al
de elopmen .50 Sho alls a e p o ing o be a
cons ain on AI de elopmen , no leas because
o in e na ional poaching o alen . B azil,
Egyp and Tunisia explici ly ci e alen ligh
as a s a egic isk. (And, indeed, Maslej e al.
(2024, 239) epo ha coun ies such as B azil
and Sou h A ica ha e expe ienced a s eady
ou low o alen o compe i o ma ke s.)
Fi h, he e is a ension o e egula o y
con e gence and di e gence, wi h implica ions
o ma ke access in bo h di ec ions o he case
s udy coun ies. Selling and sou cing a e bo h
a ec ed by egula o y compliance issues. The e
a e isks due o agmen a ion. B azil, Colombia,
Sou h A ica and Tunisia ha e aken s eps owa d
EU s yle app oaches o egula ion, while Egyp ,
Kenya, and Vie nam ha e begun collabo a ion
wi h some Chinese AI ini ia i es ( o example, ia
he Bel and Road Ini ia i e (BRI), which includes
a digi al dimension). The Uni ed Kingdom as well
has an in luence, in pa icula on AI sa e y issues.
Mo eo e , he US ma ke is a key sou ce o AI
echnology and en u e capi al o all 10 coun ies
as demons a ed in ou s a -up analysis. Ea ly
49 As no ed in he li e a u e e iew, Mu asa e al. (2024) poin o public-
p i a e pa ne ships as one means o s akeholde s o mi iga e high-cos
bu dens ( o example, wi h espec o in as uc u e p ojec s).
50 I is no able ha ac ion in his a ea is also iden i ied by Fan and Qiang
(2024), who emphasize p io i izing skills de elopmen as one o i e
p io i ies in he Wo ld Bank amewo k o posi i e AI de elopmen (see
he li e a u e e iew abo e o a discussion o his e e ence).
indica ions a e ha he US is now shi ing o a
much ligh e egula o y egime (Ca illo e al.
2025; Villaseno and Tu ne 2024). While he e ms
o he new US AI egime s ill emain unse led,
he e a e isks o US policy changes u he
ac u ing global AI go e nance. Fo he 10 case
s udy coun ies, alignmen wi h any one o he
AI majo economies could ha e knock-on e ec s
in e ms o ma ke access and compliance wi h
he o he majo economies. This is pa icula ly
impo an wi h espec o access o and con ol
o da a, key ing edien s o he AI economy.
I could a ec o he a eas as well, such as ia
compliance issues wi h espec o expo s o
equipmen o c oss-bo de sales o AI se ices.
Conclusions
This case s udy conside s he si ua ion o a sample
o in e na ionally engaged AI s a -ups and he
ole o he s a e in shaping he economic con ex
o i m-le el de elopmen . The esea ch has
iden i ied some 2,537 AI s a -ups in he 10 case
s udy coun ies ope a ing ac oss a b oad ange
o sec o s. D awing on a combina ion o global
and domes ic inpu s and inno a ion, hese i ms
a e iden i ying local and egional needs and
de eloping AI-suppo ed solu ions o comme cial
ma ke s. In doing so, abou one in 10 o hese
i ms ha e succeeded in a ac ing in es men
om en i ies based in he G7 coun y g oup. Such
in es men p o ides a channel o ans e o know-
how conce ning de elopmen and comme cial
deploymen o he echnology beyond wha migh
be a ailable domes ically. Domes ic policy is
playing a complemen a y ole o his comme cial
ac i i y. F om build ou o in as uc u e o human
capi al de elopmen , and egula ion, na ional
go e nmen s a e es ablishing he con ex and
condi ions in which he AI ecosys em is ope a ing.
As in much o he wo ld, his is a wo k in p og ess
in each o he en case s udy coun ies.
Small Scale, Bigge Impo ance
The scale o he AI s a up ac i i y in he s udy
coun ies is small compa ed o he scale o
compa able ac i i y in he leading AI na ions
such as he Uni ed S a es, China o Eu ope
and he Uni ed Kingdom. Bu i is no able in
ha en ep eneu s in he s udy coun ies ha e
41De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table A.2: The Top AI-S a -Up In es o Based in Each Case S udy Coun y, Coun s o In es men Engagemen s by Des ina ion Coun y, as o Q1
2025
In es o Name Headqua e s B azil Colombia Egyp Indonesia Kenya Nige ia Sou h
A ica
Thailand Tunisia Vie nam To al In es men S yle Suppo O e ings
Bossa In es B azil 15 0 0 0 1 0 1 0 0 0 17 Ea ly-s age seed in es o wi h
AI specializa ion; po olio spans
mul iple sec o s wi h domes ic
ocus while oppo unis ically
explo ing A ican ma ke s.
P o ides hands-on managemen
guidance and s a egic di ec ion;
o e s ope a ional suppo
o po olio companies wi h
limi ed echnical assis ance.
La in Leap Colombia 0 1 0 0 0 0 0 0 0 0 1 Ven u e Capi al S udio ocused
on so -landing pu pose-
d i en ech companies in La in
Ame ica; emphasizes ecosys em
de elopmen o e pu e e u ns.
O e s ne wo k access
and egional connec ions;
posi ions i sel as a ga eway
o in e na ional companies
en e ing he Colombian ma ke .
Fla 6Labs Egyp 0 0 2 0 0 0 0 0 0 0 2 Leading MENA egion seed
accele a o wi h s uc u ed
p og am; ocuses on suppo ing
ea ly-s age Egyp ian s a -ups
wi h egional g ow h po en ial.
Capi al- ocused in es o wi h
p og am-based suppo ; o e s
s anda dized accele a o
cu iculum a he han
cus omized assis ance.
MDI Ven u es Indonesia 0 0 0 3 0 0 0 0 0 0 3 Mul i-s age en u e und
wi h unico n ack eco d;
co po a e en u e a m o Telkom
Indonesia wi h bo h s a egic
and inancial objec i es.
P ima ily inancial in es o wi h
limi ed ope a ional suppo ;
le e ages pa en company's
egional business connec ions
o po olio companies.
Ca alys Fund Kenya 0 0 1 0 1 0 0 0 0 0 2 Impac - ocused in ech
accele a o a ge ing unde se ed
popula ions; combines
inancial inclusion mission
wi h ma ke -based app oach.
P o ides ne wo k access o
in es o s and pa ne s; emphasizes
ecosys em connec ions o e di ec
echnical o managemen suppo .
iNOVO Nige ia 1 2 1 0 0 0 0 0 0 0 4 Accele a o p og am powe ed
by UK-Nige ia Tech Hub
pa ne ship; ocuses on ea ly-
s age s a -ups ac oss mul iple
sec o s wi h egional ou look.
O e s p ima ily capi al and
p og am s uc u e; limi ed
ongoing suppo beyond
ini ial accele a ion phase.
Injini Sou h A ica 0 0 0 0 3 0 0 0 0 0 3 A ica's i s EdTech incuba o and
seed in es o ; specialized ocus
on educa ional echnology wi h
pan-A ican in es men s a egy.
Capi al ocused wi h p og am-
based suppo ; o e s
s anda dized incuba ion a he
han cus omized echnical o
managemen assis ance.
Fla 6Labs Tunisia 0 0 4 0 0 0 0 0 1 0 5 Regional seed-s age VC i m
ope a ing ac oss MENA;
s uc u ed accele a o p og ams
combined wi h seed unding
o ea ly-s age en u es.
P o ides echnical expe ise and
esou ces; o e s s anda dized s a -
up suppo p og am wi h emphasis
on echnical de elopmen .
500 S a ups
Vie nam
Vie nam 0 0 0 0 0 0 0 0 0 2 2 Tech- ocused seed in es o
a ilia ed wi h global 500
S a ups ne wo k; combines
local expe ise wi h in e na ional
in es men app oach.
O e s echnical expe ise and
global connec ions; p o ides
s anda dized accele a o suppo
wi h echnical esou ces o
ea ly-s age companies.
Sou ce: www.c unchbase.com/disco e /o ganiza ion.companies; au ho ’s abula ions and in es o da abase cons uc ion; Claude.ai ( inal able compila ion assis ance).
No e: The i m desc ip ions d aw hea ily on each i m’s sel -cha ac e iza ion and public in o ma ion. Thailand had no local op i e in es o s in sample AI s a -ups.

42 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Table A.3: The O e all Top 10 AI-S a -Up In es o s Ta ge ing he Case S udy Coun ies, Coun s o In es men Engagemen s by
Des ina ion Coun y, as o Q1 2025
In es o Name Headqua e s B azil Colombia Egyp Indonesia Kenya Sou h A ica Thailand Tunisia Vie nam Nige ia To al In es men S yle Suppo O e ings
Google o
S a ups
Uni ed S a es 26 0 0 1 4 1 0 0 0 4 36 Ea ly-s age accele a o p og am
ha c ea es global connec ions
o s a -ups; specializes in
so wa e and digi al pla o ms.
O e s men o ship, ech
esou ces, Google p oduc access
and ecosys em ne wo king
oppo uni ies bu limi ed
di ec managemen suppo .
Bossa In es B azil 15 0 0 0 1 1 0 0 0 0 17 Ea ly-s age seed in es o
wi h s ong domes ic ocus;
indus y-agnos ic bu
wi h AI specializa ion.
P o ides hands-on managemen
guidance and ac i e in ol emen
in s a egic decisions; limi ed
echnical assis ance.
Cana y B azil 12 0 0 0 0 0 0 0 0 0 12 Fi s -check La in Ame ican
seed und a ge ing inno a i e
ech companies wi h
scalable business models;
s ong domes ic ocus.
Capi al- ocused in es o wi h
limi ed ope a ional in ol emen ;
o e s ne wo k connec ions
wi hin B azilian ech ecosys em.
No e Ven u es B azil 10 0 0 0 0 0 0 0 0 0 10 Pu e- ollowe und c ea ed and
managed by en ep eneu s;
ocuses on ea ly-s age
B azilian s a -ups.
P ima ily inancial in es o
wi hou signi ican alue-added
suppo ; elies on lead in es o s
o ope a ional guidance.
Y Combina o Uni ed S a es 5 0 1 3 0 0 0 0 0 1 10 A leading global s a -up
accele a o wi h s anda dized
in es men e ms; ba ch-
o ien ed p og am wi h
s ong alumni ne wo k.
P o ides s uc u ed s a -
up cu iculum, demo day
exposu e and ounde ne wo k
access; limi ed ongoing
managemen suppo .
Techs a s Uni ed S a es 2 0 0 0 2 1 0 0 1 3 9 Global accele a o wi h egional
p og ams o ganized by indus y
e icals; highly s uc u ed
h ee-mon h p og am.
Comp ehensi e suppo
including men o ship ne wo k,
echnical esou ces and business
de elopmen assis ance
o ea ly-s age g ow h.
La i ud Ven u es B azil 8 0 0 0 0 0 0 0 0 0 8 Communi y- ocused in es men
pla o m aimed a La in
Ame ican ounde s; combines
capi al wi h in as uc u e.
O e s ounde communi y,
echnical in as uc u e and
ne wo king esou ces o
ea ly-s age companies.
Amazon Web
Se ices
Uni ed S a es 5 0 1 0 0 0 0 0 1 0 7 Co po a e in es o p io i izing
cloud adop ion and echnical
ecosys em de elopmen ;
s a egic in es men app oach.
P o ides subs an ial echnical
suppo h ough cloud c edi s,
a chi ec u e assis ance and
echnology in eg a ion expe ise.
Village Capi al Uni ed S a es 2 0 0 0 2 1 0 0 0 2 7 Impac - ocused in es o
using pee -selec ion
model; a ge s unde se ed
ma ke s and ounde s.
Combines capi al wi h business
cu iculum, pee men o ing
and in es o connec ions o
social impac en u es.
500 Global Uni ed S a es 1 0 1 1 0 0 1 0 2 0 6 Ea ly-s age global VC wi h
egional unds; high- olume
in es men s a egy ac oss
di e se geog aphies.
Limi ed ope a ional suppo
bu o e s global ne wo k
access and ollow-on unding
po en ial; egional o ices
p o ide local expe ise.
S a upboo camp Uni ed
Kingdom
1 2 2 0 0 1 0 0 0 0 6 Indus y- ocused global
accele a o ne wo k wi h
s uc u ed p og ams; ope a es
egionally wi h concen a ed
coho s in speci ic ma ke s.
Comp ehensi e accele a o
model wi h o mal
men o ship p og am,
ne wo king oppo uni ies
and indus y connec ions
o po olio companies.
Sou ce: www.c unchbase.com/disco e /o ganiza ion.companies; au ho ’s abula ions and in es o da abase cons uc ion; Claude.ai ( inal able compila ion assis ance).
No es: The i m desc ip ions d aw hea ily on each i m’s sel -cha ac e iza ion and public in o ma ion. The op 10 lis ing includes 11 companies due o a ie in he coun s..
43De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table A.3: The O e all Top 10 AI-S a -Up In es o s Ta ge ing he Case S udy Coun ies, Coun s o In es men Engagemen s by
Des ina ion Coun y, as o Q1 2025
In es o Name Headqua e s B azil Colombia Egyp Indonesia Kenya Sou h A ica Thailand Tunisia Vie nam Nige ia To al In es men S yle Suppo O e ings
Google o
S a ups
Uni ed S a es 26 0 0 1 4 1 0 0 0 4 36 Ea ly-s age accele a o p og am
ha c ea es global connec ions
o s a -ups; specializes in
so wa e and digi al pla o ms.
O e s men o ship, ech
esou ces, Google p oduc access
and ecosys em ne wo king
oppo uni ies bu limi ed
di ec managemen suppo .
Bossa In es B azil 15 0 0 0 1 1 0 0 0 0 17 Ea ly-s age seed in es o
wi h s ong domes ic ocus;
indus y-agnos ic bu
wi h AI specializa ion.
P o ides hands-on managemen
guidance and ac i e in ol emen
in s a egic decisions; limi ed
echnical assis ance.
Cana y B azil 12 0 0 0 0 0 0 0 0 0 12 Fi s -check La in Ame ican
seed und a ge ing inno a i e
ech companies wi h
scalable business models;
s ong domes ic ocus.
Capi al- ocused in es o wi h
limi ed ope a ional in ol emen ;
o e s ne wo k connec ions
wi hin B azilian ech ecosys em.
No e Ven u es B azil 10 0 0 0 0 0 0 0 0 0 10 Pu e- ollowe und c ea ed and
managed by en ep eneu s;
ocuses on ea ly-s age
B azilian s a -ups.
P ima ily inancial in es o
wi hou signi ican alue-added
suppo ; elies on lead in es o s
o ope a ional guidance.
Y Combina o Uni ed S a es 5 0 1 3 0 0 0 0 0 1 10 A leading global s a -up
accele a o wi h s anda dized
in es men e ms; ba ch-
o ien ed p og am wi h
s ong alumni ne wo k.
P o ides s uc u ed s a -
up cu iculum, demo day
exposu e and ounde ne wo k
access; limi ed ongoing
managemen suppo .
Techs a s Uni ed S a es 2 0 0 0 2 1 0 0 1 3 9 Global accele a o wi h egional
p og ams o ganized by indus y
e icals; highly s uc u ed
h ee-mon h p og am.
Comp ehensi e suppo
including men o ship ne wo k,
echnical esou ces and business
de elopmen assis ance
o ea ly-s age g ow h.
La i ud Ven u es B azil 8 0 0 0 0 0 0 0 0 0 8 Communi y- ocused in es men
pla o m aimed a La in
Ame ican ounde s; combines
capi al wi h in as uc u e.
O e s ounde communi y,
echnical in as uc u e and
ne wo king esou ces o
ea ly-s age companies.
Amazon Web
Se ices
Uni ed S a es 5 0 1 0 0 0 0 0 1 0 7 Co po a e in es o p io i izing
cloud adop ion and echnical
ecosys em de elopmen ;
s a egic in es men app oach.
P o ides subs an ial echnical
suppo h ough cloud c edi s,
a chi ec u e assis ance and
echnology in eg a ion expe ise.
Village Capi al Uni ed S a es 2 0 0 0 2 1 0 0 0 2 7 Impac - ocused in es o
using pee -selec ion
model; a ge s unde se ed
ma ke s and ounde s.
Combines capi al wi h business
cu iculum, pee men o ing
and in es o connec ions o
social impac en u es.
500 Global Uni ed S a es 1 0 1 1 0 0 1 0 2 0 6 Ea ly-s age global VC wi h
egional unds; high- olume
in es men s a egy ac oss
di e se geog aphies.
Limi ed ope a ional suppo
bu o e s global ne wo k
access and ollow-on unding
po en ial; egional o ices
p o ide local expe ise.
S a upboo camp Uni ed
Kingdom
1 2 2 0 0 1 0 0 0 0 6 Indus y- ocused global
accele a o ne wo k wi h
s uc u ed p og ams; ope a es
egionally wi h concen a ed
coho s in speci ic ma ke s.
Comp ehensi e accele a o
model wi h o mal
men o ship p og am,
ne wo king oppo uni ies
and indus y connec ions
o po olio companies.
Sou ce: www.c unchbase.com/disco e /o ganiza ion.companies; au ho ’s abula ions and in es o da abase cons uc ion; Claude.ai ( inal able compila ion assis ance).
No es: The i m desc ip ions d aw hea ily on each i m’s sel -cha ac e iza ion and public in o ma ion. The op 10 lis ing includes 11 companies due o a ie in he coun s..
Table A.4: AI Policy O e iew in Case S udy Coun ies
B azil Colombia
Na ional AI S a egies
and Agendas
I s na ional AI s a egy (2021) has pilla s o esea ch,
go e nance, wo k o ce de elopmen and in e na ional
coope a ion. Cu en AI plan 2024–2028 p oposes a
Na ional Cen e o Algo i hmic T anspa ency and
T us wo hy AI. Comp ehensi e AI bill passed Sena e
in Decembe 2024; lowe house app o al pending.
Es ablished Na ional Policy o Digi al T ans o ma ion
and AI in 2019 wi h signi ican go e nmen
commi men o implemen a ion. The policy is
comp ehensi e wi h s a egic pilla s add essing
echnology adop ion ba ie s, inno a ion condi ions,
human capi al de elopmen and AI p epa edness.
Access o AI
De elopmen Finance
Eme ging en u e capi al ecosys em is concen a ed
in São Paulo, wi h in e na ional in es o s inc easingly
ac i e. Some go e nmen g an s a ailable. In e na ional
unding access may be impac ed by di e ging egula o y
app oaches be ween B azil and majo ma ke s such as he
Uni ed S a es, po en ially complica ing c oss-bo de lows.
Go e nmen unding is a ailable ia i s en ep eneu ship
agency including a i e-yea plan and budge
alloca ion o AI; pa ne ships wi h in e na ional
o ganiza ions a e boos ing his. CEmp ende ini ia i e
p o ides suppo o AI s a -ups; inno a ion cen es
in ci ies a e also os e ing a s a -up ecosys em.
Eme ging Regula o y
F amewo k
P i acy legisla ion implemen ed in 2020 p o ides
da a p o ec ion ounda ion. B azil’s p oposed AI
egula ion adop s a isk-based app oach simila
o he EU AI Ac , equi ing algo i hmic impac
assessmen s and human o e sigh o high- isk
sys ems. The bill uniquely manda es compensa ion
o copy igh holde s when hei con en is used
o ain AI sys ems. Risk o US egula o y clash.
Regula o y sandbox app oach h ough he
Supe in endency o Indus y and Comme ce’s (Colombia’s
na ional egula o y agency) inno a ion lab. Da a
p o ec ion egime is being es ablished wi h eme ging
AI-speci ic e hical guidelines o AI deploymen .
Seeking o balance an enabling en i onmen o
inno a ion agains need o esponsible AI use.
AI Educa ion, T aining
and Skills De elopmen
An AI alen pipeline is de eloping a leading
uni e si ies. Na ional digi al skills p og am aims o
ain 100,000 p o essionals in digi al echnologies.
Bu he e is a isk o alen ligh ab oad.
“Misión TIC 2022” p og am aims o ain 100,000
p og amme s. Uni e si y p og ams in AI de eloping
wi h in e na ional academic pa ne ships.
In e na ional AI
Coope a ion
Ac i e pa icipa ion in OECD AI ini ia i es and
bila e al ag eemen s wi h Eu opean Union on AI
de elopmen . Membe o Global Pa ne ship on
AI. Signed he Ble chley Decla a ion on AI Sa e y.
Inc easing egula o y alignmen wi h he EU model.
Ac i e engagemen wi h OECD AI P inciples and
pa ne ships wi h coun ies such as Canada and Sou h
Ko ea on AI de elopmen . Such coope a ion is pu sued
o ensu e ha Colombia’s AI s ance is compe i i e.
In as uc u e Readiness Une en digi al in as uc u e wi h high connec i i y in
u ban cen es, some lag in u al a eas. Cloud compu ing
in as uc u e de eloping apidly. Da a localiza ion
equi emen s manda e cloud se ice p o ide s s o e
local copies o go e nmen da a in B azil, po en ially
posing a challenge o in e na ional AI se ices.
Signi ican in es men in digi al in as uc u e is
unde way wi h he na ional b oadband plan, hough
u al connec i i y emains challenging. B idging
he u ban- u al digi al di ide is a p io i y a ea.
Implemen a ion
Challenges
Compliance challenges om unce ain y on
en o cemen mechanisms and he u u e
egula o y au ho i y’s s uc u e.
Implemen a ion is a ec ed by poli ical
ansi ions. Public sec o ine iciencies and AI-
hesi ancy om SMEs a e cons ain s.
Indus y-Speci ic
AI Applica ions
S ong ocus on ag icul u al AI applica ions, inancial
se ices and na u al esou ces in B azil’s economic
p io i ies. Eme ging ocus on consume p o ec ion
in AI sys ems, wi h ac i e en o cemen .
Policy ocus is on public se ices, heal h-ca e
diagnos ics and ag icul u al applica ions ailo ed o
de elopmen p io i ies. Also, he e is an AI ole o
imp o e u al educa ion access and inancial inclusion.
In e na ional Regula o y
Di e gence
T ump adminis a ion’s new app oach o AI egula ion,
wi h de egula ion and heigh ened expo con ols,
may pose isk o ma ke access o B azilian s a -
ups. Also, B azil’s alignmen wi h EU-s yle egula ion
may inc ease ope a ional cos s o s a -ups.
Colombia is wo king owa d ha moniza ion wi h
in e na ional amewo ks, pa icula ly OECD
p inciples, which could help mi iga e egula o y
di e gence. US policy unce ain y poses a isk.
44 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Table A.4: Case S udy Coun y, AI Policy O e iew (Con inued)
Egyp Indonesia
Na ional AI S a egies
and Agendas
Na ional AI s a egy launched in 2021 ocusing on
capaci y building, esea ch, e hics and go e nance
wi h ambi ion o become egional AI hub.
Na ional AI s a egy (S anas KA) eleased in 2020
wi h i e p io i y a eas including heal h, bu eauc a ic
e o m, educa ion, ood secu i y and mobili y.
Access o AI
De elopmen Finance
The In o ma ion Technology Indus y De elopmen
Agency p o ides g an suppo o inno a ion.
In e na ional in es o s showing in e es in Egyp ian ech
s a -ups, hough AI-speci ic unding emains limi ed.
VC ac i i y g owing in Jaka a. Go e nmen ’s 1,000
S a -up Mo emen is p o iding some suppo o
ea ly-s age companies including AI s a -ups.
Eme ging Regula o y
F amewo k
Da a P o ec ion Law o 2020 p o ides ounda ion.
Na ional AI e hics cha e unde de elopmen
o guide esponsible AI deploymen .
Pe sonal da a p o ec ion legisla ion ecen ly
implemen ed. AI e hics guidelines unde de elopmen
h ough Minis y o Resea ch and Technology.
AI Educa ion, T aining
and Skills De elopmen
Specialized AI acul ies es ablished a uni e si ies and
Egyp ian AI cen e o educa ion. Pa ne ship wi h
in e na ional ech companies o wo k o ce aining.
Digi al alen gap add essed h ough Na ional Digi al
Talen Schola ship p og am. Uni e si y pa ne ships
wi h global ech companies de eloping AI cu iculum.
In e na ional AI
Coope a ion
Regional pa ne ships wi hin A ab League
on AI go e nance. G owing collabo a ion
wi h Chinese AI ini ia i es h ough BRI
ASEAN Digi al Inno a ion Ne wo k pa icipa ion.
Bila e al AI coope a ion wi h Singapo e and Japan.
In as uc u e Readiness Une en digi al in as uc u e wi h signi ican
in es men s in sma ci y echnologies bu
challenges in b oade connec i i y.
A chipelagic geog aphy c ea ing digi al
in as uc u e challenges. Palapa Ring p ojec
imp o ing connec i i y ac oss islands.
Implemen a ion
Challenges
Bu eauc a ic p ocesses slowing implemen a ion.
B ain d ain o echnical alen o Gul
s a es and Wes e n coun ies.
Policy implemen a ion a ies signi ican ly ac oss
di e en egions. Coo dina ion be ween mul iple
go e nmen agencies emains challenging.
Indus y-Speci ic
AI Applica ions
Focus on A abic na u al language p ocessing,
heal h-ca e diagnos ics and a chaeological
applica ions e lec ing Egyp ’s speci ic con ex .
Ma ine esou ce managemen , disas e p edic ion
sys ems and inancial inclusion e lec Indonesia’s
speci ic geog aphic and de elopmen con ex .
In e na ional Regula o y
Di e gence
Egyp ’s app oach o AI egula ion shows po en ial
di e gence wi h Wes e n amewo ks h ough i s
g owing collabo a ion wi h Chinese AI ini ia i es
ia BRI pa ne ships. This may c ea e egula o y
alignmen challenges wi h Wes e n ma ke s
as Egyp posi ions i sel as a egional AI hub
while balancing in e na ional pa ne ships.
Indonesia is wo king o align i s egula o y app oach
wi h ASEAN amewo ks while also de eloping
pa ne ships wi h Singapo e and Japan. Howe e ,
i s pe sonal da a p o ec ion amewo k may ace
challenges ha monizing wi h s ic e egimes such
as he EU Gene al Da a P o ec ion Regula ion,
po en ially a ec ing in e na ional da a lows and AI
applica ions ha ely on c oss-bo de da a sha ing.
45De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table A.4: Case S udy Coun y, AI Policy O e iew (Con inued)
Kenya Nige ia
Na ional AI S a egies
and Agendas
Na ional AI s a egy emphasizes ag icul u e, heal h ca e,
manu ac u ing and housing wi h a ocus on le e aging
Kenya’s posi ion as a egional ech hub. Consul a ion
on he s a egy inished in Janua y 2025. The d a
p omo es implemen a ion in manu ac u ing, inancial
sec o s and o he p io i y a eas. “Silicon Sa annah”
ech hub aims o os e an inno a ion ecosys em.
Na ional Digi al Economy Policy and S a egy (2020–2030)
has AI componen s. Na ional Cen e o AI and Robo ics
es ablished o coo dina e de elopmen . In Augus
2024, he d a na ional AI s a egy was published
ou lining guiding p inciples including e hics, inclusi i y,
anspa ency and isk managemen . Digi al economy:
11.30 pe cen o GDP, Q3 2024, up 14 pe cen yea -on-yea .
Access o AI
De elopmen Finance
VC in e es in Kenyan ech s a -ups is g owing.
In e na ional dono s p o ide AI- o -de elopmen
unding. Wo ld Bank app o ed US$390
million in 2023 o accele a ion p ojec .
VC in e es in Nige ian ech s a -ups is ongoing,
pa icula ly inancial echnology ( in ech). Some
go e nmen unding h ough Na ional Digi al
Inno a ion and En ep eneu ship Fund.
Eme ging Regula o y
F amewo k
The da a p o ec ion law (2019) p o ides a ounda ion. The
Task Fo ce on Blockchain and AI is de eloping e hical
guidelines. Kenya Bu eau o S anda ds published he D a
In o ma ion Technology AI Code o P ac ice in Ap il 2024
o guide esponsible AI de elopmen . Kenya Robo ics
and AI Socie y Bill in oduced in 2023 aims o c ea e
egula o y body o AI sec o . Kenya Open Da a Ini ia i e
acili a es AI de elopmen using go e nmen da a.
Nige ia’s Da a P o ec ion Regula ion (2019) p o ides
a ounda ion. Na ional AI e hics guidelines a e unde
de elopmen by he Na ional Cen e o AI and
Robo ics (NCAIR). In Decembe 2024, he House o
Rep esen a i es passed a bill o es ablish he Na ional
Ins i u e o A i icial In elligence and Robo ics S udies
Regula ion Commission, inco po a ing se e al p oposals
including he Con ol o Usage o AI Technology Bill.
AI Educa ion, T aining
and Skills De elopmen
AI academic p og ams a e linking S a hmo e and
Uni e si y o Nai obi. IBM Resea ch A ica and Google
AI Lab in Nai obi a e building esea ch capaci y. Bila e al
ag eemen s wi h Uni ed S a es and o he na ions a e
ocusing on digi al upskilling and AI capaci y building.
AI educa ion ini ia i es a leading uni e si ies. P i a e
sec o aining p og ams by companies such as Da a
Science Nige ia add essing o mal educa ion gaps.
NCAIR de eloping capaci y building p og ams.
In e na ional AI
Coope a ion
Pa icipa ion in UN AI o Good ini ia i es. Pa ne ship
wi h he Uni ed Kingdom on esponsible AI de elopmen
h ough Digi al Access P og amme. Kenya joined Pa is
Cha e on AI in Feb ua y 2025 and pa icipa es in a ious
in e na ional AI go e nance ini ia i es including he Seoul
Decla a ion on AI Sa e y and he Ble chley Decla a ion.
Limi ed o mal in e na ional AI pa ne ships, hough
engaged wi h pan-A ican AI ini ia i es. Nige ia has
signed in e na ional AI go e nance amewo ks
including he Ble chley Decla a ion on AI Sa e y (2023),
he Pa is Cha e on AI (2025) and he A ican Union’s
Con inen al AI S a egy and AI Go e nance F amewo k.
In as uc u e
Readiness
Kenya enjoys s ong mobile connec i i y in as uc u e.
High mobile money pene a ion p o ides a ounda ion
o AI in ech. By Q4 2024, mobile ne wo k connec ions
eached US$66.1 million wi h 128.3 pe cen pene a ion
a e, enabling a obus digi al ecosys em.
Elec ici y in as uc u e challenges a ec digi al
deploymen . Mobile connec i i y g owing;
b oadband access emains limi ed. In o ma ion and
communica ions sec o inc easingly impo an : 142
million ac i e in e ne subsc ibe s by Janua y 2025.
Implemen a ion
Challenges
Resou ce cons ain s o policy implemen a ion.
U ban- u al digi al di ide pe sis s. Regula o y
sys em o AI s ill in ea ly s ages.
Implemen a ion a ec ed by esou ce cons ain s
and coo dina ion challenges ( ede al/s a e).
Ea ly s age o AI egula ion ollou .
Indus y-Speci ic
AI Applica ions
Mobile-based ag icul u al ad iso y se ices, inancial
inclusion echnologies and wildli e conse a ion
e lec economic p io i ies. AI implemen a ion
p omo ed in manu ac u ing and inance.
Financial inclusion echnologies, heal h -a e
diagnos ics and na u al language p ocessing o
Nige ia’s di e se linguis ic landscape. Po en ial
o g ow h in in ech, s ong VC in e es .
In e na ional
Regula o y Di e gence
Kenya balancing A ican Union Con inen al AI
S a egy, coope a ion wi h China on AI esea ch,
EU-Sma A ica coope a ion and Wes e n-
led ini ia i es such as he Pa is Cha e .
Nige ia is na iga ing mul iple in e na ional AI
amewo ks wi h di e en app oaches, as well
as he A ican Union and he A ican Con inen al
F ee T ade A ea’s Digi al T ade P o ocol.
46 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Table A.4: Case S udy Coun y, AI Policy O e iew (Con inued)
Sou h A ica Thailand
Na ional AI S a egies
and Agendas
P esiden ial Commission on 4IR ecommenda ions
o AI adop ion. Na ional AI Policy F amewo k
published in Augus 2024 ocusing on inclusi e
economic g ow h, alen de elopmen ,
in as uc u e, inno a ion and e hical AI.
AI S a egy (2022–2027) ocuses on compe i i eness
and wo k o ce de elopmen . Thailand 4.0 policy
posi ions AI wi hin b oade digi al ans o ma ion.
The na ional AI s a egy includes an ac ion plan o
ad ance AI de elopmen ac oss sec o s including
ag icul u e, heal h ca e and inance, wi h a ocus on
building human esou ces and os e ing inno a ion.
Access o AI
De elopmen Finance
Mos de eloped en u e capi al ecosys em in
A ica. Technology Inno a ion Agency p o iding
go e nmen suppo o deep ech s a -ups.
Thailand has a g owing s a -up ecosys em wi h
inc easing en u e capi al ac i i y. Go e nmen
suppo is a ailable h ough Digi al Economy
P omo ion Agency and inno a ion unds.
Eme ging Regula o y
F amewo k
P o ec ion o Pe sonal In o ma ion Ac p o ides da a
p o ec ion ounda ion. AI e hics amewo ks unde
de elopmen h ough Depa men o Science and
Inno a ion. Na ional Da a and Cloud Policy adop ed in
Ma ch 2024 wi h p o isions a ec ing AI de elopmen .
The Pe sonal Da a P o ec ion Ac p o ides a da a
p o ec ion ounda ion. AI e hical guidelines being
de eloped h ough Digi al Economy and Socie y
Minis y. A d a Royal Dec ee on Business Ope a ions
Using AI Sys ems adop s a isk-based app oach,
ca ego izing AI sys ems in o p ohibi ed and high isk.
AI Educa ion, T aining
and Skills De elopmen
Academic AI p og ams a leading uni e si ies.
The Council o Scien i ic and Indus ial Resea ch
p o ides addi ional esea ch capaci y. Na ional
Skills Fund suppo s AI wo k o ce de elopmen .
AI educa ion p og ams a e in place a leading uni e si ies.
Digi al wo k o ce de elopmen is ongoing h ough
Thailand Massi e Open Online Cou se pla o m
In e na ional AI
Coope a ion
Ac i e pa icipa ion in BRICS coope a ion on AI
go e nance. Engagemen wi h OECD AI p inciples
despi e non-membe s a us and A ican Union’s
Con inen al AI S a egy endo sed in June 2024
Thailand pa icipa es in ASEAN’s AI go e nance
amewo k and has bila e al AI de elopmen
pa ne ships wi h Japan and Singapo e.
In as uc u e
Readiness
Mos ad anced digi al in as uc u e in Sub-Saha an
A ica, hough signi ican u ban- u al di ide emains.
In es men in elecommunica ion ne wo ks (app ox.
US$10.6bn) and da a cen es (app ox. US$1.1bn) o e he
pas i e yea s. Ch onic elec ici y supply challenges,
including load shedding and powe ou ages, c ea e
signi ican ba ie s o AI in as uc u e eliabili y,
inc easing ope a ional cos s o AI s a -ups and
po en ially limi ing compu e-in ensi e AI applica ions.
The e is a s ong digi al in as uc u e in u ban cen es,
wi h he Thailand Digi al Valley p ojec expanding
capaci y. A na ional b oadband ne wo k is unde
de elopmen . Thailand has also been wo king o
imp o e elec ical elec ici y g id eliabili y, which a ies
signi ican ly be ween u ban and u al a eas; i is deploying
sma g id echnology and ene gy s o age solu ions.
Implemen a ion
Challenges
Economic inequali y a ec ing equi able AI deploymen .
Policy implemen a ion delayed by consul a ion p ocesses.
Coo dina ion challenges be ween mul iple
agencies in ol ed in digi al policy. Skills gap
a ec ing implemen a ion o ad anced AI.
Indus y-Speci ic
AI Applica ions
Pe Sou h A ica's economic s uc u e and social
p io i ies: mining sa e y and e iciency, inancial
se ices and heal h-ca e diagnos ics:.
Thailand’s economic p io i ies o AI: medical
ou ism diagnos ics, ag icul u al moni o ing
sys ems, ou ism managemen
In e na ional
Regula o y Di e gence
Posi ioned be ween EU-s yle egula o y app oaches and
inno a ion- ocused models. G oup o 20 p esidency
o 2025 c ea es oppo uni y o shape global AI
go e nance while balancing domes ic de elopmen
needs wi h in e na ional s anda ds alignmen .
Thailand’s app oach seeks o balance inno a ion wi h
egula o y o e sigh h ough eme ging amewo ks
such as AI sandboxes and es ing cen es; he e a e
po en ial compliance challenges o global AI companies
ope a ing ac oss di e en egula o y egimes.

47De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Table A.4: Case S udy Coun y, AI Policy O e iew (Con inued)
Tunisia Vie nam
Na ional AI S a egies
and Agendas
Na ional AI s a egy (2018) posi ions Tunisia as a
egional AI hub wi h ocus on building upon a s ong
enginee ing educa ion sys em. Memo andum signed in
2022 be ween ou minis ies o ou line he de elopmen
and implemen a ion o he na ional AI s a egy.
Na ional S a egy o AI Resea ch, De elopmen
and Applica ion h ough 2030 launched in 2021
wi h aim o become egional leade in AI. The
go e nmen ad ancing i s Digi al Economy h ough
he Na ional Digi al T ans o ma ion P og am.
Access o AI
De elopmen Finance
Inno a ion suppo a ailable h ough Tunisian S a -
up Ac amewo k. In e na ional dono s p o iding
addi ional suppo o inno a ion ecosys em.
G owing en u e capi al in e es in Vie namese ech s a -
ups. Go e nmen -inno a ion unding h ough Na ional
Technology Inno a ion Fund and a ge ed AI p og ams.
Eme ging Regula o y
F amewo k
Da a p o ec ion law in place since 2004. Minis y o
Technology de eloping AI-speci ic egula ions wi h
ocus on e hics and human igh s. Na ional Au ho i y
o he P o ec ion o Pe sonal Da a es ablished,
hough wi h limi ed egula o y au ho i y.
Cybe secu i y law and da a p o ec ion egula ions
p o ide a ounda ion. Minis y o Science and
Technology de eloping AI go e nance amewo k.
Minis y o In o ma ion and Communica ions
published d a Law on Digi al Technology Indus y o
egula e AI sys ems based on isk le els (July 2024).
AI Educa ion, T aining
and Skills De elopmen
A ocus on de eloping a s ong enginee ing educa ion
p o iding he ounda ion o AI skills. Na ional AI capaci y
building p og am a ge ing 1,000 AI specialis s o 2025.
AI educa ion p og ams a leading uni e si ies. Na ional
emphasis on STEM educa ion p o iding alen pipeline
wi h go e nmen a ge o 50,000 AI specialis s by 2030.
In e na ional AI
Coope a ion
Medi e anean AI coope a ion h ough pa ne ships
wi h Eu opean coun ies. Engagemen wi h ancophone
AI esea ch ne wo ks. Tunisia signed Council o
Eu ope’s Con en ion 108 o da a p o ec ion in 2017.
G owing pa ne ships wi h Singapo e and
Sou h Ko ea on AI de elopmen . Inc easing
engagemen wi h Chinese AI ini ia i es.
In as uc u e
Readiness
Rela i ely ad anced digi al in as uc u e
o he egion, hough in es men needed
o ad anced compu ing capabili ies.
Imp o ing digi al in as uc u e and connec i i y;
high sma phone pene a ion. Na ional
p og am o digi al ans o ma ion.
Implemen a ion
Challenges
Poli ical ansi ions a e a ec ing policy
con inui y. Economic challenges limi ing
domes ic in es men capaci y.
Regula o y en i onmen s ill de eloping o keep pace
wi h echnological adop ion. Talen e en ion challenging
wi h global demand o Vie namese enginee s.
Indus y-Speci ic
AI Applica ions
Tex ile indus y op imiza ion, oli e oil p oduc ion
moni o ing and a chaeological esea ch e lec Tunisia’s
economic and cul u al con ex . AI applica ions enhancing
Public Finance Managemen In o ma ion Sys em
o de ec aud and imp o e budge e iciencies.
Based on Vie nam’s economic p io i ies and
u baniza ion challenges: manu ac u ing
op imiza ion, aquacul u e moni o ing sys ems
and u ban a ic managemen e lec ing.
In e na ional
Regula o y Di e gence
Tunisia’s da a p o ec ion amewo k shows
alignmen wi h EU s anda ds h ough adop ion o
GDPR p inciples in d a legisla ion and signing o
Con en ion 108, bu implemen a ion gaps and limi ed
en o cemen au ho i y c ea e egula o y unce ain y
o in e na ional AI de elope s and in es o s.
Vie nam’s eme ging isk-based AI go e nance has
unique ea u es p ohibi ing sys ems ha manipula e
beha iou wi hou use awa eness o classi y indi iduals
based on sensi i e in e ences. The d a Law on Digi al
Technology Indus y equi es machine- eadable labels
o AI-gene a ed con en , employs sandboxes.
Sou ces: Basic in o ma ion was ex ac ed om h ps://oecd.ai/en/dashboa ds/o e iew, supplemen ed by da a om
Digi al Policy Ale Ac i i y T acke online policy da abase; and h ps://globalailaw.com/colombia-ai-policy/; Buza,
Jenze and Bossa d (2024); Buza and Scheiwile (2024); Buza and Taha (2025a, 2025b); Deeg and Pie o ic (2024); Gia dini,
Scheiwile and Buza (2024); Lippold (2024); Filguei as and Junquilho (2023); Ma ins (2025); Villaseno and Tu ne
(2024); Ca illo e al. (2025); OECD (2024, chap e 2); Wadipalapa e al. (2024); República de Colombia (2024); h ps://
menaobse a o y.ai/en/ egional/20; Élysée (2025). The au ho ’s compila ion o his able was suppo ed by Claude.ai.
No e: The Pa is Cha e is an acco d among 10 na ions es ablishing p inciples o openness, accoun abili y and
pa icipa ion in AI go e nance (Élysée 2025).
48 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Appendix 3: OECD
Recommenda ion o
he Council on A i icial
In elligence
The OECD p inciples on AI wi hin he
OECD’s Recommenda ion o he Council
on A i icial In elligence1 s a e ha :
→ AI should bene i people and he plane
by d i ing inclusi e g ow h, sus ainable
de elopmen and well-being.
→ AI sys ems should be designed in a way
ha espec s he ule o law, human igh s,
democ a ic alues and di e si y, and hey
should include app op ia e sa egua ds — o
example, enabling human in e en ion whe e
necessa y— o ensu e a ai and jus socie y.
→ The e should be anspa ency and esponsible
disclosu e a ound AI sys ems o ensu e ha
people unde s and when hey a e engaging wi h
hem and can challenge ou comes.
→ AI sys ems mus unc ion in a obus , secu e
and sa e way h oughou hei li e imes, and
po en ial isks should be con inually assessed
and managed.
→ O ganiza ions and indi iduals de eloping,
deploying o ope a ing AI sys ems should be
held accoun able o hei p ope unc ioning in
line wi h he abo e p inciples.
1 See OECD (2019). Also no e ha as o Ma ch 25, 2025, he OECD
legal ins umen s online da abase epo s ha adhe ence o he
ecommenda ion now co e s all 38 OECD membe coun ies and he
Eu opean Union, plus nine o he s: A gen ina, B azil, Egyp , Mal a, Pe u,
Romania, Singapo e, Uk aine and U uguay.
The OECD ecommends ha go e nmen s:
→ acili a e public and p i a e in es men in R&D
o spu inno a ion in us wo hy AI;
→ os e accessible AI ecosys ems wi h
digi al in as uc u e and echnologies, and
mechanisms o sha e da a and knowledge;
→ c ea e a policy en i onmen ha will open
he way o he deploymen o us wo hy AI
sys ems;
→ equip people wi h he skills o AI, as well as
suppo wo ke s o ensu e a ai ansi ion; and
→ coope a e ac oss bo de s and sec o s o sha e
in o ma ion, de elop s anda ds and wo k owa d
he esponsible s ewa dship o AI
49De eloping Coun ies’ Business Pa icipa ion in he AI Economy
Wo ks Ci ed
Aa onson, Susan A iel. 2024a. The Age o AI Na ionalism
and I s E ec s. CIGI Pape No. 306. Wa e loo,
ON: CIGI. www.cigionline.o g/publica ions/
he-age-o -ai-na ionalism-and-i s-e ec s/ .
———. 2024b. “Da a Go e nance Is No Ready o AI.” In
The Role o Go e nance in Unleashing he Value o
Da a, 15–23. Wa e loo, ON: CIGI. www.cigionline.
o g/a icles/da a-go e nance-is-no - eady- o -ai/.
Abungu, Cecile, Ma ie Vic oi e I adukunda, Duncan Cass
Beggs, Aquila Hassan and Raqda Sayidali. 2024. Why
Global Sou h Coun ies Need o Ca e Abou Highly
Capable AI. CIGI Pape No. 311. Wa e loo, ON: CIGI.
www.cigionline.o g/publica ions/why-global-sou h-
coun ies-need- o-ca e-abou -highly-capable-ai/.
Al en, No a. 2024. “Unde s anding he S ages o S a up
Funding.” Way a (blog), May 15. www.way a.de/
blog/unde s anding- he-s ages-o -s a up- unding.
Amodei, Da io. 2024. “Machines o Lo ing G ace. Sec ion 3.
Economic De elopmen and Po e y.” Oc obe . www.
da ioamodei.com/essay/machines-o -lo ing-g ace.
Asia-Paci ic Economic Co po a ion. 2024. “Digi al Economy
Pa ne ship Ag eemen .” Submission by Chile.
2024/SOM3/CTI/004. h ps://mddb.apec.o g/
Documen s/2024/CTI/CTI3/24_c i3_004_ .pd .
A anaso ska, Daniela and Lejla Robeli. 2025. “B azil’s AI Ac : A
New E a o AI Regula ion.” GDPR Local, Feb ua y 26. h ps://
gdp local.com/b azils-ai-ac -a-new-e a-o -ai- egula ion/.
Beckley, Michael. 2025. “The Age o Ame ican Unila e alism:
How a Rogue Supe powe Will Remake he Global
O de .” Fo eign A ai s, Ap il 16. www. o eigna ai s.
com/uni ed-s a es/age-ame ican-unila e alism.
Bommasani, Rishi and Pe cy Liang. “Re lec ions on Founda ion
Models.” S an o d Uni e si y Human-Cen e ed
A i icial In elligence. Oc obe 18. h ps://hai.s an o d.
edu/news/ e lec ions- ounda ion-models.
B emme , Ian and Mus a a Suleyman. 2023. “The AI Powe Pa adox:
Can S a es Lea n o Go e n A i icial In elligence — Be o e I ’s
Too La e?” Fo eign A ai s 102 (5): 26–43. www. o eigna ai s.
com/wo ld/a i icial-in elligence-powe -pa adox.
Buza, Ma ia, Philine Jenze and S enja Bossa d. 2024. “Digi al
Policy Ale Digi al Diges : Thailand.” Oc obe . h ps://
digi alpolicyale .o g/diges /dpa-digi al-diges - hailand.
Buza, Ma ia and Luca Scheiwile . 2024. “Digi al Policy
Ale Digi al Diges : Vie nam.” Oc obe . h ps://
digi alpolicyale .o g/diges /dpa-digi al-diges - ie nam.
Buza, Ma ia and She i Taha. 2025a. “Digi al Policy Ale
Digi al Diges : Kenya.” Ma ch. h ps://digi alpolicyale .
o g/diges /dpa-digi al-diges -kenya.
———. 2025b. “Digi al Policy Ale Digi al Diges : Nige ia.” Ma ch.
h ps://digi alpolicyale .o g/
diges /dpa-digi al-diges -nige ia.
Ca illo, Pablo E., Ma in J. Mackowski, Wol gang A. Maschek, Be h
L. Golds ein, Julia B. Jacobson, Alan L. F iel and Ma hew Ki k.
2025. Key Insigh s on P esiden T ump’s New AI Execu i e
O de and Policy & Regula o y Implica ions. Squi e Pa on
Boggs. Feb ua y. www.squi epa onboggs.com/
en/insigh s/publica ions/2025/02/key-
insigh s-on-p esiden - umps-new-ai-execu i e-
o de -and-policy- egula o y-implica ions.
CB Insigh s 2021. The Top 12 Reasons S a ups Fail. www.cbinsigh s.
com/ epo s/CB-Insigh s_Top-Reasons-S a ups-Fail.pd .
C aw o d, Rode ick and Ma k Smi h. 2022. How To Sol e he
Da a Science Skills Sho age. SAS Ins i u e. www.sas.com/
con en /dam/SAS/documen s/ echnical/educa ion/
en/sol e-da a-science-skills-sho age-uk-113039.pd .
Deeg, Nils and Tania Pie o ic. 2024. “Digi al Policy Ale Digi al
Diges : Sou h A ica (2024 Edi ion).” Decembe . h ps://
digi alpolicyale .o g/
diges /dpa-digi al-diges -sou h-a ica.
den Bes en, Ma hijs L. 2020. “C unchbase Resea ch: Moni o ing
En ep eneu ship Resea ch in he Age o Big Da a.” SSRN,
No embe 3.
h ps://ss n.com/abs ac =3724395.
Elsne , Ma k, G ace A kinson and Saadia Zahidi. 2025.
The Global Risks Repo 2025. Cologny, Swi ze land:
Wo ld Economic Fo um. Janua y. www.we o um.
o g/publica ions/global- isks- epo -2025/.
Élysée. 2025. “The Pa is Cha e on A i icial In elligence in he
Public In e es .” O ice o he P esiden o F ance, Feb ua y 11.
www.elysee. /en/emmanuel-mac on/2025/02/11/ he-
pa is-cha e -on-a i icial-in elligence-in- he-public-in e es .
Fan, Qimiao, and Ch is ine Zhenwei Qiang. 2024. “Tipping he
scales: AI’s dual impac on de eloping na ions.” Digi al
T ans o ma ion (Wo ld Bank blog), June 3. h ps://
blogs.wo ldbank.o g/en/digi al-de elopmen / ipping-
he-scales--ai-s-dual-impac -on-de eloping-na ions.
50 CIGI Pape No. 323 — June 2025 • Douglas Lippold
Filguei as, Fe nando and Tainá Aguia Junquilho. 2023. “The
B azilian (Non)pe spec i e on na ional s a egy o
a i icial in elligence.” Disco e A i icial In elligence 3
(7). h ps://doi.o g/10.1007/s44163-023-00052-w.
Gao, Xueyuan and Hua Feng. 2023. “AI-D i en P oduc i i y Gains:
A i icial In elligence and Fi m P oduc i i y.” Sus ainabili y
15 (11): 8934. h ps://doi.o g/10.3390/su15118934.
G7 I alia. 2024. “Minis e ial Mee ing on Technology and Digi al: G7
Join S a emen .” Oc obe 15. www.soumu.go.jp/
hi oshimaaip ocess/pd /documen 08_en.pd .
Ge adin, Damien. 2017. “Ensu ing sound egula o y p ocesses:
Fo a p incipled app oach.” TILEC Discussion Pape No.
2017-030. July. h p://ss n.com/abs ac =3011188.
Gia dini, Tommaso, Luca Scheiwile and Ma ia Buza. 2024.
“Digi al Policy Ale Digi al Diges : B azil.” Decembe . h ps://
digi alpolicyale .o g/
diges /dpa-digi al-diges -b azil.
Go e nmen o Canada. 2024. “G7 I alia 2024– Add essing global
challenges, os e ing pa ne ships.” Fo eign Minis e s’ mee ing
communiqué, Ap il 19. www.in e na ional.gc.ca/
wo ld-monde/in e na ional_ ela ions- ela ions_
in e na ionales/g7/documen s/2024-04-19-
pa e ships-pa ena ia s.aspx?lang=eng.
Go e nmen o Japan. 2024. “Documen s o Achie emen .”
Hi oshima AI P ocess. www.soumu.go.jp/hi oshimaaip ocess/
en/documen s.h ml.Achie emen .” Hi oshima AI P ocess.
www.soumu.go.jp/hi oshimaaip ocess/en/documen s.h ml.
GOV.UK. 2023. The Ble chley Decla a ion by Coun ies A ending
he AI Sa e y Summi , 1–2 No embe 2023. Policy Pape .
www.go .uk/go e nmen /publica ions/ai-sa e y-summi -
2023- he-ble chley-decla a ion/ he-ble chley-decla a ion-by-
coun ies-a ending- he-ai-sa e y-summi -1-2-no embe -2023.
Helle , Da id and Dominik Asam. 2024. “Gene a i e AI and Fi m-
le el P oduc i i y: E idence om S a up Funding Dynamics.”
SSRN, Augus 21. h ps://ss n.com/abs ac =4877505.
Hi e , Shelby. 2024. “150 Top AI Companies: Visiona ies D i ing
he AI Re olu ion.” eWeek, Oc obe 1. www.eweek.com/
a i icial-in elligence/ai-companies/. (No e: Upda es
o his a icle a e pos ed pe iodically by eWeek.)
IFAC-BIAC. 2018. Regula o y Di e gence: Cos s, Risks, Impac s.
In e na ional Financial Sec o S udy a OECD and IFAC.
h ps://bdi.eu/media/
hemen elde /konjunk u _und_ inanzmae k e/
downloads/Business_a _OECD-IFAC_Cos s_
o _Regula o y_Di e gence.pd .
Ipsos. 2023. Global Views on A.I. 2023. Global
Ad iso su ey. July. www.ipsos.com/si es/
de aul / iles/c /news/documen s/2023-07/
Ipsos Global AI 2023 Repo -WEB_0.pd .
Kalash, S. Yash. 2024. “AI and he Digi al Economy F amewo k
Ag eemen : P epa ing ASEAN o he Nex Tech
Wa e.” Opinion, Cen e o In e na ional Go e nance
Inno a ion, No embe 11. www.cigionline.o g/a icles/
ai-and- he-digi al-economy- amewo k-ag eemen -
p epa ing-asean- o - he-nex - ech-wa e/.
Ko inek, An on. 2024. “Economic Policy Challenges o he Age o
AI.” Na ional Bu eau o Economic Resea ch Wo king Pape
32980. Sep embe . www.nbe .o g/pape s/w32980.
Ko inek, An on and Joseph E. S igli z. 2021. “A i icial In elligence,
Globaliza ion, and S a egies o Economic De elopmen .”
Na ional Bu eau o Economic Resea ch Wo king Pape
28453. Feb ua y. www.nbe .o g/pape s/w28453.
K ugman, Paul. 1995. De elopmen , Geog aphy and
Economic Theo y. Camb idge, MA: The MIT P ess.
Lippold , Douglas. 2024. AI Inno a ion Concen a ion and he
Go e nance Challenge. CIGI Pape No. 292. Wa e loo,
ON: CIGI. www.cigionline.o g/publica ions/ai-inno a ion-
concen a ion-and- he-go e nance-challenge/.
Lippold , Douglas C. and Ma k F. Schul z. 2014. “Unco e ing
T ade Sec e s – An Empi ical Assessmen o Economic
Implica ions o P o ec ion o Undisclosed Da a” OECD
T ade Policy Pape No. 167. Pa is, F ance: OECD Publishing.
Augus 11h p://dx.doi.o g/10.1787/5jxzl5w3j3s6-en
Lund, B ady and Wang Ting. 2023. “Cha ing abou
Cha GPT: How may AI and GPT impac academia
and lib a ies?” Lib a y Hi Tech News 40 (3): 26–29.
h ps://doi.o g/10.1108/LHTN-01-2023-0009.
Mackowski, Ma in, Pablo Ca illo, and Julia Jacobson. 2025.
“T ump’s AI Policy Shi P omo es US Dominance and
De egula ion.” Bloombe g Law, Feb ua y 26. h ps://
news.bloombe glaw.com/us-law-week/ umps-ai-policy-
shi -p omo es-us-dominance-and-de egula ion.
Mannu u, Nishi h Reddy, Sakib Shah ia , Zoë A. Teel, Ting Wang,
B ady D. Lund, Solomon Tijani, Chale mchai Oak Pohboon e
al. 2023. “A i icial in elligence in de eloping coun ies: The
impac o gene a i e a i icial in elligence (AI) echnologies o
de elopmen .” In o ma ion De elopmen : 1–19. Sep embe .
h ps://doi.o g/10.1177/02666669231200628.
Ma ins, Laís. 2025. “B azil’s AI Law Faces Unce ain Fu u e
as Big Tech Wa ms o T ump.” Tech Policy P ess,
Feb ua y 4. www. echpolicy.p ess/b azils-ai-law-
aces-unce ain- u u e-as-big- ech-wa ms- o- ump.
Maslej, Nes o , Lo edana Fa o ini, Raymond Pe aul , Vanessa Pa li,
Anka Reuel, E ik B ynjol sson, John E chemendy e al. 2024.