The Algorithm #4: Sovereignty is real.
Signals from Africa, India, and Latin America on who builds AI and how it impacts the communities building around it.
AI is being built, regulated, and resisted across the Global South at a pace that rarely makes it into mainstream tech coverage. Each week, The Algorithm tracks the most important developments, new models, policy moves, funding gaps, and the harder questions about who AI is actually for.
🌍 Global South
1. Rwanda Becomes the First African Country With a Dedicated National AI Agency
Ecofin Agency, June 8, 2026 | ecofinagency.com
Rwanda’s cabinet approved a National Artificial Intelligence Agency on June 8, making it the first African nation to move from an AI strategy document to a standalone institution. The agency will govern AI deployment across public administration, healthcare, education, and agriculture.
A separate AI Scaling Hub has secured 25 billion Rwandan francs (~$17 million) to drive adoption across the economy. Rwanda is now operating on a fundamentally different timeline from most of its continental peers: while 38 of 54 African countries still lack any formal AI policy, Rwanda has signed an MOU with Anthropic, hosted the Africa AI Council secretariat, and is now building the institutional infrastructure to match its ambitions. The question the rest of Africa will be watching is whether this governance-first approach attracts investment or simply adds bureaucracy.
2. A US Export Control Blackout Overnight Reminded the World Why It Needs Its Own AI
Inc42, June 13, 2026 | inc42.com
Three days after launching its most capable models, Claude Fable 5 and Mythos 5, Anthropic suspended access for all non-US users, including its own non-US employees, after a US government export control directive citing a jailbreak vulnerability.
The overnight blackout immediately reignited the world’s sovereign AI debate: if frontier model access can vanish at Washington’s discretion, what does “AI-powered” mean for international enterprises, startups, and government services built on US platforms?
The episode is the clearest real-world illustration yet of what AI dependency looks like when the dependency is weaponisable, and it landed in the same week Modi and Macron met in Nice to jointly advocate for “open, ethical AI” outside US and Chinese frameworks.
3. Rio Is Betting $65 Billion on Becoming a Global AI Hub and Web Summit Gave It a Stage
Prefeitura Rio / Breaking Travel News, June 9–12, 2026 | prefeitura.rio
Web Summit Rio 2026 drew a record 40,287 attendees and 1572 startups from over 100 countries (June 8–11), with AI & machine learning as the leading sector and a new “Machine Summit” track debuting.
On the sidelines, Rio’s city government announced “Rio AI City”: a US$65 billion, decade-long plan to position Rio as a top-10 global AI hub by 2032, anchored by a US$550 million infrastructure deal with Elea Data Centers.
The city also presented IplanRio’s WhatsApp-based AI agent, winner of Bloomberg Philanthropies’ 2025–26 Mayors Challenge, as its flagship citizen-services model, and launched Rio.IA Saúde to connect AI startups with public health institutions Fiocruz and PUC-Rio. The ambition is significant; the test will be whether investment follows the announcement, and whether the AI city serves favela residents or just Leblon.
🌐 Everywhere else
4. AI’s Environmental Bill Is Being Sent to the Wrong Address
UN News / United Nations University (UNU-INWEH), June 4, 2026 | un.org
A United Nations University report published June 4 warns that AI data centres could consume 945 terawatt-hours of electricity annually by 2030, nearly triple the combined power usage of Pakistan, Bangladesh, and Nigeria. More than 90% of AI-specialised computing capacity sits in the US and China, while over 150 nations have no significant domestic AI infrastructure. But the environmental cost is distributed inversely: rare mineral extraction for AI hardware and up to 2.5 million tonnes of annual e-waste fall disproportionately on lower-income countries. The UNU researchers call it an asymmetry of benefit and burden, the Global South powers the AI economy through minerals and labour while bearing its environmental consequences, without meaningfully participating in its governance or returns.
5. The US and China Own 90% of the Compute. Everyone Else Is Negotiating the Terms of Dependence.
Rest of World, June 9, 2026 | restofworld.org
At a Rest of World event during New York Tech Week, researchers from the Carnegie Endowment for International Peace, Cornell University, and Access Now presented a blunt assessment: the US and China control 90% of global computing power and 70–80% of global AI investment.
Current AI is built for “WEIRD” societies, Western, Educated, Industrialised, Rich, Democratic, which represent just 15% of the global population. Frontier model providers like Anthropic are already shifting to “managed-access” rollouts, selecting which countries receive access to their most capable systems.
Panellists identified potential leverage for the Global South, critical minerals, energy resources, battlefield data, but noted a structural irony: even the most advanced sovereign AI projects (India’s IndiaAI Mission, the UAE’s Intelligence Grid) remain entirely dependent on US-designed Nvidia chips. Sovereignty is real but partial, and the terms are still set in San Jose.
That’s it for this edition of The Algorithm. If you spotted a story we missed, a startup, a policy move, a community using AI in ways nobody’s writing about yet, reply to this email.