Soft Power, Hard Compute: The AI Geopolitical Reset
The new Digital Oil diplomacy is here
Artificial Intelligence (AI) is poised to define the geopolitical and economic architecture of the 21st century, much like oil did in the 20th. While developed nations scale compute capacity, model development, and digital regulation, a quieter yet equally significant transformation is unfolding across the Global South. From Riyadh to Nairobi, Jakarta to Brasília, emerging economies are no longer mere consumers of AI technologies—they are becoming strategic co-authors of the intelligence age.
This paper outlines the contours of this shift, evaluating AI’s geopolitical logic through regional analysis. It frames infrastructure needs, public-private partnerships, energy dynamics, and public sector integration as foundational to strategic autonomy. Building on my prior work around the concept of the "AI dollar," we show how emerging markets are transitioning from data importers to data originators, model developers, and policy innovators.
The Rise of the Global South: Strategic Autonomy in the Age of Intelligence
For too long, the AI race has been portrayed as a binary contest between the United States and China. Yet, as Tobias Feakin has argued, this framing misses a crucial force: the Global South. Representing 85% of the world’s population, these countries are asserting influence across three axes:
Data Diversity: Contributing unique datasets (e.g., African health diagnostics, Southeast Asian linguistic corpora) that enhance model generalization.
Policy Innovation: Countries like Brazil and Indonesia have advanced AI roadmaps emphasizing ethical governance, agricultural resilience, and smart infrastructure.
Multilateral Influence: Through platforms like UNESCO, the African Union, and ASEAN, these nations are embedding Global South priorities in international AI norms.
China, through its Digital Silk Road, offers cost-effective, integrated AI ecosystems that have found traction across Africa, Southeast Asia, and the Middle East. While these solutions promise rapid digital transformation, they often introduce long-term dependencies, embedding Chinese standards and models into public governance systems.
Conversely, the United States' recent pivot, from multilateralism under Biden to infrastructure-centric nationalism under Trump, underscores a new strategic vision. The Stargate initiative, a $500 billion partnership with OpenAI, Oracle, and SoftBank, exemplifies this domestic-first approach. Yet its closed-source model risks marginalizing the very nations it seeks to influence unless it evolves to embrace open-source and flexible deployments.
Open-source AI, advanced by companies like Meta and Google, is emerging as a critical lever. It allows for model localization, sovereign deployment, and collaborative governance, attributes increasingly valued by emerging markets seeking digital sovereignty without sacrificing transparency or ethical oversight.
Moreover, the Global South's active engagement in multilateral institutions has elevated its role from rule-taker to rule-maker. Brazil’s leadership in drafting UNESCO’s AI Ethics Guidelines and the African Union’s AI Working Group are powerful signals that AI governance is becoming more distributed, culturally responsive, and democratic in origin.
Personal Observations from the Field: AI as Infrastructure
In the early 2000s, while working in Africa and Latin America, I observed how profit and donor-driven solar panel projects promised electrification but failed to deliver systemic change. The lesson was clear: infrastructure without ownership is fragile.
Later, at General Electric, it became evident that energy resilience depended not just on kilowatts but on governance, regulation, and localized capacity.
The same is now true for AI. Frontier models without compute infrastructure, without sovereign data strategies, are empty vessels. AI must be treated not as a software service but as a national asset. Policymakers, investors, and engineers in the Global South increasingly understand this. The AI revolution is not about apps, it is about energy grids, fiber networks, talent development, and strategic alliances.
Hence, when OpenAI quietly launched its “OpenAI for Countries” initiative, most coverage focused on democratizing access to its frontier models. Yet this initiative may be one of the most consequential geopolitical maneuvers of the decade, fusing AI, infrastructure diplomacy, and strategic alignment into a soft power tool more potent than anything the digital era has seen.
Regional Analysis: Infrastructure and Influence
North America: Leads in AI infrastructure with 37.4% of global share. Nvidia's $500 billion supercomputer initiative and Pennsylvania’s AI-energy hub show a strong centralized strategy focused on domestic capacity.
Europe: Pursuing centralized sovereignty via €20 billion for AI “gigafactories.” Adoption remains high (41.17% among large enterprises), but cost of energy impedes competitiveness.
Asia-Pacific: A hybrid model emerges. Japan, Singapore, and India invest in centralized cloud but also support decentralized innovation hubs. India’s TRUST program exemplifies this.
Middle East & Africa: Saudi Arabia, via the $100 billion HUMAIN initiative, leads a centralized compute expansion with Nvidia, AMD, and AWS. Africa, especially Kenya and Nigeria, is exploring decentralized models using mobile-first AI and regional data networks.
Latin America: Brazil champions decentralized models through local LLMs and judicial AI. Argentina and LATAM 4.0 support hybrid infrastructure, centralized where viable, decentralized for community access.
The Strategic Compute Dilemma: Centralized vs. Decentralized
Centralized AI models are exemplified by U.S. and Chinese investments in megascale infrastructure, Stargate, DeepSeek, Huawei’s turnkey cloud solutions. They offer efficiency and integration, but risk dependency and data centralization.
In contrast, decentralized AI, supported by networks such as SingularityNET, Fetch.ai, Ocean Protocol, and EleutherAI, offers sovereignty, flexibility, and inclusion. Africa’s leapfrogging strategies, Brazil’s localized judiciary models, and India’s open-source deployments all point to a future where compute is as much about ideology as infrastructure.
Energy: The Ultimate AI Constraint
AI is an electricity-intensive endeavor. With data center demand projected to reach 1,000 TWh by 2030, energy sovereignty becomes essential. The global demand is not evenly distributed, and regional strategies must reflect varying electricity baselines and growth projections:
North America: U.S. data centers currently consume around 90 TWh annually, projected to exceed 180 TWh by 2030. Projects like Nvidia’s domestic supercomputing hubs will require multi-GW energy zones, often drawing on fossil-heavy grids unless paired with renewables.
Europe: EU data center electricity use is expected to climb from 85 TWh to over 150 TWh by 2030, but high industrial tariffs and decarbonization mandates add strain. Germany’s AI hubs are increasingly reliant on imported energy and off-grid wind farms.
Asia-Pacific: With $110 billion in AI infrastructure investments projected by 2028, regional energy demand could surpass 250 TWh by the end of the decade. Singapore and Japan are shifting to offshore wind and LNG co-generation to meet these needs.
Middle East & Africa: Saudi Arabia's AI-energy strategy includes dedicated solar gigawatt farms linked to HUMAIN’s data centers. The Kingdom is expected to deploy over 5 GW of new capacity for AI workloads by 2030, equivalent to 40 TWh/year. In Africa, energy access disparities make decentralized solar-powered data nodes increasingly attractive.
Latin America: Brazil and Mexico are forecasting 60–80 TWh in AI-related energy demand by 2030. Hydropower remains central, but variability and droughts pose risks. Countries are experimenting with microgrid-enabled edge compute.
Saudi Arabia's integration of solar power into its AI strategy, linking data centers directly to renewable sources with sovereign grid control, offers a replicable model for other emerging markets navigating the AI-energy nexus.
Recommendations for Strategic AI Autonomy
Energy-AI Co-location: Prioritize green energy zones tied to data centers.
Sovereign Compute Funds: Develop national GPU banks and co-finance sovereign clouds.
Fiscal Incentives: Create AI Free Zones with tax benefits for infrastructure and R&D.
Open-Source Integration: Mandate open architecture for public sector AI procurement.
Regional Alliances: Support ASEAN, AU, and LATAM multilateral AI frameworks.
Transparent Diplomacy: Invest in capacity-building, not dependency.
Conclusion: Rewriting the Intelligence Order
We are entering a new age where strategic advantage will hinge not on territory or tariffs, but on compute, code, and control over digital infrastructure. The nations that dominate this landscape won’t be those with the largest armies or oil reserves, but those that command the means to train, deploy, and govern intelligence.
The decision now facing countries is foundational: build sovereign AI stacks, complete with energy, data, and talent, or tether their futures to external platforms and regulatory regimes. Saudi Arabia’s $100 billion HUMAIN project exemplifies how nations are turning compute into a geopolitical asset, akin to oil in the last century.
But this moment is not just about hardware. It is about values, alignment, and architectural influence. OpenAI’s multilateral expansion strategy, through its “For Countries” program, redefines soft power. It exports not just access but governance models, embedding a Western-centric blueprint of the AI future.
This isn’t merely a race, it’s a redrawing of the global order. Intelligence, like energy before it, will define influence. And in this century, sovereignty may well be spelled in GPUs.
Sources:
International Energy Agency. “AI is set to drive surging electricity demand from data centres.” IEA.
Goldman Sachs. “AI to Drive 165% Increase in Data Center Power Demand by 2030.” Goldman Sachs.
Scientific American. “AI Will Drive Doubling of Data Center Energy Demand by 2030.” Scientific American.
Wang, Gerui. “Stargate’s $500 Billion AI Bet: Have We Forgotten the Hidden Cost?” Forbes.
“AI Power Needs May Strain US Grid as Stargate Project Grows.” The New Stack.
International Monetary Fund. “AI Needs More Abundant Power Supplies to Keep Driving Economic Growth.” IMF Blog.
“Nvidia to Supply 18,000 AI Chips to Saudi Arabia’s HUMAIN.” AP News.
“All About HUMAIN: Saudi Crown Prince’s AI Initiative.” Economic Times.
“US Companies Are Helping Saudi Arabia Build an AI Powerhouse.” Computerworld.
McKinsey & Company. “The Role of Power in Unlocking the European AI Revolution.” McKinsey.
“Asia Pacific AI in Energy Market Outlook.” Grand View Research.
PwC Middle East. “Unlocking the Data Centre Opportunity in the Middle East.” PwC.
“DeepSeek’s Energy Consumption Cuts AI’s Power Use by 75%.” Rinnovabili.
“AI’s Impact on Clean Energy and Data Centers in Asia-Pacific.” Sitetracker.