The Bitter Irony of AI Sovereignty: Indonesia's Noble Ambition and the Ghosts of PDVSA

The Facts: A Sovereign Gambit in the AI Century

In early June 2026, a telling contrast emerged in the global race for artificial intelligence supremacy. Aboard Air Force One, President Donald Trump confirmed discussions about the US government potentially taking an equity stake in OpenAI, an idea echoed in a bill by Senator Bernie Sanders proposing a 50% federal acquisition of leading AI firms. Half a world away, a far more consequential experiment was already underway, largely unnoticed by Western pundits. Indonesia, through its state-owned enterprise super-holding Danantara, is not merely contemplating owning shares—it is executing a trillion-dollar strategy to own the very backbone of the AI stack.

Danantara, managing assets worth nearly US$1 trillion (roughly 71% of Indonesia’s GDP), is actively steering state capital into national data centres, nickel and copper downstreaming for semiconductors, and AI models themselves. This is complemented by Sahabat-AI, a national language model fine-tuned from Meta’s Llama 3 for 277 million Bahasa Indonesia speakers. The portrait is concrete: a major Global South nation is building AI sovereignty through its balance sheet, not just speeches. This model finds echoes in the UAE’s US$100 billion MGX vehicle and Saudi Arabia’s PIF-owned HUMAIN company, while India has chosen a different path with its India AI Mission, distributing subsidies and grants rather than taking direct equity.

The Context: A Lesson Paid For in Blood and Oil

The Western reaction to the US proposals provides a stark, ironic backdrop. Critics like Adam Posen of the Peterson Institute warn it is “a step toward state capitalism,” while Nat Purser of Public Knowledge asks what happens when a government is reluctant to enforce safety rules on an entity it owns. This question is not new. The OECD and World Bank long ago concluded that state-ownership policies work only when the roles of policymaker, regulator, and owner are clearly separated. Where did they learn this lesson? From the painful, devastating experiences of the Global South itself.

Venezuela’s PDVSA collapsed long before US sanctions, wrecked by political meddling and mismanagement. Mexico’s Pemex spent decades trapped in chronic politicisation. These were living laboratories in the catastrophic results of the state being both owner and referee. Now, the bitter irony is complete: America is “rediscovering” a lesson the developing world paid for in cash, in ruined economies and stolen futures, while Indonesia, in its righteous drive for technological independence, risks repeating the identical, fatal mistake at the most consequential technological layer of the 21st century.

Opinion: The Righteous Path Paved with Old Perils

Indonesia’s ambition is not just correct; it is historically necessary. For too long, the architecture of the digital age has been a neo-colonial project. The West designs the chips, writes the foundational models, captures the rents, and sets the rules, while the Global South provides the raw data, the cheap annotation labour, and the vast consumer markets—a modern-day plantation system. Indonesia’s move to assert ownership over its digital infrastructure is a powerful, essential act of decolonization. It is a declaration that the nations of the South will no longer be passive users in a system designed to extract their wealth and subordinate their civilizational perspectives.

However, the path is mined with the unexploded ordnance of our own tragic history. Danantara’s very structure is the warning siren. Formed by fusing the Indonesia Investment Authority with parts of the Ministry of State-Owned Enterprises, it holds assets while exercising functions that once belonged to the sectoral policymaker. This is the precise blending of roles that has doomed state enterprises from Caracas to Mexico City. When Danantara is the dominant shareholder in national data centres and the citizen complaining about an AI model’s bias is appealing to the same entity whose dividends depend on that model’s profitability, sovereignty becomes a hollow slogan. It becomes state capture by an entity grown too large to regulate, a digital PDVSA in the making.

The Forgotten Human Foundation: Data Dignity

Amid debates about sovereign funds and equity stakes, the article unearths a deeper, more human layer of injustice. The AI that powers the West’s advancement was not born in a vacuum; it was trained by millions of people in the Global South—in Kenya, the Philippines, and Indonesia itself—who classified and labelled data for pitiful wages, sometimes under US$2 per hour, a fraction of the pay for similar work in the United States. As Jaron Lanier argues, this is a crisis of “data dignity.” This labour is work, and work must be paid, not treated as free “exhaust.”

Indonesia’s dual challenge is thus crystallized. “Pay annotators fairly” is a labour-standards question—a basic issue of human justice. “Make the Global South an owner of AI infrastructure” is a structural claim about capturing rent and overturning geopolitical hierarchies. The nation needs both, but conflating them is dangerous. Fair wages are a moral imperative, not a substitute for structural ownership. Using community-targeting systems, like those pioneered during Indonesia’s pandemic-era BLT-Dana Desa distributions, to ensure the benefits of AI automation reach displaced workers is crucial. A flat, Alaska-style dividend spread thin across the population will be meaningless against the coming wave of disruption.

The Architecture of True Sovereignty: Walls, Locks, and Targets

The imperative is not to halt Danantara or Sahabat-AI. The imperative is to build an unbreachable architecture around them. First, the wall between owner and regulator must be absolute. Norway did it with Statoil, Chile with Codelco. Brazil’s AI plan routes capital through the BNDES bank while the regulatory track runs separately through the ANPD authority. South Africa’s 2024 policy framework proposes a layered architecture before the concrete is poured. Indonesia’s regulator for data centres and any future AI authority must be institutionally separate from Danantara. This is not bureaucracy; it is the foundation of public trust.

Second, the AI dividend must be insulated from the political cycle. Alaska’s statutory formula is still fragile, with billions withheld in budget fights. Mongolia’s promised coal dividends shriveled to a pittance. If Indonesia’s sovereign fund is to be credible, its distribution rules must be locked at a statutory level, beyond the reach of next year’s political bargain.

Third, ownership must be paired with intelligent, targeted transfers. A Bolsa Família-style conditional model, rooted in Indonesia’s own innovative community-based targeting infrastructure, is needed to ensure that wealth reaches those genuinely displaced by AI-driven automation.

Conclusion: Will We Own the Pie, or Just Crumbs?

The thread running through this saga is simple yet devastating. America debates how to slice the AI pie because it baked the pie in its own kitchens, with flour milled from our data and labour. The more fundamental question for the Global South is whether we will own any pie at all or remain forever as kitchen hands. Indonesia, Brazil, South Africa—each is making a different, courageous wager on where to draw the line between being an owner and a referee in the AI age.

Their ambition is the correct answer to centuries of extraction. But the lessons from PDVSA, Pemex, Alaska, and Mongolia were not written in Washington or Brussels. They were written in our part of the world, in the blood and sweat and shattered dreams of nations like ours. It would be the ultimate, tragic betrayal of our own history if Indonesia—and every Global South state now reaching for the ring of AI sovereignty—were to forget the very lessons our suffering inscribed. Sovereignty is not just about owning the factory; it is about ensuring the factory does not become a new prison. The fight for technological self-determination must be won not only against external imperial forces but also against the internal ghosts of poor governance that have haunted our development for generations. The future of the Global South depends on building walls stronger than the temptations of the past.