Anthropic Seeks Gulf Funds for AI

Anthropic CEO Dario Amodei reveals the company’s reluctant decision to pursue Gulf State funding, admitting ethical compromises are necessary in the AI arms race. The move underscores the immense capital demands forcing firms to prioritize growth over conscience.

A veil of corporate discretion has been dramatically pulled back, revealing the stark realities of the global AI arms race and the moral compromises being weighed at the highest echelons of Silicon Valley.

In a candid, almost confessional, internal memo obtained by WIRED, Anthropic CEO Dario Amodei laid bare the company’s reluctant, yet seemingly inevitable, decision to pursue significant investment from Gulf States, specifically the United Arab Emirates and Qatar.

It is a pivot that underscores the immense financial pressures shaping the future of artificial intelligence, forcing even the most ethically-minded players to confront a landscape where capital often trumps conscience.

Amodei’s note to staff, dispatched on a Sunday morning, was a striking admission of a principle abandoned, or at least, severely bent.

“Unfortunately, I think ‘No bad person should ever benefit from our success’ is a pretty difficult principle to run a business on,” he wrote, acknowledging that accepting funds from Middle Eastern leaders would likely enrich “dictators.”

The sentiment, while jarringly blunt, captures the profound dilemma facing companies at the bleeding edge of AI development: the insatiable demand for capital to train increasingly sophisticated frontier models.

These models require computing power on a scale that few entities can finance, and the Gulf States, flush with petrodollars, represent an undeniable, if ethically complex, reservoir of wealth.

This isn’t merely a strategic business decision; it’s a concession to an emergent global reality.

Amodei himself bemoaned the “race to the bottom” that has seen other major AI players leap headfirst into partnerships with authoritarian regimes.

He pointed to the widespread trend of companies “outsourcing our largest 5 GW training runs to UAE/Saudi,” a dynamic he had previously “vociferously pushed for not allowing.”

The CEO’s lament that Anthropic, having failed to prevent this collective erosion of standards, is now “stuck with it as an individual company,” paints a picture of a firm forced to play a game it never wanted to join, lest it be left behind.

The context is critical: OpenAI, Anthropic’s formidable rival, has already forged deep ties in the region, announcing a staggering $500 billion data center project, Stargate, with Emirati backing, and plans for a sovereign AI capability push in Abu Dhabi.

When the competition is building data centers powered by gigawatts of electricity in the very regions you’ve deemed problematic, the pressure to conform becomes immense.

The irony of the situation is not lost on Amodei, who conceded that the move would invite accusations of hypocrisy.

In his essay, “Machines of Loving Grace,” he had previously championed the idea that “Democracies need to be able to set the terms by which powerful AI is brought into the world, both to avoid being overpowered by authoritarians and to prevent human rights abuses within authoritarian countries.”

Yet, the pursuit of Gulf State investment directly challenges this stated ideal.

This isn’t Anthropic’s first dance with Middle Eastern money; in 2024, when the failed cryptocurrency exchange FTX’s stake in Anthropic went up for sale, a significant portion was acquired by ATIC Third International Investment, a UAE firm.

However, the current shift signifies a more direct, proactive pursuit of state-level capital, moving beyond passive investment acquisition.

While Anthropic had previously declined Saudi Arabian investment due to national security concerns, the memo suggests a broader opening to Gulf capital, without specifying if its stance on Saudi Arabia has softened.

Amodei’s internal message also offered a fascinating glimpse into the internal calculus of risk.

He expressed hope that a “narrowly scoped, purely financial investment” could avoid granting outside investors “leverage” over the company.

Yet, he immediately acknowledged the inherent danger of “soft power“—the implicit promise of future funding that could make it “harder to resist these things in the future.”

This is the Faustian bargain of frontier AI: the capital is necessary for survival and advancement, but it comes tethered to geopolitical influence and potential ethical compromises.

The sheer scale of the capital involved—”easily $100B or more” in the Middle East, as Amodei noted—makes it an almost irresistible gravitational pull for companies burning billions to develop the next generation of intelligence.

Perhaps most revealing were Amodei’s candid remarks about the anticipated public backlash, which he dismissed as “Comms Headaches.”

The “media / Twitter / the outside world” is “always looking for hypocrisy, while also being very stupid and therefore having a poor understanding of substantive issues,” he wrote.

This dismissive tone, particularly concerning the media’s role in scrutinizing corporate ethics, is striking from a CEO who claims to be making a difficult but “right” decision.

His justification—that it’s “perfectly consistent to advocate for a policy of ‘No one is allowed to do x,’ but then if that policy fails and everyone else does X, to reluctantly do x ourselves”—is a cynical, yet perhaps brutally honest, summation of how competitive pressures can erode even deeply held principles.

Ultimately, Anthropic’s decision, born out of necessity in a hyper-competitive landscape, highlights a critical juncture for the entire AI industry.

It’s a moment where the pursuit of technological supremacy is colliding head-on with ethical considerations, where the grand pronouncements of responsible AI development are being tested by the relentless demands of the market.

As AI continues its inexorable march forward, the question remains: at what cost, and with whose money, will the future of intelligence be built?

Amodei believes it’s “the right one overall,” but the full implications of this strategic surrender to capital, however reluctant, are yet to be seen.