Solana Founder Yakovenko Mocks AI Slowdown Push as Trillion-Dollar Incumbents Ask the Industry to Wait

Yakovenko reduces a trillion-dollar AI debate to four words

Solana co-founder Anatoly Yakovenko has waded into the most consequential artificial intelligence policy fight of the year with a single, biting remark. As frontier lab CEOs lined up behind proposals to slow the pace of model development, Yakovenko posted a four-word response on X that framed the entire debate around money: “Profitability at [one-trillion-USD] mcap.”

The comment, published on Sept. 13, was aimed at a proposal from Anthropic chief executive Dario Amodei, who released an essay calling on frontier AI companies to deliberately slow the development of their most powerful models. OpenAI CEO Sam Altman endorsed part of the plan, while Elon Musk told Reuters simply that “Dario is right.” Yakovenko, whose Solana network has staked much of its future on high-throughput AI and DeFi applications, did not name a specific company or publish calculations behind the figure. He also did not present evidence that the endorsements were coordinated for financial reasons.

The remark lands at an uncomfortable moment for the AI industry. Anthropic has been reported to be seeking one of the largest capital raises in corporate history, with investors discussing valuations approaching two trillion USD. For crypto natives who remember exchange insiders urging restraint while accumulating positions, the optics of trillion-dollar incumbents asking everyone to slow down are easy to read.

What Amodei is actually proposing

The essay at the center of the controversy is more detailed than a simple call for restraint. Amodei proposes a three-stage framework that would embed outside evaluators inside frontier laboratories, coordinate safety standards across the industry, and eventually pursue narrow international agreements on specific risks such as cyber threats and model-control failures.

Anthropic has committed to inviting an external review team into its internal safety process, with reviewers receiving company laptops, office access and permissions comparable to internal risk staff. The plan would allow those reviewers to publish findings without Anthropic controlling the conclusions, subject to limited redactions for privileged or security-sensitive material. Altman said OpenAI would adopt a similar system and promised further details “soon,” though no evaluator contract or timeline has been announced.

The essay cites concrete concerns: model-control failures, escalating cyber risk, and the accelerating speed at which AI systems help build newer AI systems. These are arguments about safety, not valuations, and Amodei’s supporters note that slower release cadences impose real competitive costs on the companies proposing them.

Sacks rejects the coordination framework

Yakovenko was not the only tech figure pushing back. Former White House crypto and AI adviser David Sacks posted his own critique, telling Amodei and Altman directly: “You guys are the frontier.” Sacks argued that the two companies hold a “duopoly on frontier intelligence” and could simply slow their own development voluntarily without new regulatory machinery binding their competitors.

Sacks said he would support a voluntary slowdown if unreleased systems presented risks serious enough to concern their own developers. What he disputed was the need for antitrust exemptions, regulatory approval systems, or evaluators with authority over rival laboratories. His post accused the companies of pursuing “regulatory capture,” though it offered no evidence of an agreement between Anthropic and OpenAI to exclude competitors, and neither company has accepted that characterization.

Sacks also raised product liability as an underappreciated incentive, arguing that customers already punish models that behave unpredictably. He questioned whether China would participate in any global AI agreement, a verification problem that Amodei’s own essay acknowledges while still defending narrowly scoped negotiations.

Why crypto watches the AI pacing fight

The debate is not abstract for the altcoin market. Solana traded near 101 USD on Sept. 13, and its ecosystem has become a primary venue for AI-agent tokens, decentralized compute projects and machine-to-machine payment experiments. Ethereum traded near 2,525 USD and Bitcoin near 77,200 USD in the same session, with the broader Fear and Greed Index at 63, signaling greed without the extreme readings seen earlier in the week.

AI narratives have been one of the strongest performance drivers in the altcoin sector over the past two years, and any regulatory architecture that concentrates frontier model access inside a handful of incumbent labs would flow directly into token valuations. A slowdown that raises the cost of frontier capabilities could advantage decentralized compute networks, or it could starve agent-based crypto applications of the model quality they depend on. Both outcomes were being debated across crypto social media within hours of Yakovenko’s post.

There is also a familiar historical echo. Crypto veterans recall the 2024-era pattern of incumbent exchanges endorsing aggressive regulation that smaller rivals could not afford to implement. Whether or not Yakovenko’s insinuation is fair, the fact that it resonates reflects how little trust the crypto industry extends toward trillion-dollar AI incumbents asking for restraint.

What remains unresolved

No named company has confirmed that support for slower development depends on reaching a trillion-dollar valuation. Musk endorsed Amodei’s essay without specifying which parts of the three-stage plan xAI would adopt. Altman’s evaluator commitment has no published timetable, and Anthropic has not named its review team or a start date.

Lawmakers may settle the question before the industry does. A Sanders-Casar proposal in the United States seeks a permanent ban on superintelligent systems and a temporary suspension of advanced AI work until a new federal regulator writes safety rules. A Bank for International Settlements paper separately warned that AI could compress software-patching windows from weeks to minutes, raising the stakes for infrastructure operators, including those securing blockchain networks.

For now, Yakovenko’s four words remain exactly that, a characterization of motives rather than a documented claim. But the exchange marks the moment the AI pacing debate officially crossed into crypto’s territory, with one of the industry’s most prominent founders accusing the world’s most valuable startups of wanting to pause the race only after reaching the finish line.