Anthropic shutdown fuels interest in decentralized AI

Anthropic disabled access to its Fable 5 and Mythos 5 models after the U.S. government on Friday ordered the company to suspend model access for foreign nationals over national security concerns. Anthropic turned off the models for all users to comply with the order.

Grayscale’s head of research, Zach Pandl, wrote that the shutdown highlighted the risks of centralized control of advanced AI and increased interest in decentralized alternatives. In a note, he wrote that centralized control ‘drives home the need for decentralized alternatives.’

Pandl identified Bittensor as an example of a decentralized approach. He noted that in the 12 hours after Anthropic disabled the models, Bittensor’s TAO token climbed about 30% to roughly $283, a three-week high, and that TAO outperformed the broader crypto market over the past week.

Pandl described Bittensor as an ‘alternative vision for AI based on decentralized principles’ and added, ‘Think of it as Bitcoin for AI.’ He also wrote that access to artificial intelligence is becoming an economic resource and that governments and AI labs will increasingly determine who can use these tools and under what conditions.

Industry figures offered reactions to the shutdown. Colton Malkerson, co-founder of EdgeRunner AI, compared reliance on external AI providers to renting, writing that companies risk losing access and control if providers change terms or are compelled to cut service. Tech entrepreneur Brett Hurt called the U.S. order a precedent, saying a government can silence a commercial AI model overnight without public hearings or an appeals process.

Bittensor operates as a decentralized network that lets participants contribute and access AI resources through an open protocol and uses the TAO token to reward contributors. Supporters say such networks can provide redundancy and broader access if centralized labs restrict service. Critics point to challenges with performance, governance and coordination in decentralized systems.

Regulators have increased scrutiny of leading AI providers and have pressed companies for controls on access and export. Some labs may respond by narrowing user bases or applying geographic restrictions. Investors and developers tracking those regulatory and commercial shifts have shown interest in projects that offer more distributed access or less centralized governance.