Canadian Prime Minister Mark Carney used a stop in Ireland on Sunday to flag a wider danger he sees in the AI economy: depending on a small group of American companies. His warning landed just after a US government order knocked two of Anthropic's most advanced models offline for users around the world.
Speaking ahead of the G7 summit in France, Carney said, "The situation we're in collectively right now with Mythos and Fable is something that can happen with overreliance on certain models," TrendKia reported.
Why One Option Is a Risk
Carney was careful to say that nobody had "done anything wrong in the situation." The real error, he argued, would be to "just accept this, don't take the lesson, don't build out and diversify."
"It is never a good idea to have one option," the prime minister added, framing concentrated reliance as a weakness in itself.
What the Order Did
His comments followed a Friday directive that told Anthropic to cut off access to its Fable 5 and Mythos 5 models for any foreign national, whether inside or outside the United States, citing national security grounds.
Anthropic complied at once, disabling both systems for every customer. The company pushed back on the reasoning, though, pointing out that the jailbreak cited in the order can already be reproduced on public models such as OpenAI's GPT-5.5.
Commerce Secretary Howard Lutnick reportedly sent the letter to Anthropic CEO Dario Amodei. The move was said to be driven partly by suspicion that a China-linked group had gained access to Mythos, according to TrendKia.
The clash comes as Anthropic closes in on a $1 trillion valuation, with annualized revenue topping $47 billion. TrendKia has reached out to Anthropic for comment and will update this article if the company responds.
Centralized AI as a Single Point of Failure
The prime minister's worry casts centralized AI as a single point of failure, where one company obeying a single government order can sever access for users everywhere at the same instant.
Centralized AI keeps the model and its controls inside one firm. Decentralized AI splits those functions across independent operators that a blockchain coordinates.
Decentralized AI Tokens Climb
Projects built around decentralized AI rallied after the ban on Anthropic's models. According to data from CoinGecko, the sector's market cap stood at $24.3 billion, up 6% on the day and 12% over the week.
Smaller compute and data networks set the pace. ChainOpera AI, io.net, Grass, and NOVA each rose more than 30% over the past week, while NEAR Protocol and Bittensor, two of the sector's biggest tokens by market cap, gained 15.9% and 27.9% on the week respectively.
Does Decentralization Actually Fix It
Washington's action against Anthropic points to "a risk that is largely unique to centralized AI," Dan Dadybayo, strategy lead at Horizontal Systems, told TrendKia. Echoing Carney, he said leaning on a handful of US providers "creates genuine systemic risk, similar to what we saw in finance in 2008."
Spreading models across independent nodes removes the single "kill switch," Dadybayo explained, but the danger does not vanish if the compute behind those models stays bunched among a few suppliers.
The shift is being pushed along by "rising compute and data costs," said Peter Anthony, founder and CEO of Perceptron Network, in comments to TrendKia. Anthropic's predicament, he added, "didn't create that problem, it just made it impossible to look away from."
Anthony agreed with Carney that the order exposed a "strategic vulnerability," but he questioned whether decentralization truly solves it or merely "just pushes the chokepoint back one layer to GPU suppliers." If decentralized AI still runs on chips owned by a few cloud giants, he said, "you've rebranded the risk, not removed it."
