An analysis says the regulatory battle over open-source artificial intelligence is starting to resemble the early Bitcoin market around 2014.

Blockchain outlet Bitcoin Magazine reported on Tuesday that U.S. investment research firm Brownstone Research said in a recent report that a push to tighten controls on open-source AI is following a similar path to the pressure once faced by the cryptocurrency industry.

The core dynamic is to highlight the risks of open-source models while strengthening the justification for closed models. The report cited comments made in July 2023 by Anthropic CEO Dario Amodei in the U.S. Congress as a starting point. Amodei said open source is positive in most scientific fields, but the risks of open models released so far were relatively limited. He warned, however, that the direction in which open-source models are scaling is heading down a very dangerous path.

Brownstone Research said such remarks could lead to a policy structure that restricts open-source models and promotes closed models as a relatively safe option. It also said this frame is a familiar scene for the digital asset market.

It also cited early Bitcoin examples. They ranged from a case involving U.S. Representative Jared Polis, who bought the first Bitcoin in the U.S. Congress in 2014, to Senator Joe Manchin urging a ban, calling Bitcoin a dangerous currency. It also grouped into the same trend the 2023 controversy over what was dubbed "Operation Choke Point 2.0," in which regulators were said to have tried to cut cryptocurrencies off from the banking system.

The cryptocurrency industry, however, survived such pressure. The report also pointed to the U.S. Congress now moving toward clearer rulemaking through passage of the GENIUS Act and efforts to advance the CLARITY Act. Brownstone Research said decentralised AI has also entered a similar phase of conflict.

As a recent case, the report cited a trend toward tighter access controls. It said U.S. export restrictions limiting deployment of Anthropic's latest model have increased the likelihood of a shift toward a permissioned structure that allows access only after verifying a user's identity. It also said OpenAI has limited the rollout of GPT-5.6 to trusted partners, and it saw such measures potentially leading to broader permissioned access based on identity checks.

It also cited national security concerns as a background factor. Joshua Rudd, an official responsible at the U.S. National Security Agency, was reported to have explained through Senator Mark Warner that Anthropic's "Mythos" model could penetrate nearly all classified systems in hours, not weeks.

It also said the technology gap for the open-source camp is narrowing quickly. Brownstone Research said GLM-5.2 recently posted performance comparable to Anthropic's Sonnet4.6 as of February. It said open models have now caught up to within about 3 to 4 months of frontier models, and forecast that an open competitive model responding to Mythos and GPT-5.6 could emerge this autumn.

As a key driver of decentralised AI, the report cited a network-based learning structure. Brownstone Research explained that the use of peer-to-peer networks, like Bitcoin and Ethereum, to pool computing resources and use them for model training is spreading. It said the scale of distributed training has grown over two years from under 1 billion parameters to about 100 billion.

It cited early projects including Dark Bloom, which supports low-cost private inference on idle Mac computers; c0mpute, a decentralised inference network; and Pluralis, which trains AI by connecting consumer GPUs in a distributed way. It forecast that more projects will issue tokens and adopt structures that reward providers of computing resources.

The conclusion is clear. The report said that even if regulation of open-source AI tightens, the spread of decentralised AI is unlikely to be easily halted. Regulatory uncertainty and high volatility remain, it said, but the dynamic could create long-term growth potential, as in the early Bitcoin market.