OpenAI announced Friday it is limiting the rollout of its new artificial intelligence models — GPT-5.6 Sol, Terra and Luna — to a small group of "trusted partners," complying with a request from the U.S. government. This move effectively grants a select few corporate entities early access to advanced technological tools, while the company simultaneously works with the Trump administration to establish a framework for future model assessments. The decision comes as rival Anthropic reported a staggering run-rate revenue of $47 billion in May 2026, even as it was forced to disable access to two of its latest models about two weeks ago due to an export control directive from the same administration.
State Power Secures Capital Advantage
The Trump administration has adopted a "more hands-on approach" to AI regulation since President Donald Trump signed an AI executive order earlier this month. This order, while "thin on specific details," requested AI developers to "voluntarily" allow government assessment of model capabilities prior to public release. OpenAI stated it believes in "broad access" and is working to make its models generally available in the coming weeks, but conceded, "We don't believe this kind of government access process should become the long-term default." The company argued that the current process "keeps the best tools from users, developers, enterprises, cyber defenders and global partners who need them," yet it did not disclose the names of the "trusted partners" now benefiting from exclusive access. This state intervention ensures that the most powerful AI tools are first leveraged by entities deemed strategically important by the ruling class, solidifying their competitive advantage.
Anthropic remains in active negotiations with officials in Washington, D.C., with no clear timeline for its models to come back online. The U.S. government's ban on global access to Anthropic's Mythos and Fable models has created a vacuum, which Asian capital is rapidly moving to fill.
Global Capital Shifts Amidst Controls
In the wake of the U.S. export controls, two new Asian model products have emerged. Chinese cybersecurity firm 360 reportedly unveiled Tulongfeng, an AI tool it claims rivals Anthropic's Mythos. This cybersecurity-focused AI model, along with its more restricted version, Fable 5, has reportedly been banned by the Trump administration from non-Americans. Earlier the same week, Sakana AI, a Tokyo-based AI startup, launched Fugu, a frontier AI model designed for agents and capable of orchestrating access to other models through their APIs. Sakana AI explicitly advertises "delivering frontier capability without the risk of export controls," directly capitalizing on the U.S. government's restrictions on its domestic competitors.
Sakana co-founder Ren Ito, in an op-ed published last week, urged the U.S. federal government to prioritize "preserving access for America's closest allies," arguing that AI "should not become a technology that is hoarded; it should be one that is developed together." This liberal framing seeks to manage the distribution of technological power among allied capitalist states, rather than challenging the fundamental private ownership and control of such critical infrastructure. David Ha, co-founder and CEO of Sakana, described Fugu as more than a "land grab," emphasizing its design for coordinating agent usage among many models. He warned on X that "relying on a single provider for national infrastructure is a risk the recent export controls made impossible to ignore," and that "access to top models can disappear overnight." Ha advocated for "collective intelligence" as a "practical hedge against this concentration of power," a technical solution that does not address the underlying class dynamics driving the concentration of wealth and control.
Sakana, founded in 2023 by former Google researchers, develops affordable generative AI models optimized for the Japanese language and culture. The company is targeting Fugu at Japanese businesses and government agencies seeking to reduce their exposure to tightening export controls. While Sakana co-founder Ren Ito stated that "U.S. models remain important to Asia" and characterized the moment as a temporary shift rather than a "permanent realignment," the rapid emergence of local alternatives trained to understand local language and nuance is already filling the gap left by U.S. companies. The immense profits generated by companies like Anthropic, with its $47 billion run-rate revenue, underscore the vast sums of capital at stake in this global technological competition, where state power is deployed to secure and advance the interests of national capital.