OpenAI announced new artificial intelligence models on Friday, stating its compliance with the U.S. government's request to initially limit their rollout to a small group of trusted partners. However, the company explicitly declared, "We don't believe this kind of government access process should become the long-term default," signaling a clear corporate intent to undermine national sovereignty over critical technology.
This limited release of GPT-5.6 Sol, Terra, and Luna follows a noticeable increase in the Trump administration's hands-on approach to AI regulation. President Donald Trump signed an AI executive order earlier this month, asking developers to voluntarily allow government assessment of models before full release.
Two weeks ago, rival firm Anthropic was forced to disable access to two of its latest models, Mythos and Fable 5, in compliance with an export control directive from the Trump administration. Anthropic is currently in negotiations with officials in Washington, D.C., regarding the ban.
OpenAI confirmed it previewed its new models' capabilities and shared plans with the government ahead of Friday's launch. The company also stated it is working with the Trump administration to establish a framework for such assessments and to develop a repeatable process for future model releases, framing this as a "short-term step" toward broader availability and resisting long-term national oversight.
Elite Interests Challenge National Control
This assertion of national control faces direct opposition from transnational elite interests. Ren Ito, co-founder of Tokyo-based Sakana AI, published an op-ed in Project Syndicate last week, urging the U.S. federal government to prioritize access for "America's closest allies." Ito argued that AI should not be a technology that is "hoarded," but rather one that is "developed together," advocating for a globalist approach to a technology with profound national security implications.
David Ha, another co-founder and CEO of Sakana AI, echoed this sentiment, arguing against relying on a single provider for "national infrastructure." On X, Ha wrote, "Access to top models can disappear overnight," and promoted "Collective intelligence" as a hedge against the concentration of power, implicitly advocating for a distributed, post-national control structure for critical AI.
OpenAI itself stated its belief in "broad access" and expressed concern that the government access process keeps the best tools from "users, developers, enterprises, cyber defenders and global partners who need them." This highlights the transnational corporate view that national borders and sovereign controls are obstacles to their preferred model of global distribution.
Nations Reassert Sovereignty
While U.S. tech giants navigate these new national controls, other nations are moving to fill the void and assert their own technological sovereignty. Chinese cybersecurity firm 360 reportedly unveiled Tulongfeng, an AI tool capable of competing with Anthropic's Mythos. The Trump administration has reportedly banned Tulongfeng and its restricted version, Fable 5, from non-Americans, demonstrating a national effort to control the proliferation of powerful AI.
Sakana AI, co-founded in 2023 by former Google researchers, launched its Fugu model, named after the Japanese word for blowfish. Sakana has been capitalizing on the U.S. export ban, advertising "frontier capability without the risk of export controls" on its website. The company aims Fugu at Japanese businesses and government agencies seeking to reduce their exposure to tightening export controls.
Sakana AI specifically optimizes its generative AI models for the "Japanese language and culture," and they are designed to work well with small datasets. This focus on national cultural and linguistic specificity stands in stark contrast to the generic, borderless approach favored by many globalist tech entities.
Anthropic, which had been on a historic growth trajectory, reported run-rate revenue crossing $47 billion in May 2026. The extent to which this revenue depends on Asian enterprise customers is not publicly known. However, in the weeks since the U.S. export order took effect, at least two companies, one in Tokyo and one in Beijing, have stepped into the space left behind, offering local alternatives that are trained to better understand local language and nuance, demonstrating a global shift towards national-focused AI development in response to U.S. policy.