Major AI labs expand infrastructure and security capabilities while the White House drafts an executive order to regulate model releases through an FDA-style approval process.
The landscape of American digital sovereignty shifted rapidly on May 8, 2026, as the leading AI laboratories engaged in a high-stakes arms race for infrastructure and security dominance. Anthropic announced a doubling of rate limits for its Claude Code platform, fueled by a strategic partnership with SpaceX. By utilizing the Colossus One data center—a facility boasting 300 megawatts of capacity and over 220,000 GPUs—Anthropic is moving beyond terrestrial limits. The deal reportedly includes long-term plans to develop multiple gigawatts of orbital AI compute capacity, signaling a new frontier for American technological infrastructure.
Simultaneously, OpenAI has launched GPT-5.5-Cyber, internally nicknamed “Spud,” to a select group of vetted security teams. This specialized model aims to achieve parity with Anthropic’s Mythos Preview in vulnerability detection. To facilitate broader commercial adoption, OpenAI also introduced a trio of voice-focused models: GPT-Realtime-2, GPT-Realtime-Whisper, and GPT-Realtime-Translate. These releases offer granular, minute-based billing designed to compete directly with established voice AI providers.
In the private markets, ElevenLabs has demonstrated the immense scale of the generative voice sector. The company’s Series D funding has now surpassed $550 million, attracting high-profile cultural figures like Jamie Foxx and Eva Longoria. ElevenLabs reported a staggering $100 million in net new annual recurring revenue in the first quarter of 2026 alone, bringing its total ARR above the $500 million mark. This financial momentum was accompanied by a significant price reduction across its API services to maintain its market lead against OpenAI.
However, this rapid private-sector expansion is meeting increased friction from federal and state regulators. National Economic Council Director Kevin Hassett confirmed that the White House is drafting an executive order to mandate the vetting of AI models before they reach the public, a move that critics argue mirrors the slow-moving FDA approval process. This policy shift follows reports of AI models autonomously identifying zero-day vulnerabilities, raising concerns regarding national security and dual-use risks.
State-level litigation is also intensifying. Pennsylvania has filed a lawsuit against Character.AI, alleging the platform allowed AI personas to impersonate medical professionals. As the federal government weighs a centralized vetting regime, states like Connecticut and Iowa are moving forward with their own safety laws, creating a complex patchwork of regulations that could challenge the free-market principles that have historically driven Silicon Valley’s global leadership.

