The AI race is shifting from pure technical competition to a strategic political battle. OpenAI and Anthropic, traditionally rivals, are now coordinating their lobbying efforts in Washington to warn policymakers about the risks associated with open-weight AI models.

The economic threat of openness

While the public narrative focuses on safety and catastrophic risk, the underlying driver is financial survival. The emergence of high-performance open-weight models, such as Moonshot AI's Kimi K3, has proven that the capability gap between closed and open systems is closing rapidly. Kimi K3's frontier-level results have sent a shockwave through the industry, suggesting that proprietary APIs may no longer hold a monopoly on intelligence.

Furthermore, the process of distillation—where smaller models are trained using outputs from larger proprietary ones—is viewed as a direct threat to the bottom line. OpenAI has echoed concerns from Trump administration officials, arguing that this practice allows competitors to "steal" the value of their expensive training runs without incurring the same costs.

Regulatory capture as a strategy

To safeguard their market position, both companies have ramped up spending. In Q2 2026 alone, they spent a combined $3.17 million on lobbying, a 23% increase over the previous quarter. They are no longer competing solely on benchmark scores, but on who can shape the rules of the game.

This effort aligns with broader US government initiatives like Gold Eagle, which centralizes control over frontier model access. While OpenAI seeks a streamlined federal approach, Anthropic is pursuing a state-by-state strategy to ratchet up safety requirements, effectively creating barriers to entry for smaller players.

The IPO pressure and financial fragility

This aggressive lobbying is a symptom of structural instability. Neither OpenAI nor Anthropic is currently self-sufficient, relying heavily on external equity and cloud partners for survival. With potential IPOs targeted for late 2026, the companies must prove to investors that they can maintain high margins despite the rise of open-source alternatives.

If open-weight models continue to capture a larger share of enterprise workloads, the closed-model business strategy may become unsustainable. The battle is now between those who believe in a centralized, regulated AI ecosystem and those, like Meta or DeepSeek, who see open access as the primary driver of global innovation.