Analysis
OpenAI and Anthropic both increased their Washington lobbying spend in the second quarter of 2026, CNBC reported on July 21, even as legacy technology companies and defense-adjacent contractors pulled back their own political spending over the same period. The divergence is the real story: it signals a shift in where regulatory and political risk is concentrated in the tech sector, away from established incumbents and toward the frontier AI labs whose products are now central to policy debates ranging from chip export controls to copyright liability to national security.
Both companies have accumulating reasons to invest more heavily in Washington relationships. Anthropic just closed a $1.5 billion copyright settlement and is reportedly pursuing a confidential IPO near a $965 billion valuation -- exactly the moment a company wants friendly, informed relationships with regulators rather than surprises. OpenAI, meanwhile, is navigating its own board changes, an internal power struggle reportedly involving its hardware venture with Jony Ive, and renewed scrutiny after this week's disclosed model security incident on Hugging Face's infrastructure.
“Both companies have accumulating reasons to invest more heavily in Washington relationships.”
The broader regulatory backdrop this quarter has been unusually active: Treasury Secretary Bessent floated potential sanctions against China over AI model 'theft,' the Trump administration's push for American-made AI chips is reportedly squeezing TSMC's margins, and there's ongoing debate -- including OpenAI's own policy team reportedly arguing, then partially retracting, a case for restricting open-weight models -- over how aggressively to regulate frontier AI capability versus open competition.
Compared to prior lobbying cycles, frontier AI labs spending more than legacy tech and defense contractors marks a real shift in Washington's attention. Companies like Google, Meta and Amazon have run large, mature lobbying operations for over a decade; OpenAI and Anthropic building up comparable spend within just a few years of each becoming policy-relevant reflects how fast AI has become a first-order regulatory and national-security topic rather than a niche tech-policy issue.
For founders and operators in AI-adjacent regulated industries, rising lobbying investment from the two most prominent US labs is a signal that policy outcomes -- export controls, safety mandates, copyright frameworks -- are going to be actively contested and shaped over the next several quarters, not settled passively. Any startup whose business model depends on regulatory clarity in these areas should expect continued volatility.
