OpenAI and Anthropic Just Turned 'Poaching Washington' Into a New Kind of AI Moat

OpenAI has hired at least three former Trump officials and Anthropic at least two this year, according to filings and the officials' own announcement posts. The hires are concentrated in the policy and infrastructure functions that decide which lab wins federal contracts, defense work, and favorable regulatory treatment.

Anthropic is now staffing the next phase: a Political Programs Lead paying $295–345K to build and run a \"presidential engagement\" program for the 2028 cycle, with the explicit charter of mapping both parties' prospective fields and offering AI policy guidance to candidates.

Personnel poaching only matters because it sits on top of a lobbying budget that has more than doubled in twelve months. In Q2 2026, OpenAI spent $1.2 million and Anthropic spent $1.97 million on federal lobbying — a combined $3.17 million, up 23% quarter-on-quarter and a record for both companies, per disclosures tracked by CNBC and OpenSecrets. Anthropic's 2026 run-rate has nearly tripled from its 2025 full-year total of $3.13 million.

For context, the eight largest defense primes combined spent $19.68M in Q2 2026 — Anthropic alone now spends more than one-fifth of that total from a private, pre-revenue-mature balance sheet, and OpenAI is within $50,000 of Nvidia despite having no hardware revenue to defend.

The poaching and lobbying surge is not opportunistic — it is reactive. Between February and September 2026, the federal government did more to reshape the competitive position of OpenAI and Anthropic than in the prior three years combined. The labs needed people who already knew how the levers move.

Both labs have crossed the S-1 threshold. OpenAI filed a confidential registration statement on May 22, 2026, confirmed Jun 8; Anthropic filed on Jun 1, 2026, targeting a listing as early as October. Three lines on those prospectuses look materially different in light of the September–October personnel moves.

The headline numbers support a steeper valuation, but the dispersion is now wider. OpenAI's annualized revenue approached $70 billion by end-September 2026, up from $20 billion at year-end 2025, on a $852 billion private valuation. Anthropic's run-rate crossed $47 billion in May and roughly $65 billion by late July, with leaked S-1 financials showing $4.6 billion of 2025 revenue, $11.5 billion in Q2 2026, and a $42 billion 2025 net loss. At a rumored $965 billion to $2 trillion post-money valuation, every percentage point of regulatory discount is worth $10–20 billion.

The talent-and-lobbying surge ricochets through the listed AI ecosystem in three directions: the hyperscalers underwriting the labs, the chip designers building the compute, and the defense and cybersecurity vendors already inside the procurement pipeline.

Two horizons diverge sharply. The next two quarters are about filings, court rulings, and political theater. The 1–3 year arc is about which lab institutionalizes the personnel pipeline — not the next five hires.

The structural risk is the reverse: a future administration could weaponize the same personnel pipeline it ignored this cycle. A small-to-midsize bench, currently a moat, could become a target under a different White House. That asymmetry is the single most important line the S-1 risk-factor section will be forced to disclose.