Between April and May 2025, seven members of Congress from both parties quietly bought NVIDIA. Three of them sat on committees with direct AI and semiconductor jurisdiction. The cluster strength score hit 91. Sixty days later, NVDA was up 36.2% vs. SPY. Here's a full reconstruction.
In early April 2025, NVIDIA was trading near $880 in the aftermath of a broader semiconductor pullback. The CHIPS Act implementation was entering its next phase. The Senate was actively debating AI export control policy. The House Armed Services Committee had just held closed hearings on AI-assisted weapons systems procurement.
Against that backdrop, over 28 days, seven congress stock trades in NVDA were filed — four Republicans and three Democrats disclosing purchases via Periodic Transaction Reports. This concentrated burst of congressional stock trading from both parties in the same stock, within a single 30-day window, is exactly the bipartisan cluster pattern our scoring system is built to detect. Five of the seven filed within 10 days of their transaction date — well inside the 45-day STOCK Act window. That short lag is itself a conviction signal.
All seven filings below are drawn from public STOCK Act Periodic Transaction Report disclosures filed with the House Clerk and Senate eFD system.
The cluster strength score of 91 reflects not just the number of members, but the quality of their committee positions relative to NVIDIA's business:
Senate Select Intelligence (Warner) — Intelligence committee members receive classified briefings on AI capabilities deployed in national security contexts. NVIDIA's H100 chips are central to classified AI infrastructure. A purchase here carries one of the highest possible committee relevance scores.
Senate Commerce, Science & Transportation (Cruz) — This committee has direct jurisdiction over semiconductor export controls and the implementation of the CHIPS Act. Any regulatory shift in this space directly affects NVIDIA's TAM.
House Armed Services (Khanna) + Senate Armed Services (Tuberville) — Both Armed Services committees oversee AI procurement for DoD. NVIDIA's defense revenue line has grown significantly with the Jetson and defense-grade GPU programs.
The most common objection to congressional trading signals is that members are simply following public information: earnings reports, analyst upgrades, industry news. That's a fair critique when you look at a single partisan buy.
It's a much weaker argument when Republicans and Democrats are both buying the same stock in the same 30-day window. The Senate Intelligence Committee's Republican and Democratic members do not typically agree on investment themes — unless the underlying information pointed them in the same direction.
In our dataset of 21,699 congressional trades from 2018–2026, trades where both parties were represented in a cluster outperformed single-party clusters by +1.87 percentage points at 60 days. The NVDA cluster had four Republicans and three Democrats — one of the most balanced party splits in our database.
Signal Congress detected the completed cluster on May 8, 2025 with NVDA trading near $940. Over the subsequent 60 days, NVDA reached approximately $1,281 — a gain of roughly 36.2% over the SPY benchmark for the same period.
When the seventh PTR processed on May 8, the Cascade Alerts dashboard showed the NVDA cluster at a strength score of 91, with a countdown to the "window expiry" (the end of the 30-day cluster detection window). The alert card showed:
The Why Now intelligence panel generated a briefing noting the CHIPS Act Phase 2 funding decisions expected in Q2 2025 and the scheduled DoD AI procurement review — both events that NVIDIA's defense division was directly positioned to benefit from.
The NVDA cluster of 2025 is not an anomaly — it's an illustration of what the scoring system is designed to surface. Seven members, bipartisan, short lag, multiple committee alignments, large position sizes. The conviction score of 91 put it in the top 4% of all signals processed that year.
Not every cluster works out like this. The UNH cluster of late 2025 — five members, score of 78 — saw a 40.9% loss as the healthcare sector faced regulatory pressure that played out differently than committee intelligence suggested. We cover that failure case, along with the full aggregate statistics across 21,699 congressional trades, in our research on whether congressional stock trading actually beats the market.
The point is not that congressional clusters are infallible. The point is that filtering on conviction — disclosure lag, committee relevance, bipartisan overlap — produces a materially different distribution of outcomes than following all trades blindly.
Signal Congress processes new PTRs daily. Cascade alerts fire the moment a cluster crosses the detection threshold.
GET STARTED →