To facilitate the export of United States artificial intelligence systems, computing hardware, and standards globally.

HR6996, the Full AI Stack Export Promotion Act, was introduced January 9, 2026, by Rep. Randy Fine (R-FL), reported out of the House Committee on Foreign Affairs on April 22, 2026, by a 37-7 vote, and now awaits floor action. The bill explicitly finds that the U.S. is in an AI race against China and establishes a policy to 'maintain United States dominance' by facilitating export of U.S. AI models, cloud services, data centers, and semiconductors to allies. The mechanism is a consortia program within the Department of Commerce that identifies security-approved industry groups to coordinate export licensing and compliance, effectively creating a fast-track system for approved allied nations to procure the full U.S. AI stack.

The bill does NOT authorize or appropriate any funds — it is a pure policy and regulatory shift. The Commerce Department's existing export licensing infrastructure (BIS) would implement the consortia program using existing appropriations. The money trail is indirect: lower regulatory friction → increased volume of high-margin AI hardware sales (NVDA GPUs, AMD Instinct), higher cloud consumption revenue (MSFT Azure, AMZN AWS, GOOGL GCP), and more AI server system contracts (SMCI). This is not a spending bill; it's a market access expansion bill.

Structural winners are clear: AI chip designers (NVDA, AMD) gain incremental sovereign and enterprise demand as allies build national AI infrastructure without having to navigate cumbersome individual export license applications. AI server integrators (SMCI) benefit as the 'full stack' includes complete rack-scale systems. Cloud hyperscalers (MSFT, AMZN, GOOGL) win because the bill explicitly mandates that allied AI deployments run on U.S. cloud operators using U.S. data centers. Intel (INTC) is included in the 'U.S. designed semiconductors' bucket via Gaudi AI accelerators and Xeon CPUs for AI inference — but its current AI hardware market share is negligible vs NVDA/AMD, so the benefit is structural but delayed. Companies NOT mentioned but structurally exposed include Dell (DELL) and Hewlett Packard Enterprise (HPE) as AI server OEMs — they could benefit but the causal chain is weaker.

Real market data corroborates the bullish AI infrastructure narrative: NVDA is at $209.25, up 26.69% in 30 days and near its 52-week high of $216.83. AMD has surged 71.96% in 30 days to $337.11, reflecting market pricing in expanded AI compute TAM beyond NVDA's near-monopoly. Intel's 130% 30-day move to $94.75 is driven more by restructuring and foundry hopes than this bill directly — though the bill's allied export mandate would apply to any future Intel AI products. SMCI has been flat (-1.61% 7-day) at $26.32, suggesting the market is focused on its accounting/audit overhang rather than this policy development. The cloud tickers — MSFT ($424.46), GOOGL ($349.94), AMZN ($263.04) — all show 18-31% monthly gains, reflecting broad AI infrastructure demand.

Timeline: The bill has cleared committee with a strong bipartisan vote (37-7). It now goes to the House floor. With the primary sponsor in the majority party and reported out of committee before midterms (2026 midterms are November), the window for passage is before August recess — likely June-July 2026. A companion bill has not been introduced in the Senate — this is a House-only vehicle currently, which reduces near-term passage probability. If enacted, Commerce implementation of the consortia program would take 6-12 months. The market would discount the benefit immediately on passage, with hardware plays (NVDA, AMD) responding most to export volume expectations.