# OpenAI abandons China chip curbs as Nvidia protects exports *The software developer dropped advocacy for strict hardware export controls in October disclosures while Nvidia spent a record $2.2 million to preserve Asian market access.* **Published:** October 22, 2026 **By:** Mina Okoro **Section:** Trade & Industrial Policy — Cross-border **Format:** Breaking **Confidence:** high **Source:** https://eastasiabrief.com/trade-policy/openai-abandons-china-chip-curbs-nvidia-protects-exports-857 **Publisher:** East Asia Brief (https://eastasiabrief.com/) --- ## What to know - Nvidia spent record $2.2M on Q3 lobbying while OpenAI dropped China chip advocacy - Hardware alliance includes $30B Nvidia equity stake and $105B data center guarantee - House committee schedules export markup for Nov. 18; Q4 disclosure deadline Jan. 20 ![A technician works on an advanced semiconductor chip using tweezers inside an electronics testing laboratory. (AI-generated image)](https://eastasiabrief.com/media/2026-10-21-0aa57e2e303e.webp) *A technician works on an advanced semiconductor chip using tweezers inside an electronics testing laboratory. (AI-generated image)* OpenAI retreated from its push for stricter semiconductor export controls on China on Oct. 20, 2026, handing chipmaker Nvidia a clearer path to defend its Asian hardware revenue. Congressional aides and trade lobbyists confirmed the strategy reversal following third-quarter disclosures. The retreat removes a prominent domestic technology voice backing federal bans on advanced computing shipments to Chinese commercial buyers. Nvidia spent a record $2.2 million on federal lobbying during the third quarter. That expenditure lifted the chipmaker's lobbying total for the first nine months of the year to $4.7 million. The spending targeted several bills aimed at cutting graphics processing units sent abroad. Those measures include the AI OVERWATCH Act, the Chip Security Act and the Remote Access Security Act. OpenAI filed disclosure forms under the Lobbying Disclosure Act during the same period. The company registered advocacy on federal research frameworks and antitrust exemptions for threat sharing, omitting semiconductor trade limits entirely. The omitted stance marks a sharp reversal from early 2025. OpenAI urged federal regulators to restrict semiconductor shipments to adversary nations in a March 13, 2025 submission to the White House Office of Science and Technology Policy. That previous filing argued that keeping cutting-edge accelerators away from Chinese state developers was vital to preserving American leadership in artificial intelligence. The company proposed strict export controls on Tier III countries, including China. Capitol Hill aides said OpenAI representatives stopped raising semiconductor export controls during legislative consultations this autumn. The silence arrived as commercial ties between the two technology companies deepened significantly. Nvidia finalized a $30 billion equity investment in OpenAI in February. The transaction expanded a financial alliance between the world's most valuable chip vendor and its largest algorithmic software customer. Nvidia also signed an exclusive chip supply contract with OpenAI in August. The chipmaker agreed to guarantee up to $105 billion in financing for the first phase of an OpenAI-leased data center in Ohio. Securing that massive computing volume requires stable silicon manufacturing pipelines. Tightening export controls risks raising compliance costs across the global graphics processor market, where single server nodes exceed $300,000. Lobbyists for hardware makers warned lawmakers that sweeping bans on Chinese sales reduce research budgets. Nvidia relies on international revenue to fund capital-intensive development of its next-generation processor architectures. The Bureau of Industry and Security relaxed license reviews for older accelerator models in January. The shift permitted case-by-case export reviews for specific processors, triggering pushback from national security committees in Congress. Legislators drafted the AI OVERWATCH Act to counter executive licensing discretion. The bill mandates a two-year ban on high-end processor exports to China and grants Congress veto authority over Department of Commerce licenses. OpenAI's withdrawal leaves semiconductor manufacturers with fewer domestic software counterweights on trade policy. Hardware designers argue that unilateral curbs isolate American technology and accelerate Chinese domestic component substitution. The debate exposes differing operational priorities across the artificial intelligence value chain. Foundation model developers focus on immediate computing access, while silicon designers require broad export volumes to amortize fabrication expenses. Export controls also ripple through East Asian hardware supply networks. Advanced accelerators manufactured by Nvidia rely on high-bandwidth memory from South Korea and precision packaging foundries in Taiwan. Restricting shipments to Asian trading hubs reduces component procurement orders from memory producers. SK hynix and Samsung Electronics supply high-bandwidth memory stacks integrated directly onto Nvidia server substrates. A secondary front centers on cloud computing access rules. Federal regulators are weighing compliance rules requiring offshore data centers to verify customer identities before leasing American processing power. The federal filings do not state whether OpenAI dropped its export lobbying because of its commercial agreements with Nvidia. OpenAI has not published a formal statement explaining the change in its trade advocacy. The House Foreign Affairs Committee has scheduled its next legislative markup on advanced computing trade rules for Nov. 18. Registered entities must submit fourth-quarter federal lobbying disclosure reports by Jan. 20, 2027. ## Impact map | Event | Korea | China | Japan | Global impact | | --- | --- | --- | --- | --- | | OpenAI chip lobby retreat | Memory makers maintain high-bandwidth memory delivery volume for global AI servers | Buyers retain narrow window under case-by-case BIS licensing before statutory votes | Substrate and test equipment suppliers avoid immediate packaging order cancellations | Server hardware procurement costs stabilize as model developers cease pressing for broader export bans | | OpenAI chip lobby retreat | Memory makers maintain high-bandwidth memory delivery volume for global AI servers | Buyers retain narrow window under case-by-case BIS licensing before statutory votes | Substrate and test equipment suppliers avoid immediate packaging order cancellations | Server hardware procurement costs stabilize as model developers cease pressing for broader export bans | ## In this story - **Companies:** Nvidia, OpenAI - **Tickers:** NVDA - **Exposed:** SK hynix, Samsung Electronics - **Policy:** Export Controls - **Impact:** Compliance, Capex, Supply Chain ## Primary sources 1. bis.gov 2. wittywires.com 3. sifirincidakika.com 4. dealroom.co 5. cryptobriefing.com --- Cite as: East Asia Brief, "OpenAI abandons China chip curbs as Nvidia protects exports," October 22, 2026. https://eastasiabrief.com/trade-policy/openai-abandons-china-chip-curbs-nvidia-protects-exports-857