A year-end wave of AI releases, Commerce updates China-focused chip rules, and the CDAO’s $100 million deal with Anduril

Subscribe to policy.ai
Worth Knowing

A Big Month for AI Releases — Google, OpenAI, and Meta Make Waves: 2024 closed with a wave of major AI releases from the industry’s leading companies:
While none of these releases represent a GPT-3-to-GPT-4 scale breakthrough (and it seems unlikely that OpenAI’s 12-day event will conclude with GPT-5), the consistent improvements are stacking up quickly.
Intel’s Future in Flux After CEO Gelsinger is Pushed Out: Intel announced the retirement of CEO Pat Gelsinger earlier this month, less than four years into his tenure leading the company. According to Reuters, Gelsinger’s abrupt departure came after Intel’s board gave him the choice to retire or be fired. Gelsinger, an engineer by training who first joined the company in 1979 before leaving in 2009, returned to Intel in 2021 with ambitious plans to revitalize the company and, with it, American chipmaking. The centerpiece was a proposed $100 billion investment in new U.S. manufacturing facilities and the launch of Intel Foundry Services to compete with contract manufacturers like TSMC. But the market never bought in — Intel’s stock plummeted more than 60% under Gelsinger’s tenure, and last month the company was dropped from the Dow Jones Industrial Average in favor of Nvidia. But it would be wrong to lay Intel’s decline entirely at Gelsinger’s feet. The company has struggled to achieve success as an integrated device manufacturer — designing and manufacturing its own chips — when its biggest competitors have all gone fabless — designing their chips but outsourcing the production to contract chipmakers like TSMC. Intel’s foundry business has struggled to attract major customers, likely because so many of them compete with Intel’s chip design business. Gelsinger’s exit has sparked speculation that the board may pursue a split of Intel’s manufacturing and design operations — an outcome Gelsinger had resisted. Either way, Intel’s downward spiral should be cause for concern in Washington, DC. The company is a cornerstone of the Biden administration’s plan to revitalize U.S. semiconductor manufacturing through the CHIPS and Science Act — the Commerce Department just finalized $8 billion in federal subsidies for Intel last month — but with Gelsinger gone and Intel’s future uncertain, that strategy could be in danger.
Government Updates

White House Debuts Another Update on China-Focused Chip Export Rules: Earlier this month, the Commerce Department announced significant updates to its semiconductor export controls, further tightening restrictions on China’s access to high-end chips and manufacturing equipment. The new rules mark the latest evolution of controls first introduced in October 2022 and updated last year. While the original controls have had an impact — the CEO of Chinese AI lab DeepSeek has cited them as a significant obstacle — their effectiveness has been mixed. High-end chip smuggling has been widespread (despite its CEO’s comments, Dylan Patel of SemiAnalysis estimates DeepSeek has accumulated some 50,000 nominally-controlled Nvidia Hopper GPUs), and Chinese chipmakers have built substantial stockpiles of semiconductor manufacturing equipment (SME) in anticipation of future restrictions. The new 200+ page rules package (part 1 and part 2) attempts to address these issues through three main mechanisms:
  • Expanding restrictions on SME and the software used to produce advanced-node integrated circuits. Of particular note is a significant expansion of the Foreign Direct Product Rule that extends U.S. jurisdiction over any foreign-produced SME that contain “any amount of U.S.-origin integrated circuits”
  • New controls on exports of high-bandwidth memory chips, which are essential for most AI systems
  • The addition of 140 Chinese entities to the Bureau of Industry and Security’s Entity List, many of which are engaged in China’s indigenous semiconductor development
The development of these rules reflects intense negotiations between national security hardliners who have advocated stricter controls and industry representatives and allied governments reluctant to abandon the Chinese market entirely. This attempt to thread the needle has drawn criticism — CSIS’s Gregory C. Allen argued to the New York Times that the extended deliberations made the rules unnecessarily complex, may have introduced unintended loopholes, and gave Chinese companies sufficient time to stockpile now-controlled chips and equipment. Soon after the updated rules were announced, China’s Ministry of Commerce responded by imposing export restrictions on critical materials — including gallium, germanium, and antimony — used in semiconductor manufacturing. Looking ahead, President-elect Trump seems unlikely to ease restrictions on high-end AI chip exports to China, having initiated similar efforts to restrict China’s access to critical technologies during his first term. The key question is whether his administration will attempt to maintain Biden’s balancing act or further tighten the controls.

Trump Taps David Sacks as White House AI and Crypto Czar: President-elect Donald Trump tapped venture capitalist David Sacks to serve as his administration’s “AI & Crypto Czar” and to lead the Presidential Council of Advisors for Science and Technology. An early executive at PayPal and angel investor in Facebook, SpaceX, and Palantir, Sacks has built up a significant public profile in recent years as one of the co-hosts of the popular “All-In” podcast. In announcing the appointment, Trump said Sacks would “focus on making America the clear global leader” in both AI and crypto while working to “steer us away from Big Tech bias and censorship.” The exact scope and authority of Sacks’ position remain unclear. According to Bloomberg, Sacks will be classified as a “special government employee” capable of working up to 130 days per year. While not subject to Senate confirmation or asset disclosure requirements, Sacks still must recuse himself from matters where he has financial interests. Observers see Sacks’ appointment as another sign of Elon Musk’s influence within the incoming administration. Sacks and Musk (who is set to advise the president as the co-chair of the proposed “Department of Government Efficiency”) worked together at PayPal and have reportedly maintained close ties. Some observers have raised concerns that Sacks and Musk could use their positions to target rivals — including OpenAI, which Musk sued earlier this year — or privilege their own business interests. OpenAI CEO Sam Altman appears eager to smooth over any issues with the incoming administration, donating $1 million to President-Elect Donald Trump’s inauguration fund and congratulating Sacks on social media, which Musk responded to with a laughing emoji.

U.S.-China Commission Proposes “Manhattan Project” for AI Development: In its annual report, the U.S.-China Economic and Security Review Commission recommended Congress establish and fund a Manhattan Project-style program for AI to ensure the United States develops artificial general intelligence (AGI) before China. The report calls on Congress to: 1) give the executive branch broad multiyear contracting authority for funding AI, cloud, and data center firms; and 2) direct the DOD to classify AI-related items as “highest priority” rating in the Defense Priorities and Allocations System. While the commission’s recommendations aren’t binding, the proposal has garnered attention in Washington and Silicon Valley, especially because the incoming Trump administration seems like it could be receptive to the idea. In July, the Washington Post reported that Trump allies were drafting AI-focused executive orders incorporating similar Manhattan Project-style initiatives for military applications. And one of the U.S.-China Commission’s 12 members — Palantir Senior Advisor Jacob Helberg — was tapped last week to lead economic diplomacy at the State Department. Tech industry leaders, including Google CEO Sundar Pichai, have expressed willingness to participate in a Manhattan Project-type initiative. But some experts urge a more measured approach. As Anthropic co-founder and former CSET Non-Resident Fellow Jack Clark noted during a panel at the recent New York Times Dealbook Summit, a “Manhattan Project for AI” could be an unnecessary response when less-expensive government initiatives — like high-skilled immigration reform or increasing access to computing resources — might be sufficient.

CDAO Inks A Deal With Anduril for Tactical Data Platform: The Pentagon’s Chief Digital & AI Office awarded Anduril a $100 million contract to expand its “Lattice Mesh” system, which enables frontline forces to rapidly share and process data from hundreds of different sensors. Anduril followed up the contract by releasing its Lattice software development kit (SDK) last week, enabling third-party developers to build applications that can integrate directly into Lattice. As Ben Thompson noted in his Stratechery newsletter, these moves could mark an inflection point for military software development. While the DOD has historically struggled with software procurement, Anduril’s platform approach could create an accessible ecosystem more familiar to Silicon Valley. The CDAO contract reflects an evolution in the Pentagon’s data strategy. When Defense Secretary Lloyd Austin approved the Joint All-Domain Command and Control (JADC2 – since renamed “Combined Joint All Domain Command & Control” or CJADC2) initiative in 2021, DOD planners envisioned a single unified network connecting sensors across services. But as CDAO Principal Deputy Margaret Palmieri told Breaking Defense, her office now recognizes that tactical and strategic commanders have fundamentally different needs — the former requiring rapid access to limited data, the latter needing to process massive amounts of information over longer periods. Anduril’s Lattice aims to fill that tactical role, while Palantir is working on a platform for the strategic and operational levels. Earlier this month, the two firms announced plans to work together to share data between the two platforms to enhance data processing, including for AI training.

House AI Task Force Publishes Year-End Report: The House Task Force on AI released its end-of-year report on Tuesday, offering a roadmap that emphasizes sector-specific oversight over comprehensive new AI regulations. Led by Reps. Obernolte and Lieu, the AI task force was set up earlier this year to help Congress balance U.S. leadership in AI with appropriate safeguards for its development and deployment. The 24-member bipartisan group consulted with experts across academia, industry, and government to develop recommendations spanning 15 core areas from healthcare and intellectual property to government use of AI and national security. Rather than propose new regulatory frameworks, the 253-page report advocates that federal agencies take the lead on regulating AI within their sectors using their existing authorities. “We do not think it is a good idea for the United States to follow some of the other countries in the world in splitting off AI and establishing a brand new bureaucracy and a universal licensing requirement for it,” Obernolte told reporters (CSET’s Jack Corrigan, Owen Daniels, Lauren Kahn, and Danny Hague made a similar case in their June paper, Governing AI with Existing Authorities). While the 118th Congress will end next month, the report is an important signal of where Congress could be headed on AI in the next session.

In Translation
CSET’s translations of significant foreign language documents on AI

PRC Industry Response to Export Controls: Four Chinese Industry Associations Issue Statements Condemning U.S. Sanctions. This translation combines statements issued separately by four Chinese industry associations on December 3, 2024, condemning the new sanctions and export controls aimed at Chinese companies that the United States announced the previous day. All four statements encourage Chinese companies to reconsider purchases of U.S. chips and semiconductor equipment and to look elsewhere for suppliers. The Chinese Communist Party controls all industry associations in the country, so the coordinated statements should be understood to reflect the concerns of the Chinese leadership.

PRC Ministry of Commerce Export Ban: Ministry of Commerce Notice 2024 No. 46: Notice Concerning Strengthening Controls on Exports of Relevant Dual-Use Items to the United States. This notice from China’s Ministry of Commerce, issued December 3, 2024, bans the export of gallium, germanium, antimony, and superhard materials to the United States, and enacts stricter export control checks on dual-use graphite materials. This move immediately follows the U.S. announcement on December 2 of new export controls and sanctions designed to prevent sales of advanced U.S. microchips and semiconductor manufacturing equipment to China.

PRC Cybersecurity Standards Draft: National Standard of the People’s Republic of China: Cybersecurity Technology – Basic Safety Requirements for Generative Artificial Intelligence Services (Draft for Feedback). This draft Chinese national standard is designed to improve the safety and security of generative AI services. The standard addresses some cybersecurity concerns associated with generative AI, but primarily focuses on preventing AI systems from generating content the Communist Party finds objectionable, such as pornography, bullying, hate speech, defamation, copyright infringement, and criticism of the Party’s monopoly on power.

If you have a foreign-language document related to security and emerging technologies that you’d like translated into English, CSET may be able to help! Click here for details.

Job Openings

We’re hiring! Please apply or share the role below with candidates in your network:
  • Director of Analysis: CSET is seeking applications for a Director of Analysis position. This leadership role combines organizational strategy, research oversight, and project & team management, will oversee CSET’s research agenda, manage a diverse team of 25+ researchers, and ensure high-quality, policy-relevant outputs.
Please bookmark our careers page to stay up to date on all active job postings. You can also subscribe to receive job announcements by updating your subscription preferences in the footer of this email.

What’s New at CSET

REPORTS

PUBLICATIONS AND PODCASTS

EMERGING TECHNOLOGY OBSERVATORY

EVENT RECAPS

  • On December 11, Matt Sheehan of the Carnegie Endowment for International Peace and Daniel Schiff, Assistant Professor and Co-Director of the Governance and Responsible AI Lab at Purdue University joined CSET’s Mina Narayanan and Zachary Arnold to discuss potential forecasts for AI governance efforts around the world in 2025. The webinar also featured the exclusive introduction of the Emerging Technology Observatory’s AGORA, a new tool designed to help researchers and the public examine AI-relevant laws, regulations, and standards. Watch a recording of the event.

IN THE NEWS

What We’re Reading

Annual Report: Military and Security Developments Involving the People’s Republic of China, U.S. Department of Defense (December 2024)

Working Paper: Under Pressure: Attitudes Towards China Among American Foreign Policy Professionals, Michael B. Cerny and Rory Truex (December 2024) 

Report: How Federal Funding Terms and Conditions Could Encourage Safe Artificial Intelligence Development, Ryan Consaul and Gregory Smith, RAND Corporation (November 2024)

Paper: Whack-a-Chip: The Futility of Hardware-Centric Export Controls, Ritwik Gupta, Leah Walker, and Andrew W. Reddie (November 2024)