• Intel (INTC) CEO Lip-Bu Tan reveals that knowledgeable sources indicate the U.S. now trails China in open-source AI development.
  • The warning comes amid escalating U.S.-China AI rivalry, with China making strides in efficient models like DeepSeek.
  • Intel's stock has surged over 75% in 2025 under Tan's leadership, driven by AI optimism and restructuring efforts.

Intel CEO Lip-Bu Tan, in a candid assessment, stated that knowledgeable people have informed him the United States is now behind China in open-source artificial intelligence. This revelation, shared in recent discussions, underscores the intensifying global scramble for AI supremacy, with China leveraging state-backed scale and innovation to challenge traditional U.S. leadership.

Tan, who took the helm at Intel in March 2025, has been steering the semiconductor giant through a transformative period marked by layoffs and a strategic pivot toward AI. His comments, delivered against a backdrop of rising geopolitical tensions, highlight concerns that China's advancements in open-source AI—exemplified by models like DeepSeek, which uses significantly less compute power—could erode U.S. competitive edges. "We're hearing from experts that the momentum has shifted," Tan noted, emphasizing the need for urgency in response.

The broader context includes recent U.S. policy shifts, such as the loosening of export controls on Nvidia (NVDA)'s H200 chips to China, which could accelerate Chinese AI compute capabilities by two to three years by 2026. This move has sparked bipartisan pressure in the U.S. to reverse course, with lawmakers viewing AI leadership as a national security imperative. Meanwhile, China's President Xi Jinping has labeled AI an "epoch-making" technology, urging a "whole-of-nation" approach to indigenous development in the country's 15th Five-Year Plan (2026-2030).

Intel, once a dominant force in semiconductors, has seen its stock soar over 75% in 2025, fueled by optimism around its AI-focused initiatives, including the development of edge AI and agentic AI systems. The company is advancing its 18A process node products, with real shipments targeted for 2026, positioning it as a key player in the race. However, Tan's warning suggests that software innovations, like China's efficient open-source models, may outpace hardware advancements, complicating the competitive landscape.

Industry observers point to China's DeepSeek as a case in point: it rivals offerings from OpenAI and Meta while using 11 times less compute, demonstrating how software efficiencies can bypass hardware limitations. This has fueled debates in the U.S. about the need for policy reversals and increased investment in open-source frameworks. "Without a concerted effort, we risk ceding ground in a critical frontier," one analyst remarked, echoing Tan's concerns.

In the short term, Intel is betting on edge AI—on-device processing—as a "last credible bet," with products like Arrow Lake Refresh NPUs enabling Copilot+ desktops. The company's restructuring under Tan aims to make it more agile, akin to rivals like AMD (AMD) and Nvidia, but challenges persist as China's state-backed model excels in scaled deployment. Efforts to reach Chinese officials for comment on the open-source advancements were unsuccessful, but sources indicate ongoing provincial investments despite warnings against duplication.

Looking ahead, 2026 is poised to be a pivotal year, with experts urging humility in assessing China's opaque progress. The U.S. must monitor innovation, private sector roles, and defense ties closely to narrow the gap. As Tan put it, "This is a marathon, not a sprint," but the starting gun has already sounded in the open-source AI race.