The newly appointed CEO of Intel (INTC.US), Lip-Bu Tan, who took the helm of the U.S. chip giant in March, has strongly refuted recent reports suggesting that a new Intel employee leaked trade secrets from Taiwan Semiconductor Manufacturing (TSM.US)—the world’s largest chipmaker—to Intel. Tan emphasized that the company respects all semiconductor firms’ intellectual property (IP).

Media reports this week focused on the career move of 75-year-old Lo Wen-jen, who retired from TSMC earlier this year and recently joined Intel. Some outlets cited unnamed sources claiming the executive took proprietary chip manufacturing technology from his former employer shortly before leaving.

「These are baseless rumors and speculation. We respect all chip manufacturing IP,」 Tan stated during an interview at the Semiconductor Industry Association awards ceremony in San Jose, where TSMC CEO C.C. Wei and former Chairman Mark Liu received the prestigious Robert N. Noyce Award.

TSMC, with a market cap exceeding $1.15 trillion (Taiwan stock exchange) and $1.5 trillion (U.S. ADR), is the undisputed leader in chip foundry services, far surpassing Intel. Its proprietary data and advanced process technologies are not only valuable trade secrets but also strategic assets. Reports indicate TSMC has launched an internal investigation into Lo’s alleged actions to determine if any laws were violated.

Before retiring in July, Lo oversaw TSMC’s process strategy and played a key role in mass-producing cutting-edge chips, including those powering AI accelerators from NVIDIA and AMD. Prior to joining TSMC in 2004, Lo worked at Intel, focusing on advanced technology development and managing a major fab in Santa Clara, California.

Tensions between Intel and TSMC have grown in recent years as Intel strives to catch up in the chip race. While Intel is both a TSMC customer and competitor, TSMC remains the exclusive foundry for Apple, NVIDIA, AMD, Qualcomm, and Broadcom’s most advanced chips. Former Intel CEO Pat Gelsinger repeatedly warned that U.S. over-reliance on TSMC for cutting-edge semiconductors is 「too risky,」 fueling friction between the two.

For Intel, mastering leading-edge, high-volume chip manufacturing is critical. The company is investing heavily, even purchasing ASML’s first High-NA EUV lithography machine, to surpass TSMC and Samsung in 2nm and below technologies. Under Tan, Intel’s aggressive roadmap—"5 nodes in 4 years"—aims to transition from Intel 7 to Intel 3, 20A (2nm), and 18A (1.8nm).

With stagnant growth in PC/server CPUs, Intel is pivoting to foundry services (Intel Foundry), directly competing with TSMC. Success in 18A and 16A technologies could position Intel to manufacture advanced AI accelerator chips for clients like NVIDIA, AMD, and Qualcomm.

Qualcomm (QCOM.US) may become Intel’s first major foundry customer. Citi analysts noted Qualcomm’s job postings seeking candidates with Intel’s embedded multi-die interconnect bridge (EMIB) packaging expertise, suggesting potential collaboration on AI ASICs for data centers.

Qualcomm’s newly announced AI200/AI250 accelerators, targeting rack-scale AI inference clusters, could drive significant growth in its data center chip business by 2026. These ASIC-based solutions, competing with Google’s TPU, emphasize energy efficiency and total cost of ownership (TCO) for AI operators.