Intel Faces $18.8 Billion Loss, CEO Considers 18A Shift

Intel's new CEO, Lip-Bu Tan, has been swift in implementing cost-cutting measures and exploring new growth avenues since taking charge in March. His initial steps included restructuring the business, reducing middle management, and hiring new engineering talent to stabilize Intel's finances. The urgency of these actions was underscored by Intel's $18.8 billion net loss for 2024, marking the company's first unprofitable year since 1986.

One of the key areas of focus for Tan has been the manufacturing process known as 18A, which was championed by his predecessor Pat Gelsinger. Tan has expressed doubts about 18A's ability to attract new customers and is considering a significant shift towards the contract manufacturing business, known as the "foundry" business. Sources reveal that 18A, which includes a novel transistor design and power-delivery scheme, may be shelved for external sales. This decision could result in a substantial write-off, with analysts estimating hundreds of millions, if not billions, in losses. The appeal of 18A seems to be diminishing as rivals like TSMC advance with their N2 node.

Intel has declined to comment on these developments, stating that the executive team is committed to strengthening the company's roadmap, building trust with customers, and improving its financial position. The company has identified clear areas of focus and is taking actions to turn the business around. Intel is also considering selling its networking and edge-computing operations, known as NEX, as part of its broader strategy to focus on more profitable businesses.

Tan's strategy involves redirecting resources to Intel's next-gen 14A process, which he believes could give the company a competitive edge if major partners buy in. The company is urgently appealing to customers and mapping out options for the board's upcoming meeting, possibly as soon as this month. A final verdict on whether to stop pitching 18A externally may not come until the autumn session, given the high stakes and complexities involved.

Even if the marketing of 18A is curtailed, Intel plans to continue using the process for in-house designs and existing commitments. This means that Panther Lake laptop chips, praised as the most advanced processors ever made in the US, will still be produced on the 18A line in late 2025. Intel must also honor contracts to produce chips for Amazon and Microsoft on the same node, given their tight deadlines.

The broader picture remains uncertain. Intel's hopes hinge on delivering 14A swiftly and tailoring it to marquee clients' needs. If successful, Intel could reclaim ground lost to TSMC over the past two decades, especially in the booming realms of mobile computing and AI. However, any slippage in 14A development or an unfavorable board decision might prompt the company to stick with 18A's original roadmap.

Tan's deliberations highlight the significant risks and investments required to revive Intel's foundry ambitions. As the board weighs his proposals, the chip world watches closely. A course correction now could shape the future of American semiconductor manufacturing for years to come, as well as the company's profitability.