Q1 '26 Foundry Earnings: Hit Or Miss?

Key Takeaways

• Intel Q1 revenue $13.6B, up 7% YoY, but net loss $4.28B.
• Intel Foundry Services posted $2.4B operating loss on $5.4B sales.
• Foundry revenue up 20% YoY, fueled by EUV mix and 14A.
• Core Ultra “Panther Lake” CPUs on 18A lifted CCG and DCAI earnings.
• Musk’s $20‑$25B Terafab will adopt Intel’s 14A process.

Pulse Analysis

Intel’s first‑quarter earnings paint a picture of revenue growth offset by deep losses, especially within its foundry division. While overall sales climbed to $13.6 billion, the company posted a $4.28 billion net loss, largely because Intel Foundry Services (IFS) recorded a $2.4 billion operating deficit on $5.4 billion of sales. The loss reflects the ongoing challenge of scaling external wafer production, where external customer revenue remains modest at $174 million, even as the unit’s internal chip sales drive top‑line numbers. Analysts see the gap as a critical hurdle for Intel’s ambition to become a true third‑party foundry.

On the technology front, Intel’s manufacturing roadmap shows encouraging signs. Yield rates on the 18A node, a 2 nm‑class process introduced in 2025, have risen above internal targets, and the next‑generation 14A node is reportedly ahead of schedule in both maturity and performance. This progress underpins the company’s partnership with Elon Musk’s Terafab project—a $20‑$25 billion Texas fab that plans to use Intel’s 14A process for high‑performance AI workloads. Coupled with a surge in advanced packaging demand, such as EMIB‑T, Intel is positioning its fab capabilities to serve emerging AI and high‑bandwidth applications, potentially narrowing the gap with rivals like TSMC.

Nevertheless, Intel still lags in dedicated AI silicon. The firm’s current portfolio lacks a competitive AI accelerator, prompting collaborations with Google on a custom IPU and a strategic partnership with SambaNova, which recently secured $350 million in funding that includes Intel participation. SambaNova’s fifth‑generation RDU chip promises five‑fold speed advantages for large‑scale models, and the joint heterogeneous solution aims to blend GPUs, Xeon CPUs, and RDU units for next‑gen inference. If these alliances mature, Intel could convert its process leadership into a more compelling AI offering, reshaping its market position as AI workloads dominate future semiconductor demand.