Intel Foundry's Journey: Investment, Innovation, and the 2027 Target
Intel's semiconductor manufacturing business, or Intel Foundry, is currently in an investment-heavy phase. That means they're spending billions of dollars each quarter on developing new manufacturing technologies and increasing their production capacity. It's a costly endeavor, but Intel is gambling on a more lucrative tomorrow for this business segment, which will ultimately make a profit rather than a loss, possibly as soon as 2027. That timeframe coincides with some encouraging new technology launches, including their 14A manufacturing process and the start of production on a more advanced node called 18A-P.
The 18A Node: A Stepping Stone to Success
Intel recently confirmed again that their first product using the cutting-edge 18A (think 1.8-nanometer class) process is on track. It will be a general-purpose computer processor, codenamed Panther Lake, and we can look forward to seeing it on the shelves later this year, with production next year. This same 18A tech will also be used for their Xeon server chips, named 'Clearwater Forest,' and even for certain products made for other companies.
From a business perspective for Intel Foundry, however, the 18A node is more of a demonstration of its capabilities to potential external customers. If 18A is successful as a process and is stable, Intel hopes it will serve as an incentive for other companies to sign on for its follow-ons, including the 18A-P and the even more advanced 14A (1.4-nanometer class) nodes.
David Zinsner, Intel chief financial officer, gave a little insight into this at a recent conference. "I think we do need to see more external volume come from 14A versus 18A," he said. He discussed how they work with many potential customers, build test chips, and then some of those customers might drop off. "So, committed volume is not significant right now, for sure," he acknowledged. Zinsner emphasized that Intel needs to first prove the technology with its own products – literally "eating their own dog food" – before they expect a surge of external customer commitments.
Zinsner also acknowledged a potential cost hurdle. If Intel goes ahead with its current plans to use next-generation High-NA EUV lithography tools for its 14A process, the initial costs will likely go up. The hope, however, is that the advantages these new tools introduce will ultimately balance out the higher costs.
The Path to Breaking Even in 2027
Like other contract chipmakers, Intel doesn't reveal its client list. Intel also plans to manufacture more of its own chips in-house, such as its upcoming Panther Lake and Nova Lake CPUs. That will help Intel's margins and put its high-cost fabrication plants to better use. All of those factors contribute to Intel's expectation that its Foundry business will breakeven in 2027 and be profitable thereafter.
"We still think on track to hit breakeven sometime in 2027," said Zinsner. He clarified that when they initially set this goal way back in 2024, they gave a broader range, but most people, along with Intel, have tended to settle on 2027 as the goal.
Amazingly, to hit this breakeven point, Zinsner stated that Intel Foundry needs just a few billion dollars ("low- to mid-single-digit billions") of revenue from outside customers annually. While most of the 18A volume will be for Intel products themselves, the 14A node will rely more heavily on adoption by other entities. Intel's profitability plan also includes revenue from premium chip packaging services, legacy manufacturing nodes (e.g., Intel 16), and partnerships with companies like UMC and Tower. They will continue their 'smart capital' approach, midway between in-house chip production and outsourcing to third-party foundries, and have their own Foundry unit compete for internal business to foster efficiency and cost control.
Managing External Orders and Market Perception
Intel is not sounding overly optimistic about the short-term prospects of its foundry with external customers, admitting that orders are somewhat thin at the moment. This comes alongside rumors that Intel Foundry is set to be partnered with industry behemoths like NVIDIA to make future chips.
The chip business, in general, has seen some weakness for Intel. With a new CEO, Lip-Bu Tan, there is optimism for a turnaround, but Intel Foundry does not start without its challenges. Zinsner's description of the demand for future processes as not "significant" at present suggests that, at least in the short term, Intel's most advanced chipmaking will be committed largely to its own requirements. That could be a competitive handicap if it limits the scale and revenue growth of the foundry business.
"We get test chips, and then some of the customers fall off of the test chips. Committed volume is not material at present, definitely." - Intel's CFO
These comments could either dash earlier rumors of keen interest from the likes of NVIDIA for the 18A process, or perhaps Intel is being cautious in not validating these reports until there is additional concrete progress. Securing external customers is critical for Intel to ramp its foundry business and compete effectively.
In spite of these guarded statements, Zinsner reaffirmed the 2027 breakeven goal, subject to external customers taking up their processes. It's also interesting that Intel hasn't completely discounted the use of other foundries for its own chips; TSMC, for example, is set to be involved with the future Nova Lake desktop CPUs. This implies that Intel Foundry being a purely internal project is not the preferred result.
There is still lots of hope for Intel's 18A process, especially given recent news. NVIDIA, for instance, is reportedly looking for alternatives to TSMC, and both Intel and Samsung are the top contenders in this race. The next two to three years will be critical in determining if Intel's massive foundry investment will pay off.
