Starting in the first quarter of 2024, Intel (INTC +1.14%) will report its foundry operations as a separate business unit. This is part of the company's plan to grow into a major player in the foundry market by 2030. Intel's product divisions will treat the foundry more like an external supplier, and the foundry will treat Intel's product divisions like customers.
In preparation for this switch, Intel recast its financial statements for the past few years to match the new reporting structure. The foundry business is highly unprofitable right now, which shouldn't be too surprising. Almost all of Intel's foundry revenue is internal, the company has been investing heavily in new manufacturing technology and facilities, and the post-pandemic PC market crash gutted demand for the company's chips.
The stock market reacted poorly to the disclosure, sending Intel stock tumbling lower on Wednesday. Intel's foundry segment reported an operating loss of nearly $7 billion in 2023. The company's assurances that the foundry segment would reach breakeven around 2027 and achieve a 30% adjusted operating margin by 2030 did little to brighten investors' moods.
NASDAQ: INTC
Key Data Points
Why profits can grow rapidly once the ball gets rolling
One thing to note is that the foundry segment only recently began operating as its own business unit. That $7 billion loss in 2023 is partly a consequence of Intel's manufacturing costs historically being spread out among its product divisions.
While Intel's recast financial statements are helpful for investors, they may not be all that meaningful as a guide to future results. The product divisions had little incentive to utilize the manufacturing side of Intel efficiently in the past. Beyond the low-hanging fruit of cost efficiencies that can be unlocked now that the foundry is its own business unit, the switch to serving external customers, as well as internal customers, will lengthen the useful life of Intel's manufacturing assets.
When Intel was only making chips for itself, usage of a particular process node would fall off quickly once a new process node came online. As the capital requirements for leading-edge semiconductor foundries explode, this business model just doesn't work anymore.
Under the foundry model, Intel can extend the lifespan of its process nodes. The Intel 3 process, which the company will use for its upcoming server CPUs, will also be available to foundry customers. The company is planning on multiple revisions to Intel 3, adding new features and enhanced performance. Once the node is no longer suitable for cutting-edge products, it can still appeal to customers looking for mature, cost-effective manufacturing.
Intel will play a similar game with the Intel 18A process, which will be ready by the start of 2025, and its next-generation Intel 14A process. Both process nodes will see revisions, extending their useful lifetimes well beyond what would be possible if the company only planned to manufacture its own chips.
Around one-third of market-leader TSMC's revenue comes from process nodes that are 10 nanometers (nm) or older, and 24% of revenue still comes from process nodes that are 28nm or older. (For perspective, TSMC launched its 28nm node in 2011.) Improvements have been made over time to these older nodes, but these investments made long ago continue to pay off. Intel will look to replicate this model.
Intel's targets aren't farfetched
While some investors may have trouble believing that Intel will be able to pull its foundry business from a $7 billion annual loss to a 30% operating profit margin by 2030, this shift in how the company utilizes its manufacturing assets will be a powerful profit driver. Intel 3, Intel 18A, Intel 14A, and other future nodes will be churning out revenue for a very long time. The company will spend around $100 billion in the U.S. alone on building and expanding manufacturing facilities, something that wouldn't be feasible if new process nodes didn't have long lifetimes.
Once Intel's foundry business starts generating meaningful revenue from external customers, the profit picture can improve quickly. The company's targets may look ambitious, and it certainly has a lot to prove, but Intel's goals are realistic.
