There are multiple components to Intel's (INTC 1.11%) turnaround story, but at the center is the company's aggressive and costly push to rival Taiwan Semiconductor Manufacturing in the semiconductor foundry market. Intel gave up its once-untouchable manufacturing edge through chronic delays and missteps, enabling TSMC to pull ahead and TSMC customers like AMD to sell chips that beat Intel on performance and efficiency.
The process of transforming itself from a company that solely makes chips for itself to a semiconductor foundry that also makes chips for others has been (and will continue to be) a long, drawn-out affair. Intel took the critical step this year of separating its manufacturing operations into a distinct business unit with its own profit and loss statement. In the past, manufacturing costs were spread across Intel's various product segments.
Don't worry about massive foundry losses
Intel formally unveiled the details of its new reporting structure in April, which led to a steep sell-off in the stock. The foundry segment was broken out for the first time, and the numbers didn't sit well with investors. The unit posted a $7 billion operating loss on $18.9 billion in revenue in 2023.
A $7 billion loss in a segment that is critical to Intel's future looks like a disaster, but there are some important things to know.
First, nearly all of Intel's foundry revenue right now is internal, meaning it's coming from the company's other segments. While there was little external revenue in 2023, Intel still invested heavily in its manufacturing operations. The costs associated with those investments are now being allocated to the foundry segment, but it will take time for these investments to pay off.
Second, Intel has only just begun operating its manufacturing operations like a distinct business unit. The results from 2023 are from another era when manufacturing costs were simply spread throughout the rest of the company.
Already, Intel has realized meaningful cost savings by having the foundry treat the product segments like customers. The number of production expedites requested by the product segments, which are expensive and hurt manufacturing efficiency, are down 95% now that product segments need to directly pay for them. There are billions of dollars in annual cost savings that can be unlocked thanks to the new model.
The path to profitability
Bringing in meaningful external revenue will be necessary for Intel's foundry segment to eventually turn a profit. The company expects the unit to reach breakeven around 2027 and for the operating margin to eventually reach 30%.
Intel has $15 billion worth of foundry deals already signed, including wafer manufacturing and advanced packaging. The company expects external revenue to top $15 billion by 2030, up from barely anything today.
While the dramatic shift from losses to profits that Intel envisions for the foundry segment may seem far-fetched, an important change in the economics of the company's manufacturing investments will help enable this turnaround.
When Intel was only making its own products, the company would invest in a new process node, use it for a time, and then largely throw it away. Under the foundry model, process nodes can be utilized for much longer periods as mature nodes are repurposed for different types of chips. Investments that Intel is making today will pay off for many years, far longer than the investments it made in the past.
The foundry party for Intel begins in earnest next year when the Intel 18A process debuts, but it may take until 2026 for meaningful capacity to drive significant revenue. Intel 18A is on track, according to the company, and it already has a handful of customers lined up, including Microsoft.
While the foundry appears to be a train wreck right now looking at the numbers, that $7 billion annual loss isn't particularly meaningful. For long-term investors, Intel is a turnaround stock worth buying.
