The investment firm Third Point LLC is pushing for Intel to consider splitting its foundry business from its chip design company. This seems to be a standard tactic for the firm, which has a long history of pushing various companies to engage in spinoffs, mergers, and acquisitions. In this case, Third Point believes Intel should consider spinning off its manufacturing arm. It anchors these arguments in the problems we’ve seen Intel struggling with over the past few years — the repeated delays to 10nm and the more recent delay at 7nm.
Some of Third Point’s… points are hard to disagree with. It’s true that Intel has struggled to execute in recent years. It’s been long enough, frankly, that “recent years” doesn’t really capture the scope of the problem. Intel’s 14nm was the first node to be delayed due to manufacturing issues, which means you can argue that the company has been struggling with node transitions for the past six years. That’s not trivial.
But there are a few specific reasons to think Third Point is stirring the pot. First, Intel has already stated that it’s considering tapping third parties for cutting-edge CPU production. The company is going to have to talk about these issues in its investor presentations in 2021 — there’s no way to avoid it.
Intel enjoys some of the best profit margins in the industry, and owning its own factories has always been key to that equation. At the same time, it was only by delivering higher-performing products more reliably than any other firm that Intel was able to justify the high cost of maintaining its own dedicated facilities. So long as the high costs appeared justified by high performance, Intel could argue in favor of retaining its own manufacturing. If Intel is unable to compete with the likes of TSMC and Samsung, how can it possibly justify owning its own manufacturing facilities, as opposed to agreeing to build hardware at a pure-play foundry?
It’s a fair question, but it oversimplifies the practical difficulty of Intel moving its silicon volume to any other manufacturer. First, neither TSMC nor Samsung has enough spare capacity to absorb Intel’s manufacturing demand. Intel could undoubtedly partner with TSMC on leading-edge nodes, in much the same fashion that Apple does today, but that’s going to require TSMC to build out larger leading-edge facilities. That kind of buildout can take 12-18 months, while a new facility may require 3-5 years of construction time.
Second, it’s not clear that splitting Intel’s foundries from the design aspect of the company makes sense for either. Intel’s manufacturing rules are specifically designed for Intel microprocessors. Intel does not always use the same design rules for a given node that TSMC did, and it emphasizes high-performance silicon rather than ultra-low-power chips. The company had trouble finding customers for its client foundry business because Intel’s design rules are restrictive.
A hypothetical spinoff (Intel Foundry) could continue to build chips for a chip designer (Intel Engineering, let’s call it), but Intel Foundry would have to overhaul its own fabs to appeal to other customers. That, again, is going to take time. Literally nothing about the AMD / GlobalFoundries spinoff suggests we should assume this will go well. It might be the right move, but it’s likely to be a costly one.
Intel’s biggest problem is that confidence in its manufacturing prowess has waned to the point that conversations like this are happening at all. Semiconductors require long investment cycles that pair poorly with the quarterly focus of the market. Before it delayed its 7nm node, Intel had announced it would regain process leadership with its own 5nm node in 2023. Presumably, the best-case date for that now is 2024. If 5nm were to shift outwards again, Intel might need until a hypothetical 3nm to regain superiority, with an extremely theoretical launch date of somewhere between 2026 and 2028.
All of these facts are already weighing on Intel based on comments made by Bob Swan when he announced the 7nm delay. The case Third Point makes for a spinoff isn’t as cut and dry as the company would like to make it sound, but whatever decision Intel makes on these issues will have a profound impact on the company in years to come, for good or ill. We’ll almost certainly hear more about these issues during the company’s quarterly conference call in January 2021.
Intel’s Desktop TDPs No Longer Useful to Predict CPU Power Consumption
Intel's higher-end desktop CPU TDPs no longer communicate anything useful about the CPUs power consumption under load.
Review: The Oculus Quest 2 Could Be the Tipping Point for VR Mass Adoption
The Oculus Quest 2 is now available, and it's an improvement over the original in every way that matters. And yet, it's $100 less expensive than the last release. Having spent some time with the Quest 2, I believe we might look back on it as the headset that finally made VR accessible to mainstream consumers.
New Intel Rocket Lake Details: Backwards Compatible, Xe Graphics, Cypress Cove
Intel has released a bit more information about Rocket Lake and its 10nm CPU that's been back-ported to 14nm.
Intel’s Raja Koduri to Present at Samsung Foundry’s Upcoming Conference
Intel's Raja Koduri will speak at a Samsung foundry event this week — and that's not something that would happen if Intel didn't have something to say.