Semiconductor heavyweight Intel (INTC -9.20%) announced last week that it would be separating its programmable chip segment from the rest of the business. The field programmable gate array (FPGA) business will be operated under its own CEO, with an eye toward a partial IPO within three years.
Intel's stock rose slightly on the news when announced, but the move could be interpreted both positively or negatively. Either way, the FPGA separation follows several spinoffs and capital raises Intel has made over the past couple years.
Moreover, it's also possible -- or even likely -- Intel isn't going to stop here. Under CEO Pat Gelsinger, Intel appears to be on a path to a long-term reorganization, which could eventually result in Intel breaking up into several different companies.
The breakup would make Intel and its constituent parts smaller, but could boost shareholder value. With the stock well below all-time highs, it could be a smart move.
The FPGA business will get its own leader
In the announcement, Intel said the FPGA separation will "give PSG the autonomy and flexibility it needs to fully accelerate its growth and more effectively compete in the FPGA industry, which serves a broad array of markets, including the data center, communications, industrial, automotive, aerospace and defense sectors."
The new unit will begin operations as a stand-alone business January 1, 2024, and be led by Sandra Rivera, who previously headed up Intel's Data Center and AI segment. The DCAI segment has been executing much better of late, so it's not a surprise to see Rivera "promoted" to CEO of a unit that could one day become its own stand-alone public company.
Intel acquired the FPGA business via its acquisition of Altera back in 2015 under prior management. FPGAs can be programmed through software and firmware after installation. In this respect, they are more flexible than custom ASICs (application-specific integrated chips), but are generally less powerful and efficient for specific tasks.
Image source: Getty Images.
Reasons for the separation
It was thought that Intel's core CPU business might have synergies with FPGAs, as Intel would be able to perhaps bundle total integrated systems rather than sell one-off chip sales. Certainly, the fact that Advanced Micro Devices bought FPGA competitor Xilinx in 2022 indicates that there could be reason for CPU/GPU designers to also have FPGA capabilities under the same company umbrella.
So the move may merely be designed to raise money for cash-strapped Intel through an IPO.
After all, Intel has engaged in several divestitures under Gelsinger, as it aims to build its own foundry ecosystem. In 2021, Intel sold its NAND flash business. In 2022, it executed a partial IPO of its Mobileye (MBLY -5.50%) self-driving car system subsidiary. And this summer, it sold a minority stake in its IMS nanofabrication business, which makes multi-beam mask writing tools. Intel even brought in co-investors to build out fabs for its foundry this year, with partner Brookfield Infrastructure Corporation now owning some of those economics.
With Intel's core PC business still depressed and the company still lagging behind on leading-edge processes until 2025, Intel may feel it needs to tap other sources of cash as it moves forward with its ambitious and expensive foundry plans. So, the FPGA unit separation and IPO may be geared solely toward that fundraising objective.
But there may be a larger plan at work
Even beyond the separation and fundraising angle, it appears Intel could be on a path to break itself up. After all, it's possible that its various units may add up to much more than Intel's stock trades for today.
The memory, FPGA, Mobileye, and IMS units were never really considered "core" Intel businesses, and Intel doesn't get much credit for owning them. So, it's not a surprise to see Intel spinning them off to get public market valuations.
But even Intel's "main" businesses in CPUs, AI GPUs and accelerators, and the new foundry could separate as well. Unlike most other semiconductor companies in the world, Intel is currently both a designer of processors as well as a manufacturer.
However, Pat Gelsinger is now building out Intel's foundry ecosystem to serve third party designers, while also admitting that Intel designs may also be manufactured by outside foundries, depending on capability and cost.
But if that's the case, is there a reason for Intel's foundry and CPU/GPU design teams to be under the same corporation?
In fact, about one year ago, Gelsinger outlined his vision for an "internal foundry" model for Intel. Under the new model, Gelsinger said Intel will have, "a fundamental shift in mindset," in which its manufacturing arm will become even more independent of Intel's designers.
Gelsinger gave the example that in the past, Intel's designers would often ask for additional product stepping from the manufacturing side to test out ideas, which would lead to inefficiencies at the foundry level. Therefore, Gelsinger pointed out more separation and independence for the foundry, putting Intel requests on par with third party customer requests, would lead to greater efficiencies.
But if that's the case, why have an integrated design/manufacturing company at all?
Intel could break up, and it could be good for the stock
In general, "pure play" companies in public markets tend to be valued higher than conglomerates that own various businesses. It's not clear why that is, but it happens.
Therefore, if -- and this is a big if -- Intel can execute on its foundry buildout while catching up to design competitors on the leading-edge, it may be able to generate more value for shareholders by breaking up into its individual parts.
After all, a lot of the reasons given for these current spinoffs are in the name of "efficiency" and "focus." So why not just continue with this trend?
In any case, Intel investors should consider the possibility of an eventual break-up as a potential upside scenario for the stock.
