The recent news that Intel (INTC) will be a chip foundry for Altera (ALTR) and conjecture that this is just the first of several foundry relationships has brought a new aspect to how we should look at Intel the company. Much of Intel analysis in recent months has been in relation to ARM Holdings (ARMH) as the world becomes increasingly ARM and Intel architecture-centric. Intel and ARM comparisons never made much sense to me as their business models are night and day. (Do we compare apples and hamburgers because they are both round?) Comparing Intel to other foundries makes more sense since if Intel were to make a major foundry push, it would be competing directly with existing IC foundries.
If we compare Intel to anyone in the chip foundry business, it should probably be TSMC (TSM), who is the largest and most technically advanced of the foundry players. From a techie point of view, Intel is generally considered to have world-class foundries, and the Intel technical advances are based on what they call a tick-tock model. Focus one release cycle on improving the processor architecture, then focus the next on improving the IC process, then back again to improving the architecture. The result has been decades of both the most sophisticated processors and the most advanced IC manufacturing processes in the world. Intel and TSMC regularly are in the horse race to be first with the newest and smallest IC technologies. Intel usually wins but not for long.
So assuming a general technical parity in foundries, it is worthwhile to look at the respective business models for Intel and TSMC.
1. Now comparing a U.S. and a Chinese (Taiwanese actually) company is not always 'espectively higher than Intel. So relatively, the market values TSMC's business model significantly higher than Intel. Further confirmation is the market has kicked Intel down some 20% over last year's highs, whereas TSMC has seen fairly steady rises in valuation.
2. On Profit and Operating Margins, TSMC does quite a bit better than Intel, with 30% plus margins compared to Intel's 20% plus margins. Part of that is cost differential of China vs. U.S. companies, but a majority is the business focus itself. Despite the huge capital investments required, the IC foundry business done right can be very profitable.
3. On ROA and ROE measures, TSMC and Intel are roughly tied. TSMC numbers are little higher, but not enough to be worth justifying.
So what can we take away from this?
First, an increasing focus on the merchant IC foundry, may not only increase Intel's overall margins, but if TSMC is a standard, it may increase the respect and valuation that the market gives the company.
Second, the nature of the foundry business is that it is very capital intensive, but incremental costs are relatively minor. A foundry that is busy 24 hours a day is much more profitable than one that is not... and a lot of that extra business drops down to the bottom line very quickly. If part of Intel's current issues are foundry utilization-related, expanding its merchant foundry business is a quick and easy fix to improve its margins using its existing infrastructure.
Third, Intel would be entering a business where it already has a world-class reputation. Since Intel has leading processes, many companies, in addition to Altera, will be more than ready to have Intel making their ICs. The model of a captive foundry making chips for 3rd party IC companies is not new. IBM (IBM) and TI (TXN) - other companies with major and probably undervalued foundry assets, have a history of making chips for third party IC companies. IBM has had an on-again off-again relationship as a foundry supplier for Xilinx (XLNX), Altera's alpha competitor, going back many years, TI was long time foundry for Sun SPARC processors.
Of course, it is very doubtful that Intel will take anything other than incremental steps into the IC foundry business. Despite opinion to the contrary, Intel's processor business should still be alive and well for beyond the foreseeable future, and will utilize much of their foundry resources. It will be interesting to see Intel's position on being the foundry for ICs that are based on competing ARM processors, but it is good to see the company taking steps to leverage its foundry infrastructure, which is one of its crown jewels, into a new business model. It should bode well for the future of the stock.