|Overview of the Foundry Model|
|Trusted sources we've spoken to in the semiconductor industry have implied that TSMC is considering a partnership with Apple that would realign the manufacturer's technology roadmap and fundamentally alter the balance of power between the foundry and its other customers. Morris Chang, TSMC's CEO, spoke about the possibility of closer collaboration with its customers in general terms last Friday, but at the time we thought the likelihood of an Apple alliance was unlikely. We've since been given reason to think otherwise.
A dedicated alliance with Apple that gives the company first access to 20nm production and/or a dedicated fab could fundamentally redefine the foundry-customer relationship and have serious impact on both TSMC's other customers and its competitors.
Here's what Chang actually said. The question came from Dan Heyler, an analyst with Bank of America.
Q: (Heyler) Did you think you need to dedicate some times to product specific areas that are very high volume?
The Fracturing Foundry Model
The easiest way to explain why Chang's remarks are so important is to recap the evolution of what we call the foundry model. From the 1960s to the late 1980s, the overwhelming majority of semiconductor design firms owned their own production facilties. This changed over time as manufacturing techniques became more standardized and the industry as a whole synchronized with what came to be known as Moore's Law. The rising cost of node deployment and the rapid, capital-intensive factory upgrades that were required led IDMs (Integrated Device Manufactures) to seek ways to outsource manufacturing and design. Thus was born the foundry model. Small-to-medium companies that couldn't afford the high capital expense of keeping up with Intel or IBM would be able to pay a "pure-play" foundry to handle their manufacturing. By aggregating the business of many manufacturers, the foundry would be able to afford to keep up with the few remaining vertically integrated IDMs. TSMC, founded in 1987, was the first pure-play foundry.
The foundry model worked well for twenty years because technology ramps were regular, power consumption and die size scaled consistently, technology roadmaps were unambiguous, and higher production costs were offset by increased yield per wafer. For TSMC, that began to change at 40nm. The 28nm ramp has been smoother than 40nm's legendary problems, but it still took the foundry nearly a year to ramp 28nm volume production, with Chang predicting that the foundry would "fully meet" demand for 28nm products in Q1 2013. All current indications are that the 20nm ramp will be at least as difficult; Chang told investors he expects pilot 20nm production in 2013 to be "very small" with 2014 a "ramp year" for 20nm SoC production.
Past and estimated future costs of next-generation nodes. Original image courtesy of EETimes
The high cost and long ramp time of new nodes have led companies like Nvidia to call for increased cooperation and risk-sharing on new production. On the one hand, TSMC appears to be acknowledging that its customers are concerned that the old foundry model is no longer able to properly address scaling challenges. A direct partnership with Apple in which the latter gains early, guaranteed access to 20nm production in exchange for a substantial amount of up-front financing, however, probably isn't what anyone else had in mind.
|An Apple/TSMC Deal Makes Sense -- But Disrupts Everybody Else|
|From Apple's perspective, a deal like this makes a tremendous amount of sense. Apple and Samsung clearly haven't mended their fences; the enmity between the two companies' is high enough that we're surprised they haven't broken manufacturing ties already. The Cupertino-based company can easily afford a massive investment in TSMC, particularly when said cash would buy access to a next-generation process node and a production guarantee. Apple has a long history as an early adopter of technology and strongly prefers to launch products based on that perception.
As for TSMC, an agreement of this sort would win it additional capital for investment and possibly accelerate its node deployments. Investors are clearly concerned about the amount of money the company spends on capital expenditures, and while Chang handled the questions adroitly, the foundry plans to introduce FinFET designs at 16nm and EUV at the 10nm node. Both of these technologies are significant changes that could prove difficult to adopt -- EUV (Extreme UltraViolet lithography) remains deeply problematic even after years of research.
Prioritizing Apple Would Change Everything
For everyone else who uses TSMC, this type of arrangement between the foundry and Apple would be an ugly kick in the guts. TSMC would undoubtedly try to downplay the impact of sudden demotion to second-string by offering more attractive terms or by claiming that Apple's investment would allow the company to accelerate deployment to such a degree that the impact would be negligible. Such arguments are unlikely to fly far given the current state of competition, in which Qualcomm, Samsung, Texas Instruments, and a number of other companies are fighting Apple tooth-and-nail.
Ultimately, TSMC's biggest customers may find themselves short of options. Moving designs from one foundry to another takes a considerable amount of time and effort might not solve the problem. If a combined TSMC+Apple definitively pulls ahead of GlobalFoundries on node deployments, opting for the latter would do Qualcomm little good. All the same, there have been rumors of shifting partnerships in the foundry world for months -- Qualcomm has reportedly tapped GlobalFoundries for future 28nm designs, Nvidia and Samsung may be working on a Tegra device, and AMD apparently settled on GloFo for its 28nm follow-up to Trinity, codenamed Kaveri. Kabini -- the 28nm follow-up to Brazos -- is less certain.
Rumors and questions about what Apple might do with its record-setting cash reserves have floated for months, and an investment of this sort could be one of the best ways the company could put those assets to work. It fits the company's ruthless competitive nature, its desire for a technological edge, and its emphasis on guaranteed availability. It's the exact opposite of the equal access arrangement the foundry model supposedly enabled, but for companies trying to catch Intel's manufacturing lead, it may be the only way to do so.
Don't expect any announcements in the next week, or even next month -- Apple is notoriously secretive, and TSMC will want to float this idea slowly, with probable implementation at the 20nm node rather than the current 28. Don't be surprised, however, if the number of companies exploring partnerships with non-traditional partners continues to rise. In today's cutthroat market, the risk of being 3-6 months behind Apple at a given node are too high for most companies to take -- even if the offer comes with lower costs and better yields.