This might be true for a few value-added parts (CPUs for Apple or Windows), and I can see companies having people knowledgeable about chip design on their staff but I don't see them developing chips internally. No system integrator wants to do chips themselves, but would rather let a few semi companies duke it out for low-margin sockets similar to how bums are paid to fight for alcohol or a few spare dollars. Motorola (Freescale) and Philips (NXP) all sold off their semiconductor divisions. Granted, those companies have been stinkers themselves, so maybe they aren't great examples. To give a better example, Cisco outsources many of their networking chips to Broadcom.
Semiconductors are capital-intensive businesses and will only get more so with the complexity required for more advanced nodes. Yes, the front-end design is not hard, but with complications like double-patterning, physical design has become more expensive and not a core competency for companies like Microsoft or Google, not to mention the added expense of CAD licenses.
That said, I wish this article were true. Semi companies have been sh*tty to work for the past decade, and I'd rather be eating M&Ms and gourmet food at Google than having my free soda and vacation taken away as has been done at my current employer.
Welcome system OEMs to the cut throat semi industry. Now, they will actual design the chips that meet all of the unrealistic performances and in a ridiculous time frame. Then productize these chips at a rock bottom price but still meet all insdustry standards for quality and reliability. Wake up?? We already lived the nightmare for sometime.
It occurs to me that a "side benefit" of the system house submitting their own database directly to the foundry is it would seem to be a trivial matter for them to "use the opportunity" to not only hide their own IP, but also to conceal unlicensed IP from other sources because the IP theft would be very easy to conceal and VERY difficult to detect at the chip level. It doesn't even have to be deliberate, even a legit design's IP done directly to silicon would be very hard for the "patent trolls" to litigate over, or for a competitor to "rip off". So there are really multiple reasons for this dynamic, and most all of them are in favor of seeing it accelerate in the future. I hope nobody thinks I'm being too cynical?
@Truc Giang: You may have a valid point here but I would have to differ and digress on one point in your comments: rock bottom prices. This was true for Semi Chip vendors but not any more for OEMs. They 'package' the individual chip components into the end product -bet it an iPhone... Android phone... iWhatEver. That is a clear advantage OEMs have -a real product!
Another point is on the teardown and costing of this by industry experts -most often when they estimate the BoM cost, they are STILL applying the Semi Chip vendor's pricing model (whether it is fabless or IDM). I don't think this is accurate in the case of OEM's. Short of knowing what Apple's internal pricing model looks like, these are best guesses for all it is worth. So your point about cheap components need further exploration.
I would also argue that there are differences in the operating models of OEM's chip development-to-production path vs. that of a fabless company. It is worthwhile to have some one dig up more on this... Or, the OEMs could suffer the same nightmare you allude to!
@Truc: Funny and too true. Chip makers have not only been doing the work of understanding the process tech, market and chip design--they have even been doing system level reference designs and much of the software stack to hand to OEMs who basically need to just decide to brand and have it manufactured. I feel your pain!
Well let's think through this. Now that chipmaking is past its days, there is consolidation in this sector for sure. And people figuring out system-making is better, more profitable will lead to a proliferation of system-makers. This will help start off a cost reduction spiral ultimately leading to commoditization at the system level, fulfilling the expectation of Truc. But this time the evolution cycle time for this to happen will be much faster, with low-cost crowd funding replacing classical VC funding.
I should have guessed this article would cause some good discussion when my CEO read a draft and choked on his coffee!
This "OEMs making their own chips" trend seemed innocuous to me, and it's been very obvious to me and my company's sales team that this is occurring.
I think there is sufficient evidence in the marketplace to claim that some of the most innovative consumer product companies are "re-verticalizing", at least for their most important products that require differentiation. What I don't have a clear answer for is, "Why?"
I have a hypothesis that it is actually the software that is driving systems companies to design their own chips. When I was at TI, we offered operating system board support packages and driver software along with our OMAP phone chips. Software was not a core competence (buzzword alert!) of TI, and it took many years and lots of money to do it sufficiently well. TI were experts on the chip, but not on software.
When we look at a company like Google, Facebook or Microsoft, these companies are experts at software, but are looking to create innovative battery-operated devices. If they buy merchant silicon, they have to buy a chip that was designed for no particular OS, tool chain, application or form factor in mind. If they design their own chip, they have total control over all these things as well as exclusive access to the end product.
I don't think every OEM will choose to design their own chips, only the ones that can get an advantage through innovation (higher pricing) or significantly lower costs. I imagine the economic hurdle rate to design one's own chip is quite huge. Apple and Microsoft have determined that some of their product lines meet this hurdle rate, and Google and Facebook may have, too.
In any case, we're lucky to be in an industry that innovates not only with technology, but also with new business models. It keeps all of us from being replaced with computers ;-)
Blog Doing Math in FPGAs Tom Burke 15 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...