Apple the Platform Company
Over the past 10 years, systems companies have trended away from chip development for two simple reasons: 1) design and fabrication costs – too expensive unless amortized across huge volumes 2) semiconductors are a saturated market (lots of “merchant” chip players fighting for business).
In 2007 we saw an example of this play out in mobile, when Nokia divested its chip designers and wireless IP to STMicro. Ironically, Apple then moved in the opposite direction by buying a chip company (PA Semi).
Why would Apple do this? (effectively deciding to compete against the world’s top chip makers – Intel, Qualcomm, TI, Infineon, etc). Let’s analyze in the context of hardware, software, platform control, buyer/seller bargaining power, and marketing.
Hardware: Apple is celebrated for its minimalistic approach to hardware design. This runs contrary to the incentives of chip companies, who benefit from a “kitchen sink” approach – i.e. integrating every conceivable feature so as to sell to as many customers as possible. By designing its own chips, Apple avoids paying for unused silicon (saves chip size, cost & power). They can also optimize for their minimum feature set (e.g. video at 720p vs 1080p) and focus on custom cell logic which maximizes performance per watt.
Software: a BIG potential win for Apple and here’s why: As SoCs (system-on-chips) increasingly go multicore, hardware companies face stiff parallel software challenges. It will take years for software to adapt to dual-core models, and until it does, there will only be a ~30-40% gain in performance from offloading background tasks to the second CPU core, versus a theoretical improvement of 100%. Why the big difference? Only the best programmers understand parallel-compute structures. Apple can wield tremendous advantage here by having its own chip and software guys collaborate – Apple pioneered parallel software efforts in GCD and OpenCL.
Platform Control: Key to Apple’s platform strategy is control. By owning its own chips, Apple can put in special hooks so people can’t access hardware internals – e.g. they can design more fool-proof DRM to instill confidence for Hollywood, and they can tighten security holes to keep out jailbreakers/hackers. If they use outside chip vendors, they can’t fully control these aspects of their platform.
Buyer / Seller Bargaining Power: Seller and buyer power are largely driven by a company’s market share. As Apple’s share increases in mobile they wield unbelievable power over suppliers. Component vendors are falling over themselves to get inside iPhone/iPad (I wrote about Intel here). Apple’s agreement with Samsung (Fab partner for A4) allows it to benefit from massive scale (Samsung is #2 chip-maker worldwide), get quick-turns, test chips, and best pricing. Samsung does this because Apple will continue to buy flash memory in massive volumes (so this ultimately ties back to platform halo effect). Since A4 is ARM-based (ubiquitous in mobile), Apple avoids any peril of building a platform around a proprietary/unpopular architecture (e.g. the 90’s battle of Mac/PowerPC vs PCs/Intel).
Marketing: Steve Jobs made it no secret in calling out Apple’s A4 by name during the iPad launch. Why did he do this? (unusual for Apple to advertise the “guts” of a product). Answer is simple – Apple wants consumers to believe that by building its own chips, its products will have a noticeable edge in performance and battery life against the competition. This is subtle but very important. Advertise it, the Apple fanboys will write about it, and consumers will pay extra.
So the move to vertically integrate chip development helps Apple erect barriers and become a dominant platform company. This spells a larger trend – it is no longer adequate to simply be a device or software company to succeed.