← Acquired

P.A. Semi + AuthenTec

Episode 2017-01-242017-01-24EnglishMembers only
P.A. Semi + AuthenTec
Listen to this episode00:00 / 1:08:15
Reading settings
Reading preferences
Text size
Download in-depth read
Read offline

English · In-depth read · Current version

Downloads are not available yet.

Text language

In-depth read · Key ideas, reasoning, and examples, organized by topic.

§01Two acquisitions, one thesis, and a warning about expertise

Acquired's thirtieth episode pairs Apple's purchases of P.A. Semi and AuthenTec, timed to the tenth anniversary of the iPhone. Ben Gilbert frames the pairing as a confluence: the two deals produced the A-series chips and the TouchID sensor, features that work together and that Apple has used to differentiate its devices.

The hosts open with an explicit disclaimer that matters for how the rest of the episode should be read: neither is a semiconductor expert, David says they are playing experts on television, and he invites listeners who know the field to correct them on a later show. The episode's value is in the deal logic and the history, not in chip architecture detail.

Ben then supplies the minimum technical background the episode needs. Desktops, laptops and servers have used Intel's x86 architecture; ARM popularized a different instruction set that is more restricted but far lower power. Intel has tried low-power parts, with Atom as the example, but ARM became dominant in phones and tablets because mobile devices have strict power budgets when they cannot be plugged into a wall.

§02Dan Dobberpuhl: the textbook, StrongARM, and SiByte

P.A. Semi's origin story is really a founder story. David traces the company to Dan Dobberpuhl, whom he describes as a celebrity in chip design: a lead designer at DEC in the 1970s and 1980s, author of a widely used textbook on chip design, and later the person who built DEC's California team for the StrongARM processors — higher-performance versions of ARM-based parts aimed at mobile devices in the mid-1990s. The attempt did not work commercially and the division was sold to Intel, where it became XScale and powered many of RIM's BlackBerries.

Dobberpuhl's reputation, in David's telling, is specifically about pushing the limits of the power-to-performance tradeoff — the exact axis on which mobile chips would later be judged. That framing matters because it explains why Apple would eventually want the team rather than the product.

After leaving Intel in the late 1990s, Dobberpuhl founded SiByte, which made chips for routers and networking equipment during the internet build-out, when the largest chip market was Cisco and 3Com gear. SiByte was acquired by Broadcom in late 2000 for over $2 billion before shipping a product. David's detail here is the sharpest illustration of reputation as an asset: Broadcom's CEO had encountered Dobberpuhl through his college textbook and said SiByte had the best engineering team he had ever seen.

Palo Alto Semiconductor was founded in 2003, initially under that name, and moved to Santa Clara shortly after — which is why the company kept the P.A. Semi name rather than renaming itself. It raised from Bessemer, Venrock and Highland, with Texas Instruments later investing heavily, and it built very low-power, high-performance chips on the Power instruction set rather than ARM.

§03The WWDC 2005 collision

The episode's best narrative beat is the collision between P.A. Semi's sales effort and Apple's Intel transition. At WWDC 2005 Steve Jobs announced that the Mac had been running a secret second life: because of the NeXT acquisition, Mac OS had been able to run on Intel chips all along, so Apple could switch whenever strategy required it. Ben notes the theatrical reveal — that shipped Mac OS X versions already contained universal binaries capable of running on x86, and Jobs told the audience they already had it and Apple had simply disabled it.

Keep reading as a member.

You've read the free preview. Become a member to read the rest, with side-by-side reading and downloads where available.