Print This Post

Inside Apple and PA Semi Part 1: The Micro Look

PA Semi Last week, Apple made a tiny debit from their enormous cash reserves to buy small semi conductor design firm PA Semi.  At approximately $278m, the deal barely shifts the multi billion dollar cash account on the balance sheet; still it has left many analysts, watchers and writers asking questions.  The first and most obvious is “why?” Why did a company that favors buying companies in the early stages of development buy a company with an established customer list? And why given Apples’ focus and success with consumer-focused products choose to invest in the challenging and cyclical semiconductor industry?

Another more macro set of questions (in a companion post to this article) query whether the deal represents a possible shift in Apple’s acquisition policies.  Was this the start of a more acquisition friendly Apple; one interested in taking advantage of weaknesses in the current financial markets to fortify underlying tech assets?

Apple remains characteristically tight lipped on both the micro and macro points.  Spokesman Steve Dowling said only “Apple buys smaller technology companies from time to time, and we generally do not comment on our purposes and plans.”  In last week’s earnings conference call, CFO Peter Oppenheimer shared a similar perspective.

Without direct insight, we’re all left to speculate and make educated guesses.  This Part 1 of this two part article will look at the micro issues – the PA Semi buy and its logic.  Part 2 (which is found here) will look at the larger issue of Apple’s general M&A perspective.

PA Semi – so what’s the deal?  
Some are speculating this wasn’t a purchase driven by existing technology but instead aimed at acquiring intellectual capital.  The theory surmises Apple’s interested in acquiring IP and engineers to help them create their own customized lower-cost, lower-power consuming chip set for use in the iPhone or future devices.

Another angle of analysis theorizes Apple wants to take the chips PA Semi has already designed and make them exclusive components to Apple products.  Taking the chips off the market, and controlling supply, they say, will give Apple a significant competitive advantage.  It will make it harder for other firms to catch up, or duplicate Apple’s products.

However it’s sliced, the purchase seems to come down to one of these categories:  Buying product (as speculated by Forbes) or buying IP (Macworld) to build something new.

Palo Alto Semiconductor (PA Semi, for short) is a specialist “fables” microprocessor company. They aren’t in the business of manufacturing chips, they design them.  Before the acquisition put them on the global map, few outside the semiconductor industry, or involved in engineering devices likely knew them.   Semiconductor design firms aren’t usually household names.

The company was founded in 2003 by Dan Dobberpuhl, a former standout chip designer for Digital Equipment Corporation in the 90s.  Today, the company has a staff of about 150 and was well funded by Bessemer Ventures, Highland Capital Partners and Venrock.

PA Semi’s core technology is a series of high performance low-power chips named PWRficient processors.  As the name suggests, efficiency in electrical power consumption and resource allocation were principle parts of the design.  The company’s corporate white-papers suggest the chip family uses as little as 1/5th the power of comparable Intel Xeon or AMD Opteron X86 chips.

More than 50 patents were filed to support PA Semi’s specialized designs. The foundation on which their chips are built, however, is licensed.  PWRficient chips are built around the “Power Architecture” chip platform. This is the same architecture that powered pre-Intel Macs (Power PC chips).  It’s also the backbone of Cell Processors (the powerful chips custom created by Sony, Toshiba and IBM to power the PS3 gaming platform).

Apple exited the Power PC chip market in favor of Intel chips in 2006.  Intel’s processors were lower cost (as a result of manufacturing scale).  The Intel chips have also had significant impact on Mac Sales.   This last quarter, for example, Apple’s 36% record setting profit increase was powered in no small part by the sale of 2.29million computers, a 51% year over year improvement.

The question raised is – why does Apple want a company focused on a chip design that Apple already abandoned?  Especially given the decision to switch to “Intel Inside” has brought incredibly positive financial results.

The possible answers break into three categories.  Either Door Number  1. Apple has specific product plans for the technology (eg, they are buying tech); Door Number 2. Apple is buying Intellectual property in the form of smart engineers; or the neutral Door Number 3: Apple sees opportunity in getting both.

Speculation here has hinged on how PA Semi’s chips could be used to power Apple’s existing products.  To be sure, there’s no question chips that consume less power which allows portable devices greater battery life.  Another upside of drawing less power is these chips will produce less heat.  A third angle is sizing.   “System on a Chip” designs, especially those producing less heat, can be clustered together into a smaller space, thereby giving more design leeway for the final packaging of a product (itself, one of Apple’s signature elements).

New Macs in an ultra-portable format are one theory that’s been put out there.  The universal binary architecture that Apple currently uses could, presumably, be built around a PowerPC again.   Such a prospect doesn’t make much sense, however.  The switch to Intel is a significant factor in consumer shifts from PCs to Macs.   Additionally, Intel has shown a willingness to work with Apple to meet design constraints. For the MacBook Air, as a case in point, Intel made efforts to adapt chip sizes to fit the computers ultra thin form factor. For Apple to abandon a positive partnership seems beyond remote given the circumstances.

Another big speculation is the iPhone.  Low power “system on a chip” configurations could mean longer battery lives or smaller phones; wins in both categories.   Such a configuration could even theoretically power the kind of dual screen clamshell designs that have been patented and speculated on.   Adding to this buzz, an in house chip design could add a larger barrier to block copycat competitors from trying to either match or reverse engineer Apple’s hardware innovations and designs.

Tempting, even compelling, as the iPhone applications are, today, all of Apple’s iPhones and iPods are powered by ARM based processor technology.  The marketplace for these components is broad.  There is already a host of system on a chip and low-power chip vendors in (or entering) the market.   As much as exclusivity could be an advantage, and makes good copy, it seems likely to be too expensive to warrant a shift. A partial re-design? Or part of a basket of components? That would make more immediate sense.

[Note: All of the product assumptions built around existing chips assume the Power Architecture licenses underlying the PA Semi chips can be transferred to Apple. There's some speculation that might not be possible. For the sake of argument, and reviewing product applications, this presumes Apple can (and will) sort out any such IP issues.]

If not Macs or iPhones, then What?
The Mac and iPhone represent the biggest growth opportunities in the current product line, and therefore get the most attention but Apple’s got more in its hip pocket.  Could one of those products – Apple TV? Time Capsule? AirPort?  be the bigger beneficiary of this purchase?

It seems partly a question of whether this deal applies to existing products, derivative products or something new entirely.

PWRficient chips, it turns out, have advantages that go beyond just power savings.  One of the upsides is security.  The designs have integrated acceleration for encryption and decryption processing.  This could be advantageous for storing and securing data without taxing a device’s software.  For one application, imagine a smooth running password vault on your iPhone, more efficient device-based eCommerce.  Network storage on a home network could potentially benefit as well.

Another advantage of the chipset in PA Semi’s arsenal is the way the architecture can handle modular storage architecture.   This with security applications make for potential NAS products.  In fact, NEC and Mercury are already PA Semi customers.  Of course, Apple isn’t really in the enterprise storage market.  Besides the fact that the Xserve servers Apple makes are a niche product, they are on the Intel platform just like the Macs. It would be a surprise for that to change.

Given Apple’s strengths in consumer focused products, it’s hard to imagine they’re aiming to get deeper into the enterprise server market.  It’s possible, just doesn’t seem the best way to build on the momentum they have.   A more likely theory is that PA Semi technology could get mashed up into a new product that offers a broader array of home networking functionality. Maybe it takes the form of a newer, bigger sibling, to the Time Capsule products revealed earlier this year. 

Another natural extension might be a next generation Apple TV.  Currently the system runs an Intel 1.0 GHz "Crofton" Processor.  In its current configuration, which lacks a cooling fan, the device is quiet but known to run at high temperatures (about 110degrees when in use according to tests reported at Digital Trends)Apple has said this is normal and no cause for concern.  Still, a lower power chip architecture could conceivable address the issue in future models and a PWRficient chip set could do just that and improve storage architecture for the device too.  Another bonus, the chip can be attached to a larger than normal number of Input/Output (I/O) devices. That could potentially expand home media opportunities. 

Another class of product theories could look toward embedded applications or in a farfetched, but interesting, idea lead to some form of Apple gaming platform.  It would be “out of left field” but its not without at least one interesting element.  Turns out – Power Architecture based chips, those of the same fundamental foundation as PWRficient chips, power all of today’s home gaming consoles.

•The Xenon chip in the Xbox 360 is a 3 core Power PC chip
•The Broadway chip in the Wii is a slower Power PC derived chip reported to run at about 729Mhz.
•The Cell Processor in the Sony PS3 was a custom developed evolution of the PowerPC built by Sony, Toshiba and IBM.  

An Apple game platform is far from likely, but it makes for amusing speculation.

The second theory of the deal is an engineering purchase, an IP buy.  It’s equally plausible. Maybe even more so if Power Architecture licenses prove non transferable.

PA Semi features a headcount of approximately 150 talented engineers.  The group includes specialists skilled in processor design, software engineering, ASIC and system architectures.  Many of these engineers, including some of the firm’s top tier executives, have tremendous experience.

As it stands, Apple engineers already are often tasked to create custom motherboards and system architectures in order to allow products to fit Apple’s external (or usability) design criteria.

There’s no question, adding a large group of talented staff to the pool of resources already in Cupertino would be valuable.    An in-house semiconductor design team could both help Apple work with existing suppliers to customize chips for exclusive use, or potentially, build chips from the ground up.  

Engineer talent? Product Enhancements? New Chips or Re-engineering old? Where is Apple headed?

The safe answer is to say it’s a little of both, or a lot of both.  The safe, diplomatic prediction is “flexibility.”  But as an experienced former Silicon Valley CEO was quick to point out, “such perfect synergies never happen.”  To his eye, this will be an either/or proposition with only a trickle of overlap.

A new Apple TV? A new storage product? Designers to build for the future?  Reality is, where Apple is headed with PA Semi is a mystery (probably just as Apple wants it).   Whenever, or whatever, though, one thing is actually somewhat certain – the deal will figure prominently into Apple’s plans.   Steve Jobs not only had a direct hand in negotiating the deal, according to reports, many of the final meetings actually took place in his own home.  The purchase of PA Semi wasn’t an arm’s length transaction.  This was the first step in a friendship.    

[More on the Macro question of whether PA Semi is a sign of changing M&A policies at Apple can be found in part two of this article, here]


Related Articles
Apple Q2 Earnings: The Oppenheimer Effect
More iPhone Mania: Sales Suggest Second Phone?
iPhone 2: June or Sooner?
$100m KPCB iPhone Fund
•iPhone and Blackberry Battle for Mobile Markets
Apple Q1 2008 Earnings
Apple Q4 2007 Earnings
Apple Macworld Announcements
The New Macbook Air: A first look
Top Ten Beatles Songs for their iTunes Debut (if it happens)

Comments are closed.