Login

What happened to AMD?

AMD shadow ceo rory read stock 1020

There's blood in the water, and the sharks are circling around. AMD is in trouble, to the point where buyout rumors sound credible even when they turn out to be false. AMD's having trouble putting its processors into computers people want to buy. The company's losing money and some key personnel are abandoning ship. Under new management and after two rounds of layoffs, the stock price is practically as low as it's ever been, except once in late 2008 when the company was struggling to rid itself of tremendous debt amid the global recession.

It wasn't always this way.

Seven years ago, AMD processors ran circles around the Intel chips in their price range. AMD had arguably surpassed Intel in engineeringBy 2005, AMD had not merely become competitive, it had arguably surpassed Intel in engineering chips. With the Athlon 64, AMD led consumers to 64-bit processing, and the company led the charge to dual-core processors shortly thereafter. And when AMD bought ATI in 2006 for $5.4 billion, it didn't merely acquire a graphics card business and some talented employees, but a host of new markets as well: smartphone graphics, game console graphics, and the ability to build a new type of hybrid CPU-GPU processor that AMD dubbed the "Fusion APU." AMD put chips into the Xbox 360, the Nintendo Wii, and a host of Dell computers.

5870_beauty_full__1__copy

It was the heyday of AMD, but not the only period of prosperity in recent years. In late 2009, AMD managed to obtain a $1.25 billion settlement from Intel over anti-competitive practices, and launched GPUs which all but leapfrogged rival Nvidia. In 2010, Sony starting putting AMD chips into VAIO systems, and in 2011, Apple tapped AMD for graphics in both laptops and desktops. After three years of denying any interest in the netbook market, AMD swooped in hard, finally launching those AMD Fusion APUs in a variety of inexpensive laptops and selling millions and millions. In the desktop market, Intel's Core 2 and Core i series managed to keep a performance edge, but they allowed AMD to compete at lower price points and typically offer more cores for the money.

However, the days when AMD was firing on all cylinders are slipping away. The question is why.

PC industry decline

AMD management primarily blames the macroeconomic situation, which roughly translates to "the PC industry isn't doing as well as it used to." Whether due to a global recession, or component shortages from flooding in Thailand, or people putting more money towards tablets and smartphones, there's a number of factors in play... not to mention a drop in spending leading up to the launch of Windows 8, where consumers and companies may have been waiting to buy computers so their purchase wouldn't become obsolete. AMD's management is right to point out these industry-wide setbacks. But that isn't the whole story.

2012-06-06_19-44-45verge-1

Execution woes

Primarily, AMD has stumbled, repeatedly, during recent attempts to combat Intel. The company released a host of products, each with an unfortunate issue. While AMD's Opteron server processors may feature in the fastest supercomputer in the world, AMD's Bulldozer architecture for those chips failed to impress in either the server or the consumer desktop realm. When Nvidia popularized switchable graphics in notebooks, AMD couldn't piece together a compelling competing solution quickly enough. Several missed passes in a row When the company shipped its Llano chips, they didn't sell well, partly because they required a new motherboard chipset that AMD didn't allocate properly. As a result, the company has a huge surplus of Llano, which it can't easily sell because the successor platform, Trinity, is already on the market. AMD's Trinity is in trouble too, though. It wasn't ready in time to compete with Intel's Ivy Bridge, so notebook manufacturers designed the current wave of thin laptops and Windows 8 systems with Intel in mind, and AMD more of an afterthought.

Intel dominance

Meanwhile, while AMD hasn't been able to execute properly, Intel is a well-oiled machine. The company's not merely coming out with chip after chip on its famous "tick-tock" cadence (re-design, then shrink) but actually adding new technology each and every year. With Ivy Bridge, the company improved its integrated graphics, encroaching on AMD's APU turf, while keeping CPU performance far enough ahead to keep Trinity looking like the underdog. In case you're curious, Trinity still seems to be ahead in graphics performance and battery life, but crucially, not both at the same time. Reportedly, that integrated GPU is rather power-hungry. Intel also incentivizes manufacturers to use its own chips, of course, providing hundreds of millions of dollars in marketing and design assistance for its ultrabook standard, while AMD won't or can't afford that kind of effort.

2012-06-06_19-48-22verge

Branding confusion

AMD dismantled all of its old branding over the past couple of years, getting rid of the ATI moniker entirely in 2010, and replacing the Athlon, Phenom and Sempron lines with a series of alphanumerics like C-50, Z-60 and A8-3850 instead. Since then, the company's pursued branding initiatives like "AMD Vision" and built a series of buzzwords and product codenames that are even more difficult to follow than the CPU and GPU brands we've already listed. Intel naming schemes aren't exactly consumer-friendly either, but they make a certain degree of sense, while AMD traded names that people knew for ones that felt arbitrary by contrast.

Leadership struggles

AMD's management team has been in flux for quite a while. CEO Hector Ruiz stepped down in July 2008 in the middle of a series of layoffs, and was later implicated in an insider trading scandal that caused him to step down from AMD spin-off Globalfoundries in late 2009. In January 2010, the board of directors ousted CEO Dirk Meyer because he didn't have an aggressive enough vision when it came to tablets and mobile devices. CFO Thomas Siefert was acting CEO for seven months until a replacement was found. Siefert resigned in September, CFO once again, and he took a chunk of AMD's share price with him. Former Lenovo President and COO Rory Read took over as CEO in August of last year. The rapid turnover can't have inspired confidence in the company, but there's more to it than that.

"Ignoring" the netbook platform since two-thousand eight

We've heard rumblings from inside AMD that the collapse has been like a chain of dominoes ever since Dirk Meyer left, particularly since it was Meyer who was credited with the company's return to profitability a few years back. The company has been in billions of dollars of debt since the $5.4 billion ATI purchase, and AMD later admitted that it overpaid for the graphics firm. Under Meyer, the company spun off Globalfoundries, and sold off businesses, including a digital TV chip business to Broadcom and its mobile graphics business to Qualcomm.

However, a source within AMD tells us that while Meyer was good for the company, he was indeed complacent, and in a rather public way. In 2008, Meyer said AMD was "ignoring the netbook platform," and the company wound up late on the scene. Qualcomm also turned that mobile graphics business into the Adreno graphics part of its profitable Snapdragon platform (and recently hired away AMD's Eric Demers to run it) after AMD sold it for a (relative) song, but perhaps hindsight is 20/20.

2012-10-29_01-42-04verge

However, that same AMD insider says it was a lack of a leader, not just a lack of leadership, that led to AMD's stumbles. While he characterized Rory Read as a zealot on a crusade — pushing out AMD's old guard in favor of like-minded individuals blindly marching towards innovations like modular chips — he doesn't blame Read for most of the issues that AMD has seen. "lots of people just floating in a purgatory place" Rather, there were "disconnects with leadership" due to the seven-month gap between Meyer's ousting and the time Read arrived at AMD. Our source characterized it as "lots of people just floating in a purgatory place." The AMD insider believes that AMD's current strategy to focus more on the server, embedded and gaming markets is actually a good idea, and talked up the company's work on building modular chips, but believes that AMD might not last long enough to realize those goals without a better short-term plan.

The cash crunch

Last month, analysts had the same worry. They cited AMD's shrinking cash compared to its expenses and over $2 billion in remaining debt, but both the company and other analysts insist that there are still options. Though AMD has fully spun out Globalfoundries, it still has a symbiotic relationship with the company due to its orders of chips, and it may have the power to renegotiate its contracts since Globalfoundries would be adversely affected if AMD had issues. Sterne Agee analyst Vijay Rakesh also issued an investor note this week suggesting (among other things) that AMD could sell or leaseback its campus, and that AMD could sell or license some of an estimated (by Sterne Agee) $2.2 billion worth of patents in order to raise money.

Getting back in the saddle

In some ways, that gets us right back where we started, though: AMD weighing difficult options including selling off important assets — intellectual property — that it claims will be critical for future business strategies. AMD wants to lead the market for microservers using new 64-bit ARM cores, and offer up pick-what-you-like custom chips for embedded applications in the long term. How does AMD turn things around until those businesses can take off?

Nathan Brookwood, an analyst at semiconductor consulting firm Insight 64, said that it might just require running the traditional business without mistakes for a while. "The execution miscues, it's like watching the Raiders game the other day. Just when you think it's going to be there, there's an interception," he told The Verge. "What AMD really needs to do over the next several quarters is to execute without mistakes, and if so there's a clear opportunity for them to regain some of their physical health," he said.

2012-10-29_01-52-17verge-1

The last time AMD was in dire straits, Brookwood recalled, an acquisition was what got the company out. In 1995, when its Pentium competitor didn't meet expectations, AMD bought Nexgen to acquire talent that helped it design future processors. If AMD plans to do more with ARM processors, that strategy might already be helping them again: In August, AMD brought back Jim Keller, who led development of the Apple A4 and A5 SOCs. Also, though many people have left, it's not like the company is running out of people. AMD still has around 10,000 employees. It's a big firm.

Brookwood also told us that while he takes executives at their word that they aren't actively looking for a buyer, he believes AMD has some definite value to the right company. There's some speculation that AMD would be difficult for another company to purchase because of Intel lawyers, who would take it as an opportunity to terminate the special x86 license that allows AMD to produce chips. Brookwood pointed out that Intel also relies on a license to AMD's 64-bit extensions to produce its own processors. It's a mutually beneficial, mutually destructive relationship, and Brookwood believes it could protect any theoretical buyer of the company.

So now, we wait and see whether AMD can deliver compelling enough products, soon enough, to get it out of this ditch... or if some benefactor decides it wants to compete with Intel and Nvidia by paying for the privilege.

The Verge
X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Spinner

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.