How Intel will fare in the post-Otellini era
The retirement of Paul Otellini as CEO of the world's largest semiconductor company will mark a big shift for Otellini, but Intel's path will likely continue in its current direction, a direction he helped shape. As the industry shifted from high performance desktop PCs to handheld mobile smart devices, Otellini drove Intel to focus on efficiency over performance. Ultrabooks, tablets and mobile phones now ship with Intel inside, something that would have seemed impossible when Otellini first took the helm.
As the first non-engineer to run Intel, Otellini brought much-needed marketing chops gained from an earlier stint as Intel's Executive VP of sales and marketing. When he took the CEO job, Otellini had already been steeped in the technical intricacies of CPU product design as head of the Intel Architecture Group. During that time, Intel produced some of the highest clock speed desktop CPUs on the market under the Netburst brand. Netburst CPUs were also notorious for being extremely power hungry.
Indeed, Otellini realized from the very beginning of his tenure that high clock speeds and the biggest, baddest CPUs wouldn't lead the world's largest semiconductor company forward. From the very beginning of his stint as top dog, Otellini was more focused on product positioning than product science—a philosophy that perfectly suited his roots in not engineering but rather (gulp) sales and marketing.
His lack of engineering background didn't prevent Intel from sending him to take charge of the as general manager of the Intel Architecture Group starting in 1998, about as engineering focused as it gets at Intel. It was a surprise move for the executive who had been Executive VP of sales and marketing. During that time, Intel produced some of the highest clock speed desktop CPUs on the market under the Netburst brand. Netburst CPUs were also notorious for being extremely power hungry.
Otellini became CEO of Intel in 2005, and that began the steady shift away from extreme clock speeds towards efficiency. At the time, a group of designers at Intel's Israeli facility had built a new CPU for laptops called the Pentium M. It was a somewhat subversive effort, and outside Intel's mainstream architecture group. Otellini took notice of the Pentium M's success, since Netburst was starting to bump up against some serious power constraints. At the time, Intel's chief competitor, AMD, had started shipping 64-bit processors that outperformed all but the very highest frequency Netburst-based processors, while using substantially less power.
Under Otellini's charge, Intel outward focus shifted away from producing power-hungry PC processors with high clock speeds to building power efficient processors suitable for mobile PCs, tablets and smart phones. By early 2006, the next Pentium M CPU, whose name had changed to Core, hit the market. The shift to more efficient processors allowed Apple to adopt Intel even as its own G series CPUs began running out of steam. Apple had been unwilling to use the hotter Netburst CPUs, which weren't a good fit for Apple's minimalist designs.
The final nail in Netburst's coffin arrived when Intel shipped the first desktop Core 2 Duo CPUs in the second half of 2006. At the time, raw CPU performance was still the watchword, but Core 2 heralded the beginning of the shift to power efficiency instead of pure performance. Today, Intel's various design groups put efficiency, as measured in performance per watt, ahead of more straightforward performance metrics. As Intel's Per Hammarlund noted at a session on Intel's upcoming Haswell CPU during the 2012 Intel Developer Forum, new CPU features are considered only if they don't consume more power, or at least, increase the performance per watt.
During Otellini's seven year tenure, Intel also honed its manufacturing chops, pushing its process technologies down to 22nm, and soon, 14nm. These are the highest density mainstream semiconductor processes in existence today, and Intels' manufacturing chops have as much to do with the company's success as its architectural designs. On the architectural side, Intel began building system-on-chip products suitable for tablets and high end smart phones, well aware that these mobile smart devices inevitably cannibalize PC sales.
Otellini's tenure hasn't been without its stumbles, most of them legal in nature. The company went through a messy and public legal war with AMD over Intel's hardball marketing tactics, which was finally settled in 2009 when Intel paid out $1.25 billion to its smaller competitor. Intel also paid Transmeta $150 million, plus an ongoing annual $20 million payment to settle a patent dispute. Other legal disputes outside the US, including Japan and the EU, have focused on Intel's aggressive marketing.
It's unlikely that Otellini's departure will signal any substantial changes. Legal wrangles aside, Intel under Otellini saw steady revenue and earnings growth, even as Intel aggressively implemented new, higher density manufacturing processes that cost billions to implement. The future of all tech products lie with lower power consumption and increasing mobility, and it's unlikely we'll see Intel shift away from that focus. If anything, a new CEO will likely push Intel harder towards mobile designs, with desktop processors borrowing technology from the low power side rather than vice versa.
More interesting will be how Intel's sales and marketing will change. Relatively few people care about the CPU inside their smart phone or tablet, and it's unlikely that any Intel marketing campaign will make end users care. Add in the fact that even high end smart phones tend to be built with low cost, commodity chips. That means Intel will need to compete on price, something it's never been fond of doing. On top of that, it doesn't really have the marketing leverage on the mobile side that it has in the PC business, so potential OEM customers don't care about issues like x86 compatibility or the chipset ecosystem. What phone OEMs want is price and guaranteed delivery. Intel's manufacturing muscle may play well, giving Intel marketers some leverage. But it's up in the air, and with ARM owning the lion's share of the smart device market, Intel has a long uphill battle. Will it be able to push PC technology down far enough so that it can compete on price and power, even as revenues from the PC side continue to slide? That's the conundrum the next Intel CEO will need to face, and it won't be an easy puzzle to solve.