As the world’s leading maker of CPUs for home computers and servers, Intel seems as though it can do no wrong. Yet, in 2001, Intel made one of the most costly stumbles in microprocessor history when it launched the Itanium high performance CPU.
Itanium: The Supercomputer that Wasn't
The product of seven years of development and a budget that exceeded $2 billion, Itanium was billed as the ultimate processor for supercomputing and enterprise service applications. And although it was certainly one of the more capable chips on the market, it failed to meet the needs of most customers. For some, it was too expensive. For others, it was too difficult to optimize.
Long before Itanium hit the market, some members of Intel’s senior management team had begun to question the project. “This will end up being one of the world’s worst investments,” warned David House, former Intel chief of corporate strategy. According to Nicholas Blachford, a technical writer specializing in CPU design,
“Intel’s implementation was very bad. Intel screwed up badly during development and instead of a low cost ultra fast CPU, ended up with a massively complex and hideously expensive beast of a CPU which didn’t give the expected stellar performance.”
Moreover, with Itanium, Intel had abandoned the backward compatibility that had been a cornerstone of the company’s strategy for more than 20 years. As a result, potential customers had to rewrite their software to run on the new platform.
When I first wrote about Itanium in 2002, few programs were capable of running on the processor. Poor real world performance further doomed Itanium to marginal market share relative to its direct competitors, such as the DEC Alpha. In 2001, only 500 servers worldwide incorporated the new chip.
Itanium has improved in recent years, with the release of Itanium 2 and other derivatives, reaching 200,000 units per year by 2007. Today, Intel is far from abandoning the design. The latest version, known as Tukwila, employs a four-core processor design and a highly efficient 65nm manufacturing process. Still, one can hardly view Itanium as anything but a financial and PR disaster for Intel.
By 2008, Intel was languishing. Net income had dropped 24% and the company was forced to cut its payroll by 20,000 (nearly 20% of its workforce). Nevertheless, Intel had deep pockets, deep enough, in fact, to not only weather the economic downturn, but to create a whole new line of highly innovative products.
Atom to the Rescue
One of those new products was the Atom microprocessor. Atom was the converse of Itanium in almost every way.
The company first hired ethnographers to study how people actually use their products in real life. They quickly learned that basic personal computers had advanced beyond the needs of many customers, who used their computers primarily for browsing the web, reading emails, and basic word processing. Instead of creating more powerful processors, Intel’s marketing team proposed to build a cheap, energy efficient chip that could power small Internet devices for casual users. Although many of the company’s engineers remained skeptical, Intel CEO Paul Otellini liked the concept.
The first challenge was to convince the company’s engineers that building a less powerful processor was worth the time and effort. Most engineers adhered to Moore’s Law, and viewed the path to victory as an endless “space race” with rivals like AMD, IBM, and Sun Microsystems. However, with Otellini’s backing, Atom was a done deal.
In early 2008, Intel launched the Atom processor as the CPU of choice for a new line of low priced, ultra-portable computers known as netbooks. But instead of engaging in a traditional advertising campaign, Intel focused its marketing efforts on educating retailers and users on the functions and benefits of netbooks.
It is one of the key lessons we discuss in the book. Namely, “if a product is really new or a radical innovation, promotional efforts should focus on educating the public about the product and its benefits.” It proved successful for Nintendo when it introduced the DS and Wii, and it has proved equally successful for Intel.
By the end of 2008, consumers were beginning to warm up to the idea that less was more when it came to netbooks. In real world terms, lower price and less power translated into greater portability and longer battery life. The real success of the netbook, however, did not come until 2009. Despite the economic downturn (or perhaps because of it), sales of ultra-portables have exceeded Intel’s wildest expectations.
“The time is right for the netbook,” according to ABI Research.
A confluence of social and technological factors has created a kind of “perfect storm” that will lead to a market explosion for netbooks over the next few years. The firm forecasts worldwide shipments of nearly 35 million this year, rising to an estimated 139 million in 2013.
For Intel, however, netbooks are only the beginning. The chip is now finding its way into an array of devices. Today, companies are coming to Intel with ideas for the Atom that Intel never dreamed of, from Internet connected sewing machines to “smart” farm tractors. The chip is already powering a host of devices, including ATMs, smartphones, and vehicle navigation systems.
The Atom processor may soon find its way into "smart" tractors
One the surface, Atom looks like the microprocessor equivalent to the Nintendo Wii, a simple gaming console that speaks to needs of the average person. However, unlike Nintendo, Intel is not standing still. While Nintendo seems comfortable with the success of the Wii and DS, Intel is developing new lines of products, each suited to the unique needs of particular market segments.
Intel continues to produce some of the most powerful processors on the planet, with applications ranging from advanced weather forecasting to quantum computing at Los Alamos National Laboratory. It is launching efficient 32-nm i5 processors for high performance laptops (The i5 is capable of running four threads simultaneously and employs a new proprietary design known as Turbo Boost). And Intel continues to release improved versions of Atom in a continual effort to fend off competition in the netbook and mobile device markets. On December 17, 2009, Intel announced a total of 17 new processors, more than enough to keep the company at the forefront of the microprocessor market for the foreseeable future.
Prof. 
