On Sunday, Pride celebrations shut down a rainbow swath of San Francisco. In the shadow of the city’s iconic Coit Tower, chipmaker Intel held a nerdier and more select party.
At the five-hour event, 100 attendees from startups, venture capital, and tech giants drank in semiconductor-themed cocktails and detailed explanations of how sand is processed into silicon chips. It was a celebration of how exponential upgrades from the chip industry have propelled progress in technology and society over the past 50 years—and an argument that the party’s not over.
“It's going to keep going,” said Jim Keller, a semiconductor rock star who joined Intel last year as senior vice president of silicon engineering, and a cohost of the event. “Moore’s law is relentless,” he added, referring to the 54-year-old assertion by a former Intel CEO that the number of transistors that could be fit onto a silicon chip would double on a predictable schedule.
Sunday’s event was aimed at making clear that storied, troubled Intel can still deliver huge growth in computing power as the industry has provided in the past half-century. Keller, a chip industry veteran whose silicon creations have helped shift the trajectories of Apple and Tesla, joined Intel at the end of a rough decade. It missed out on the market supplying chips for mobile devices. Those pocket-sized gadgets undermined sales of PCs, where Intel had a de facto monopoly.
Intel still dominates the market for server chips that power cloud computing, but its two most recent generations of chip technology arrived late. In April, the company announced it was abandoning work on chips for 5G wireless devices, walking away from the next big wave in mobile technology, and a deal that placed Intel modems in some iPhones. The following month, Intel told investors that it expects profit margins to fall over the next two years.
Those concerns weren’t much discussed Sunday, where the focus was on the history, and the future, of technology. Intel staff stood by microscopes where the curious could peer at the tiny sculptures that are modern transistors, capable of toggling electrical currents on and off billions of times per second. In addition to Keller, speakers included Raja Koduri, Intel's chief architect, and Mike Mayberry, the company’s chief technology officer. Koduri has said he helped recruit Keller, whom he knew from when they both worked at Apple, to Intel last year.
The story of computing is bound up with Intel and Moore’s law. For decades, Intel maintained the doubling pace by inventing new materials, processing techniques, and designs for ever-smaller transistors. More recently, the pace slowed and the futures of Intel and computing have appeared less tightly bound.
The latest generation of technology the company has on the market, known as 14 nanometer, was around a year late when it fully launched in 2015. The next generation, 10 nanometer, has also missed its original schedule. Taiwan's TSMC is already shipping its roughly equivalent generation of technology, including in chips inside iPhones.
In 2016, a biennial report that had long served as an industry-wide pledge to sustain Moore’s law gave up and switched to other ways of defining progress. Analysts and media—even some semiconductor CEOs—have written Moore’s law’s obituary in countless ways.
Keller doesn’t agree. “The working title for this talk was ‘Moore’s law is not dead but if you think so you’re stupid,’" he said Sunday. He asserted that Intel can keep it going and supply tech companies ever more computing power. His argument rests in part on redefining Moore’s law.
“I’m not pedantic about Moore’s law talking just about transistors shrinking—I’m interested in the technology trends and the physics and metaphysics around that,” Keller says. “Moore’s law is a collective delusion shared by millions of people.”
Keller said Sunday that Intel can sustain that delusion, but that smaller transistors will be just one part of how. On the conventional side, he highlighted Intel’s work on extreme ultraviolet lithography, which can etch smaller features into chips, and smaller transistor designs based on nano-scale wires due to arrive in the 2020s.
Keller also said that Intel would need to try other tactics, such as building vertically, layering transistors or chips on top of each other. He claimed this approach will keep power consumption down by shortening the distance between different parts of a chip. Keller said that using nanowires and stacking his team had mapped a path to packing transistors 50 times more densely than possible with Intel's 10 nanometer generation of technology. “That’s basically already working,” he said.
In January, Intel showed a new chip design called Lakefield that stacks multiple chips on top of one another to fit more computing power into a given space. The company is also part of a shift to a new, Lego-block-style approach to making chips, where modules dubbed chiplets are recombined to make development faster.
Doesn’t this mean that progress will be harder won, less predictable and more expensive? Keller describes the future differently: “Weirder and cooler,” he says.
Despite the metaphysical musing, Keller is a serious technical leader. He has ushered other notable names through major technological paradigm shifts.
At AMD in the early 2000s, Keller co-authored the specification for an era of more powerful, 64-bit processors. He later joined Apple, where he helped the company design its first mobile processors. That strategy has helped Apple keep the iPhone ahead of competitors with features like face unlock. Before joining Intel, Keller led Tesla’s in-house chip design operation to support Elon Musk’s ambitions for self-driving cars.
Stacy Rasgon, a semiconductor analyst with Bernstein, says Keller’s track record suggests he will make an impact at Intel. He also says Intel’s troubles, mostly built up under the watch of CEO Brian Krzanich, who stepped down last year, are formidable. “Keller is right, wonderful amazing things can be done—but there needs to be a business case to do it,” Rasgon says. That last part has got trickier.
Rivals such as Taiwan’s TSMC, which makes chips for clients including Apple, and AMD—which competes with Intel’s server chips—have shown themselves more nimble and efficient with R&D spending, Rasgon says. Intel has acquired several companies that make specialized chips to run AI software, but faces competition from Nvidia, whose GPUs have become an AI standard; Google and Amazon also are designing their own AI chips for use in their data centers.
Keller hasn’t been at Intel long enough to make much of a visible mark on the company’s response to those challenges. It takes years to research, design, and produce new chip technology. When asked how Intel’s output might change under his leadership and interpretation of Moore’s law, his answer is vague. “Make faster computers,” Keller says. “That’s what I want to do.”
Rasgon says a true assessment of his impact should be possible in about five years. “This stuff takes time,” he says.