Enterprise

Moore’s Law is all but dead. This new transistor design could keep it alive and reshuffle the industry.

Intel, Samsung and TSMC are racing to achieve a generational leap in transistor technology. It might reshuffle the industry pecking order.

Moore’s Law is all but dead. This new transistor design could keep it alive and reshuffle the industry.

Video: Intel via YouTube

A once-in-a-decade shift is underway involving one of the most elemental building blocks of computer chips, and it has the potential to reshuffle the pecking order of chip giants for years to come.

Intel, Samsung and TSMC are racing to achieve a generational leap in transistor technology. This leap must occur to realize anything close to the computing requirements demanded by the ideas behind the metaverse, to produce AI that isn’t a joke, to make truly self-driving cars or even make apps load faster.

This next-generation design is called “gate-all-around.” With new materials, and redesigned manufacturing tools that cost tens of millions each, the new gates accomplish one thing: They more tightly control the flow of electricity received by each transistor. Modern chips can have upwards of 30 billion transistors on a single device, and in some cases tens of billions more. In 2025, Gartner expects chip manufacturers to generate roughly $5 billion in revenue from the new technology, up from nothing last year.

Chip companies must deliver substantially more computing horsepower every year to get to a version of the future that’s been promised by the tech titans. Doing so requires some of the most complex, expensive manufacturing equipment on the planet, and the development of even more creative ways to improve fundamental aspects of chip construction. That means making already atomic-sized features even smaller. This process, loosely described as Moore’s law, has kept the chip industry humming for a half-century, but it’s getting harder.

“The rate at which we're shrinking is for sure slowing down, big time,” Applied Materials Vice President Kevin Moraes said.

Chip manufacturers improve high-volume production and performance with a combination of advanced tools such as extreme ultraviolet lithography machines and techniques that help squeeze more features onto each piece of silicon. Typically manufacturers tout improvements to their techniques and technologies as process nodes that use smaller and smaller nanometer numbers.

But another way to tackle the increasingly difficult problem is to further refine the fundamental building block of each chip: the transistor.

“Process nodes are just one measure of the progress,” Jack Gold, principal analyst at J.Gold Associates, said. “I’d argue that even more important is the design of the transistors themselves, and not just the process node. To that end, Intel and others have been advancing the design of the transistors, particularly in the way the gates are designed.”

Better gates, faster chips

A gate is the tiny portion on each transistor that controls whether a transistor receives electricity — kind of like using your foot on a garden hose to turn water on or off — in order to represent the zeros and ones that make up bits of data. But gates, like hoses, can be imperfect, and some of the electricity can slip through even with the most advanced designs.

“As you start making things smaller, you learn that there are some electrical characteristics with smaller elements that didn’t work as well as they did when they were bigger,” FeibusTech analyst Mike Feibus said. “There’s a rule of thumb that the smaller you get, the more current leakage you’re going to get, which is more heat.”

Because chipmakers know some electricity will evade the gate, it’s becoming harder to continue to shrink features to achieve the power consumption and performance expected of new designs. And to build more efficient gates and transistors that generate less power, chipmakers spend billions of dollars inventing the next, better way to develop them.

“Gate-all-around-based designs will have significantly better performance and efficiency than [existing] designs, potentially shifting the competitive position for many high-performance products,” MLCommons Executive Director David Kanter said.

Surrounding all four sides of a part of the transistor with material instead of the current three-sided design allows the gate to better regulate the flow of electricity. At the atomic scale chips operate at, having better control over the flow of electricity gives designers new options: most importantly, making them even smaller.

Smaller transistors allow designers to squeeze more of them onto a single chip and adding more of these tiny features roughly translates to an improvement in the chip’s ability to perform calculations.

“With better heat and power characteristics, that means you can turn up the juice — with all else being equal — you will have higher clock rates and you won’t need as exotic cooling for high-performance designs, so that can cut costs,” Feibus said.

The manufacturer that successfully deploys the new generation of gate tech in high-volume production will be able to manufacture chips with a leap in computing horsepower that’s impossible with the current process technology.

Fastest out of the gate

All three of the largest chip manufacturers are trying to figure out how to take advantage of this promising technology. This issue isn’t producing one batch of chips with the new feature; it's producing hundreds of thousands, or millions of chips, at the scale at which the most advanced technologies are needed.

Samsung may have a leg up. One of the company’s executives, Kinam Kim, co-authored an early technical paper about the new gate technology nearly 20 years ago. The company was one of the earliest to commit to producing next-generation gates, discussing its path to move to the tech in 2017. In April, Samsung said the company was on track to produce chips at high volume with the new gates this year.

However, there is reason to be skeptical of its progress. Samsung, which did not return a request for comment, has had widely reported problems with ensuring its latest manufacturing process can produce a sufficient number of working chips and has pushed back its planned launch date for the new gates several times, according to Gartner chips analyst Gaurav Gupta.

So even if the company does roll out the new design on certain chips, it might not be a viable option for its high-volume customers who demand reliability — a six-month delay to a new iPhone launch would be disastrous to Apple’s business, for example.

“Samsung has always been saying they will be the first to the gate-all-around at three nanometers,” Gupta said. “And they have been pushing it — next six months, [another] six months more, [another] six months. They’re saying now they are expecting it by the end of the year.”

Intel says that its current plans call for the technology to roll out in 2024 for high-volume customers. There is reason for skepticism with Intel’s efforts too; the company went through years of delays on the last several generations of manufacturing improvements. It was also the last to adopt the EUV lithography tech that is widely viewed as necessary for future improvements. An Intel spokesperson said the company remained on track to hit its roadmap goals.

But Intel was the first company to successfully make chips at high volume with the current generation of gates, which suggests that it has the technical institutional knowledge to do it again. Under CEO Pat Gelsinger, the business has a far clearer direction than the years of dysfunction that preceded him. Still, Intel has a lot to prove, and part of Gelsinger’s plan includes what Gupta described as a “very aggressive” strategy to regain its manufacturing leadership.

As it now stands, TSMC will be the last major manufacturer to move to new gate technology. Executives are tight-lipped, and the company didn’t respond to a request for comment, but TSMC has disclosed that it plans to roll it out sometime in 2025. People familiar with the company’s operations said that, overall, TSMC is a conservatively run business that prefers to avoid risks, if at all possible, and moving to a new, unproven gate technology is a big risk.

But the new gates may matter less to TSMC than its rivals. The majority of its customers buy smartphone chips, and the company has proven that it can make them with the EUV tools that Intel and Samsung have struggled to deploy. As a result, TSMC may be able to use other techniques for the time being to eke out more performance from its silicon until it can ship the new gate designs.

“The thing is, it's all about power and performance, and you see 90%-plus of customers at advanced nodes in the foundry business are with TSMC,” Gupta said, adding that if TSMC’s manufacturing using existing gate technology is superior to Samsung’s, TSMC won’t lose customers.

Sticking with the current generation of transistors has its advantages for chip designers too. New gates require some aspects of chip designs to get scrapped, and industry analysts said that the road to make a chip with new gates could easily exceed two or three years. So if a chipmaker such as Qualcomm, AMD or Nvidia were to make a bet that the new gates can be made this year, and that doesn’t happen, they could be badly hurt.

“They call it bleeding edge for a reason,” Feibus said. “Sometimes you get cut up.”

Fintech

Judge Zia Faruqui is trying to teach you crypto, one ‘SNL’ reference at a time

His decisions on major cryptocurrency cases have quoted "The Big Lebowski," "SNL," and "Dr. Strangelove." That’s because he wants you — yes, you — to read them.

The ways Zia Faruqui (right) has weighed on cases that have come before him can give lawyers clues as to what legal frameworks will pass muster.

Photo: Carolyn Van Houten/The Washington Post via Getty Images

“Cryptocurrency and related software analytics tools are ‘The wave of the future, Dude. One hundred percent electronic.’”

That’s not a quote from "The Big Lebowski" — at least, not directly. It’s a quote from a Washington, D.C., district court memorandum opinion on the role cryptocurrency analytics tools can play in government investigations. The author is Magistrate Judge Zia Faruqui.

Keep ReadingShow less
Veronica Irwin

Veronica Irwin (@vronirwin) is a San Francisco-based reporter at Protocol covering fintech. Previously she was at the San Francisco Examiner, covering tech from a hyper-local angle. Before that, her byline was featured in SF Weekly, The Nation, Techworker, Ms. Magazine and The Frisc.

The financial technology transformation is driving competition, creating consumer choice, and shaping the future of finance. Hear from seven fintech leaders who are reshaping the future of finance, and join the inaugural Financial Technology Association Fintech Summit to learn more.

Keep ReadingShow less
FTA
The Financial Technology Association (FTA) represents industry leaders shaping the future of finance. We champion the power of technology-centered financial services and advocate for the modernization of financial regulation to support inclusion and responsible innovation.
Enterprise

AWS CEO: The cloud isn’t just about technology

As AWS preps for its annual re:Invent conference, Adam Selipsky talks product strategy, support for hybrid environments, and the value of the cloud in uncertain economic times.

Photo: Noah Berger/Getty Images for Amazon Web Services

AWS is gearing up for re:Invent, its annual cloud computing conference where announcements this year are expected to focus on its end-to-end data strategy and delivering new industry-specific services.

It will be the second re:Invent with CEO Adam Selipsky as leader of the industry’s largest cloud provider after his return last year to AWS from data visualization company Tableau Software.

Keep ReadingShow less
Donna Goodison

Donna Goodison (@dgoodison) is Protocol's senior reporter focusing on enterprise infrastructure technology, from the 'Big 3' cloud computing providers to data centers. She previously covered the public cloud at CRN after 15 years as a business reporter for the Boston Herald. Based in Massachusetts, she also has worked as a Boston Globe freelancer, business reporter at the Boston Business Journal and real estate reporter at Banker & Tradesman after toiling at weekly newspapers.

Image: Protocol

We launched Protocol in February 2020 to cover the evolving power center of tech. It is with deep sadness that just under three years later, we are winding down the publication.

As of today, we will not publish any more stories. All of our newsletters, apart from our flagship, Source Code, will no longer be sent. Source Code will be published and sent for the next few weeks, but it will also close down in December.

Keep ReadingShow less
Bennett Richardson

Bennett Richardson ( @bennettrich) is the president of Protocol. Prior to joining Protocol in 2019, Bennett was executive director of global strategic partnerships at POLITICO, where he led strategic growth efforts including POLITICO's European expansion in Brussels and POLITICO's creative agency POLITICO Focus during his six years with the company. Prior to POLITICO, Bennett was co-founder and CMO of Hinge, the mobile dating company recently acquired by Match Group. Bennett began his career in digital and social brand marketing working with major brands across tech, energy, and health care at leading marketing and communications agencies including Edelman and GMMB. Bennett is originally from Portland, Maine, and received his bachelor's degree from Colgate University.

Enterprise

Why large enterprises struggle to find suitable platforms for MLops

As companies expand their use of AI beyond running just a few machine learning models, and as larger enterprises go from deploying hundreds of models to thousands and even millions of models, ML practitioners say that they have yet to find what they need from prepackaged MLops systems.

As companies expand their use of AI beyond running just a few machine learning models, ML practitioners say that they have yet to find what they need from prepackaged MLops systems.

Photo: artpartner-images via Getty Images

On any given day, Lily AI runs hundreds of machine learning models using computer vision and natural language processing that are customized for its retail and ecommerce clients to make website product recommendations, forecast demand, and plan merchandising. But this spring when the company was in the market for a machine learning operations platform to manage its expanding model roster, it wasn’t easy to find a suitable off-the-shelf system that could handle such a large number of models in deployment while also meeting other criteria.

Some MLops platforms are not well-suited for maintaining even more than 10 machine learning models when it comes to keeping track of data, navigating their user interfaces, or reporting capabilities, Matthew Nokleby, machine learning manager for Lily AI’s product intelligence team, told Protocol earlier this year. “The duct tape starts to show,” he said.

Keep ReadingShow less
Kate Kaye

Kate Kaye is an award-winning multimedia reporter digging deep and telling print, digital and audio stories. She covers AI and data for Protocol. Her reporting on AI and tech ethics issues has been published in OneZero, Fast Company, MIT Technology Review, CityLab, Ad Age and Digiday and heard on NPR. Kate is the creator of RedTailMedia.org and is the author of "Campaign '08: A Turning Point for Digital Media," a book about how the 2008 presidential campaigns used digital media and data.

Latest Stories
Bulletins