Chip Giants Intel and Nvidia face new threats from Amazon to Google and Apple

Amazon has unveiled a new chip this month that it says promises to speed up the way algorithms that use artificial intelligence learn from data. The company has already designed other processors for its cloud computing arm, called Amazon Web Services, including the brains of computers known as CPUs.

The pandemic has accelerated the rise of cloud computing, as companies have generally adopted the type of digital tools used by these remote servers. Amazon, Microsoft, Google and others have experienced strong growth in the cloud during the remote work period.

Business customers are also more eager to analyze the data they collect about their products and customers, fueling the demand for artificial intelligence tools to make sense of all this information.

Google was one of the first engines of the technology giants, which launched an AI processor in 2016 and has since upgraded the hardware. Software giant Microsoft, the number 2 cloud behind Amazon, has also invested in chip designs, including a programmable chip to manage AI and another that improves security. He now also works on a central processor, according to a person familiar with his plans. Bloomberg News previously reported on Microsoft’s CPU effort.

Boosting the movements of the tech giants are changes in the way the semiconductor world works and a growing sense that Moore’s law, a key assumption in the industry about constantly improving chip performance, is losing relevance. As a result, companies are looking for new ways to get better performance, not always measured in speed, but sometimes lower power consumption or heat generation.

“Moore’s Law has been around for 55 years, and it’s the first time it’s slowed down very materially,” said Partha Ranganathan, vice president and engineer of Google’s cloud unit, which pursues specialized chips.

The large size of the cloud giants presents a challenge for traditional chip producers. In the past, semiconductor manufacturers tended to design their high-performance semiconductors for generic applications, leaving customers the ability to adapt and get the most out of chips. Older customers now have the financial muscle to drive more optimized designs.

“Although Intel in the 1990s was an order of magnitude larger than all of its customers, now the customer has a higher scale than the supplier,” said James Wang, analyst at New York money manager ARK Investment Management . “As a result, they have more capital and more experience to take components home.”

Nvidia, now the largest chip maker in the United States by market capitalization, is worth $ 330 billion and Intel is $ 207 billion. The big clouds, Amazon, Microsoft and Google-parent Alphabet Inc.,

GOOG -0.97%

each top value in a trillion dollars.

Tailor-made efforts are partly possible thanks to the increase in contract chip manufacturers, who make semiconductors designed by other companies. This deal helps tech giants avoid the multimillion-dollar cost of building their own chip factories. Taiwan semiconductor manufacturing Co.

TSM 0.37%

, in particular, has jumped to the forefront of chip production technology.

The changes have benefited chip design firm Arm Holdings Ltd., which sells circuit designs that anyone can use after paying a license fee. Apple is a big Arm customer, as are all the big tech companies that make their own chips.

It is estimated that Amazon, Google, and Microsoft operate millions of servers on global data center networks for their own use and are leased to their millions of cloud computing customers. Even small improvements in performance and minimal reductions in the cost of powering and cooling the chips are worth the effort when handed out across these vast technological empires. Facebook Inc.

FB 0.70%

he has also explored working on his own chips.

David Brown, vice president of AWS, said making their own processors was an obvious choice for Amazon, given the performance gains it could achieve by ceasing to be compatible with older software and other standard features of Intel chips that don’t need large data center operators.

“We were able to build one [chip] this is optimized for the cloud, so we’ve been able to eliminate a lot of things that just don’t need it, ”he said. Amazon’s chip-making efforts began largely with the acquisition of an Israeli company called Annapurna Labs , about five years ago.

Custom chips are gaining favor in consumer products as well. Apple this year began using its own processors on Macs after 15 years coming from Intel. Google has incorporated its AI chip into its Pixel smartphones.

So far, the lost business for traditional chip makers has been modest, said Linley Gwennap, a chip industry analyst. The market share of all custom-based arm-based CPUs is less than 1%, he said. Google’s AI chips are, by far, the processors designed by higher-tech technology companies, he said, that comprise at least 10 percent of all AI chips. Intel still supplies the vast majority of CPUs that go to data centers.

Headlines don’t stay inactive in the race for the supremacy of the cloud and AI chips either. Nvidia agreed this year to buy Arm in what would be the largest acquisition of the chip industry. And Ian Buck, who oversees Nvidia’s data center business, said the company works closely with its larger customers to optimize the use of their chips in their hardware configurations.

Intel said about 60 percent of central server processors sold to large data center operators are customized according to customer needs, often disabling chip features they don’t need.

And Intel has invested in AI processors and other specialized hardware, including the purchase of Israel-based Habana Labs last year for about $ 2 billion. AWS recently agreed to put Havana’s AI training chips in its data centers, as Amazon develops its rival chips that it believes will work best when they appear next year.

Remi El-Ouazzane, strategy director for Intel’s data platform group, said Havana’s chips at AWS could challenge Nvidia, the dominant player in an AI training market that said it will be worth more than $ 25 billion in 2024. “It’s a great new opportunity network for Intel and it’s a very big market,” he said.

—For more analysis, reviews, tips, and WSJ technology headlines, sign up for our weekly newsletter.

Write to Asa Fitch at [email protected]