Arm Holdings to sell its own chips, eyeing sales goal of $19.2 billion

SAN FRANCISCO – Arm Holdings, which made its name licensing technology to semiconductor makers, will begin selling its own chips for the first time, adding a business that it expects to generate about US$15 billion (S$19.2 billion) annually within five years.

Meta Platforms will be the first major customer for the company’s chip, called an AGI CPU, Arm said on March 24 at an event in San Francisco. The product will have as many as 136 cores – a measure of processing power – and draw 300 watts of electricity, Arm said. Taiwan Semiconductor Manufacturing Company will produce the chips.

As part of the move, Arm laid out aggressive sales targets for the coming years. The Britain-based company expects revenue from the new chip business to eclipse sales from its current operations, which focus on selling intellectual property (IP).

That will help generate total sales of roughly US$25 billion within five years, five times the level today, Arm said. The IP business will continue growing, hitting about US$10 billion by that point, the company predicted. 

Earnings will reach US$9 a share in that time frame. Analysts have estimated that earnings will be US$1.75 a share in the current fiscal year, excluding some items.

Under chief executive Rene Haas, Arm has shifted from its roots as a provider of smartphone technology and taken a greater role in the data centre market. The change is meant to help the business get more of the money generated by what is often complex and expensive work.

The shift also helps Arm benefit from bigger-ticket purchases. Even the most expensive smartphone chips cost tens of dollars. The highest-end data centre semiconductors can run in the tens of thousands.

The new sales outlook sent Arm shares up as much as 7.8 per cent in late trading. They had climbed 23 per cent this year through the close.

Chief financial officer Jason Child said that while Arm’s current offerings – semiconductor designs and technology licenses – are more profitable in terms of margins, selling actual chips will bring in far more profit dollars.

On a theoretical US$1,000 chip, Arm gets about 5 per cent in licensing revenue when customers use its instruction set – the basic code a chip uses to communicate with software. That is pure profit. But if the client uses Arm’s designs, the profit payoff is about US$100. And if Arm makes the chip itself, it gets about US$500 in gross profit dollars, he said.

Arm decided to make the new chip because customers asked for it, Mr Haas said. The product – a central processing unit (CPU), often described as the brains of a computer – is designed to work alongside the accelerator chips offered by companies such as Nvidia. It helps coordinate work between computers, prepares data and runs elements that provide a response to users making AI queries, Arm said.

“The product that we’re building is not only compelling – but we actually have customers who are lined up to buy it,” Mr Haas said in an interview. 

The company said its product offers greater power efficiency compared with traditional CPU designs from Intel and Advanced Micro Devices. That means that data centre owners will be able to wring more computing power from the same footprint and electricity budget, Mr Haas said.

Arm’s increasing reach is a direct threat to the so-called x86 data centre products made by Intel and AMD, Mr Haas said. Taking share from those traditional stalwarts in a rapidly expanding market will allow both his company and its customers to grow, he argues.

“The market is plenty big enough for multiple players,” Mr Haas said.

Arm faces plenty of competition in data centre processors. A number of start-ups and established companies have sought to challenge Nvidia’s dominance in the field with a variety of approaches. And Nvidia itself just introduced a new CPU line-up, targeting the category that Arm is now entering. Mr Haas said his chip is aimed at a different part of the market than Nvidia’s latest addition.

Arm’s chip move also threatens to complicate its relationship with customers. Most of the biggest buyers of data centre silicon, including Meta, have their own in-house chip programmes. And almost all of them license technology and designs from Arm.

Data centre operators buy chips from a range of suppliers. That includes Meta, which recently signed long-term deals with Nvidia, AMD and start-up Cerebras Systems. The social networking company plans to use the AGI CPUs with its other chips. 

“We worked alongside Arm to develop the Arm AGI CPU to deploy an efficient compute platform that significantly improves our data centre performance density,” Mr Santosh Janardhan, head of infrastructure at Meta, said in a statement.

Other companies – including OpenAI, Cerebras and SK Telecom – also plan to deploy the AGI CPU in their infrastructure, Arm said. Off-the-shelf systems using the chip are out now from sellers such as Quanta Computer and Super Micro Computer. They should be available in greater volumes in the second half of this year, Arm said. 

Under Mr Haas, Arm has increased its revenue by more than 20 per cent a year. Annual sales topped US$4 billion for the first time in 2025. 

At the same time, Arm has maintained a startlingly high level of profitability. Gross margin, the percentage of revenue left after deducting costs of production, was 98 per cent in its most recent quarter.

Most of Arm’s peers in the chip industry have much higher sales but lower margins. Even Nvidia, with its near lock on sales of AI accelerators, has margins in the mid-70 per cent range. But Arm generates a tiny fraction of the revenue: Nvidia is on course for annual sales of US$356 billion this fiscal year, according to Wall Street estimates. 

SoftBank Group, which owns a majority stake in Arm, is also ramping up its own efforts to get into AI data centres. That push has involved acquiring chip start-ups and investing heavily in data centre owners. BLOOMBERG