The U.K.-headquartered chip design company Arm Holdings Plc. is reportedly planning to launch its first-ever complete semiconductor after securing Facebook’s parent company Meta Platforms Inc. as one of its first customers.

In an exclusive report today, the Financial Times suggests that Arm is looking to build its own hardware for the first time, in a move that will bring it into direct competition with many of its customers.

The company is an ubiquitous name in the chipmaking world, but it has never made its own chips. Instead, it licenses to others a key technology called an instruction set, which is a sort of blueprint for building powerful and energy-efficient chips. More recently, though, the company has begun selling more complex core designs that customers can customize more easily.

Arm’s unique status in the chipmaking industry has helped it to earn a nickname as the “Switzerland” of chip technology firms, as it has always been perceived as dealing with customers such as Apple Inc., Nvidia Corp., Qualcomm Inc., Intel Corp., Amazon Web Services Inc. and Microsoft Corp. neutrally, never favoring any particular company.

But that could change, as the massive amounts of money being spent on artificial intelligence chips prove too tempting for the company to ignore. Meta alone has said it’s planning to spend up to $65 billion on AI infrastructure this year, and though much of that money will be spent on Nvidia’s graphics processing units, it will also buy plenty of others, including central processing units from the likes of Intel and Advanced Micro Devices Inc. It’s also thought to be working on its own chip designs.

The Financial Times says Arm is looking to design a CPU for servers, rather than compete with Nvidia and AMD in the GPU market.

Nvidia tried to acquire Arm from its owner SoftBank Group Corp. in 2020 for $40 billion, but the deal was blocked on regulators’ concerns about Arm’s key role in the chip market. Arm responded by instead going public, and its market capitalization has since grown to more than $173 billion. SoftBank remains its biggest shareholder.

This year already, Arm’s stock has increased 29%, as it’s increasingly seen as a key enabler of AI systems. The company’s executive team has told investors that it’s looking to sell more advanced technology to customers, as part of a plan to expand its revenue streams.

The move to build its own chips is not really so surprising, as SoftBank founder Masayoshi Son has made Arm the centerpiece of his plans to create a vast infrastructure network for AI. Last month, Son appeared at the White House alongside U.S. President Donald Trump, OpenAI Chief Executive Sam Altman and Oracle Corp. founder and Chief Technology Officer Larry Ellison, where they unveiled the Project Stargate initiative, which plans to invest $500 billion in AI infrastructure. SoftBank and Abu Dhabi state fund MGX will provide much of the funding, and Arm was also revealed as a technology partner for that initiative.

Another clue came from Arm CEO Rene Haas (pictured) during the company’s latest financial earnings call earlier this month, when he cited the spending plans of Meta, Google (which will invest $75 billion on AI infrastructure) and Microsoft (which will spend $80 billion) as a big opportunity for the company.

“No one is pulling back,” Haas pointed out, referring to how those giants are doubling down on their AI spending amid the rise of Chinese AI startup DeepSeek Ltd.

Moreover, SoftBank is currently trying to acquire another chipmaker, Ampere LLC, which is heavily backed by Oracle and makes CPUs for data center servers. According to the Financial Times, that deal is critical to Arm’s plans to build its own chips.

Though the profits from selling its own chips could be extremely lucrative, Arm also risks alienating some of its major customers if it goes ahead with this move, said analyst Rob Enderle of the Enderle Group.

“Competing with your licensees is a good way to lose those licensees if you’re not careful, and this does represent a risk to those customers who license Arm technology,” the analyst said. “It means they’ll be competing with Arm itself while also using its technology, but of course, Arm will have a significant advantage as it owns that technology.”

The decision could well be a calculated move for Arm, though, as the AI industry is lucrative enough that it may be able to gain enough direct sales to offset any licensing revenue it loses, Enderle added.

“There is always a risk when you license something, as the one who owns the license may eventually decide it wants your revenue and profit, and there is little recourse to such customers if it does decide to move in on their turf,” he said.

Even so, Holger Mueller of Constellation Research Inc. said the move appears to be a risky gamble for the company, as any switch from operating purely on licensing revenue would dilute the company’s profit margins.

“This will expose Arm to supply and production issues and greater stock volatility due to the roller-coaster nature of chip market sales, and it will upset its existing customers,” Mueller said. “It’s surprising, but in any case Arm’s management is aware of all of this, so it must be able to see an upside.”

The Financial Times suggests Arm may announce its plans by the summer or possibly even earlier, but the company itself refused to comment on the report.

However, it’s certainly clear which direction the industry is headed. Four days ago, Reuters reported that OpenAI is looking to develop its own chips to reduce its reliance on Nvidia’s GPUs. It’s said to be in the advanced design stage already, with the next step likely to involve transferring that design to the Taiwanese semiconductor manufacturer Taiwan Semiconductor Manufacturing Co. for experimental production and testing.

Meanwhile, Forbes reported this week that Meta is looking to acquire the South Korean AI chip startup FuriosaAI Inc. as part of its own plan to develop an alternative to Nvidia’s GPUs. That deal could close as early as this month, the report suggested.

Of course, the cloud infrastructure giants AWS, Google and Microsoft have all developed their own AI accelerators, which they offer to cloud customers as an alternative to Nvidia’s hardware.

Photos: Arm/YouTube