Arm, a UK-based chip designer whose technology can be found in smartphones and PCs, says AI demand is fueling revenue. “We are also seeing strong momentum and tailwinds from all things AI,” said Rene Haas, Arm CEO, on a conference call with analysts and investors on Wednesday, Feb.7.
Arm’s shares are up 55% as of 11:48am ET, following the company’s third-quarter earnings that beat Wall Street expectations.
In the last three months ending Dec. 31, Arm’s revenue was up 14% to $824 million (pdf), driven by royalty revenue and better-than-expected licensing revenue, which comes from selling more complete chip designs that semiconductor companies use in their own chips.
The boost in the latter can be attributed to long-term license agreements with “leading technology companies” that are using Arm’s most advanced CPUs to run AI products, according to the company. Tech giants like Google and Microsoft need a lot of computing power to run the most advanced AI models.
What Arm sees on the horizon for AI and phones
Most of the generative AI chatter has revolved around chatbots, but the technology is moving into the device space. Arm provides the technology needed to power Google’s latest large language model Gemini Nano (goodbye Bard), which runs on the Pixel 8 smartphone. Phone makers Samsung, Vivo, and Xiaomi have also announced new Arm-based smartphones that demonstrate generative AI capabilities.
So what does a generative AI-boosted phone look like? Features include live call translation, intuitive search features—such as a person being able to circle something on her phone to search—and AI-enhanced photo and video capabilities like being able to erase something in the background of a photo.
“We’re seeing from the design standpoint more and more compute technology being pushed into those phones such that they are AI capable and AI ready because this field is moving very, very fast,” Haas said. “You know, a year from now, who knows what type of AI applications it might be able to run on a smartphone... Nobody wants to be caught behind with not enough performance when the new application comes out.”