Oracle’s approach to chip design differed from hyperscale rivals in that it owned a minority stake in Ampere Computing, which produces Arm-compatible CPUs for other companies in addition to Oracle, instead of employing its own teams who develop processors solely for in-house use.
Oracle Chairman Larry Ellison said Wednesday that the company sold its stake in chip designer Ampere Computing “because we no longer think it is strategic for us to continue design, manufacturing and using our own chips in our cloud data centers.”
Japanese investment giant SoftBank Group acquired Ampere for $6.5 billion in an all-cash transaction last month to boost its AI capabilities as the investor pursues major AI infrastructure projects, like the Stargate Project in the United States.
[Related: How Arm Is Winning Over AWS, Google, Microsoft And Nvidia In Data Centers]
Oracle said it received a $2.7 billion pre-tax gain from the sale of its Ampere stake—which reached 29 percent in May 2024, according to a previous regulatory filing.
The company’s approach to chip design differed from hyperscale rivals Microsoft, Amazon Web Services and Google in that it owned a minority stake in Ampere, which produces Arm-compatible CPUs for other companies in addition to Oracle, instead of employing its own teams who develop processors solely for in-house use.
But the sale of Oracle’s Ampere stake nevertheless stood in contrast to the approach of rival hyperscalers, who have pursued the development of their own chips to meet the high demand for AI and to bring down computing costs.
Ampere Has Spent Past Few Years Building A Channel Presence
While Oracle was a big buyer of Ampere’s CPUs, the chip designer has spent the past few years trying to build a business selling products through channel partners.
In an interview with CRN back in May, Ampere Chief Product Officer Jeff Wittich said the company hopes to grow its channel business “very fast” but acknowledged it didn’t represent a large percentage of sales volume at the time.
The executive was talking about the new Ampere System Builders program, which seeks to create more flexibility and lower costs for AI and cloud computing infrastructure by bringing together several IT infrastructure vendors to speed up the development of server platforms.
“What we hope is the next step is that we get a lot of attention and interest out of those channel partners, and we can start to pull them into the program,” Wittich said.
Ellison: Oracle Is Now Committed To ‘Chip Neutrality’ Policy
Ellison explained the rationale for the Ampere stake sale in Oracle’s second-quarter earnings release. As a consequence of this move, he said, the company is “now committed to a policy of chip neutrality where we work closely with all our CPU and GPU suppliers.”
“Of course, we will continue to buy the latest GPUs from Nvidia, but we need to be prepared and able to deploy whatever chips our customers want to buy,” said Ellison, who is also Oracle’s CTO, in a statement. “There are going to be a lot of changes in AI technology over the next few years and we must remain agile in response to those changes.”
While Oracle has been a significant buyer of GPUs and related components from Nvidia, the company announced a deal with AMD in October to launch the “first publicly available AI supercluster” powered by the chip designer’s Instinct MI450 GPUs. The deployment is set to debut in the third quarter of next year with an initial 50,000 GPUs.
The semiconductor industry is seeing competitive shifts elsewhere, with Qualcomm planning to re-enter the server CPU market and, separately, launch a pair of AI accelerator chips over the next two years. Google could also shake up the market with the company reportedly exploring the deployment of its TPUs—traditionally used for the company’s own infrastructure—inside data centers run by customers, including Meta, which is a major Oracle customer.
