Microsoft
and
Nvidia
, beyond vying for the world’s most valuable company title, have reportedly been clashing over AI chip design for months. The dispute eventually reached Microsoft
CEO Satya Nadella
, who intervened to resolve it.
According to a report by The Information, Nvidia CEO Jensen Huang has tight control over how chips are allocated and installed by its clients, which include Microsoft.
While the Nvidia CEO previously explained that the company dictates how chips are allocated to prevent companies from stockpiling them during periods of limited supply, this control reportedly caused a disagreement with Microsoft. Reportedly, these tensions have been building for some months behind the scenes
Why alleged tensions brewed between Microsoft, Nvidia
According to a report, Nvidia’s efforts to manage chip allocation extended beyond simply deciding who received the chips. The company reportedly attempted to dictate how these chips were installed within data centres, specifically focusing on the server rack configuration.
This additional control over how Microsoft utilised the purchased hardware is what allegedly led to a feud between the two tech giants.
The report further details that the disagreement escalated into a feud when Nvidia tried to convince Microsoft to adopt its next flagship chip, the GB200, within the server rack design specified by Nvidia.
How this is a problem for Microsoft
Installing the chips the way Nvidia wanted may reportedly have hindered Microsoft’s ability to switch to different
AI chips
. Citing , citing people involved, the report said that when the dispute over server racks did not resolve for several weeks, it reached Microsoft CEO Satya Nadella’s desk, the report said.
Microsoft eventually won out as Nvidia backed down and agreed to the
Big Tech company
’s demands, The Information reported.
Reportedly, Nvidia CEO is worried whether companies like Microsoft and Amazon Web Services are going to run out of data centre space, which means a dearth of space will lead to less sales of AI chips.