What's going on with Qualcomm's server team?

A few years back, when IBM was already losing a lot of hardware folks to attrition ,Qualcomm gutted IBM's POWER teams in Raleigh and Austin to start their own server chip effort. Earlier this year, Qualcomm decided that this was a dead end and decided to defund the effort. They announced this to the team relatively early this year and it made the news in May when they publisized the fact that they were looking for a buyer.

Qualcomm looked like the best hope for producing an ARM server chip with decent performance -- they had a very good team and the first chip they produced, which was made with the intent of being a practice run, not something that would be competitive Intel, was already more interesting than a lot of competing offerings. I was hoping they'd find a buyer quickly, so work wouldn't get disrupted too much, but it looks like that's not going to happen.

I'm told, by people who should be in a position to know, that after some shopping around, Ampere Computing was willing to buy the team, but this basically ended up being blocked by the Chinese joint venture Qualcomm is involved in. They didn't literally block the deal, but they demanded the same sort of access they currently have after the acquisition, which caused Ampere to back out.

Paul Jacobs also made a reasonable bid for the team, but I'm told that was killed by a Qualcomm exec for political reasons. The backstory here is that Paul Jacobs is the former CEO and a former board member, who was booted from the board recently. He lined up funding and a list of customers that would ensure the team's continued existence for some time, but I'm told that an exec was worried that they would look bad if the team was successful after Qualcomm decided to pull the plug, so they rejected the deal and, in order to guarantee that no buyer could make them look bad, they laid off half of the server group and stuck the other half into R&D to do work that's TBD.

Parts of the team have gone various places (e.g., a bunch of the core design people got picked up by Microsoft to work on a RISC-V chip) and will probably do interesting work, but it's a shame that the team couldn't be kept together. It's really hard to assemble a good chip team, and Qualcomm had a great one.