Microsoft AI Watch posted an update
Microsoft AI chief Mustafa Suleyman says the company’s “new silicon” delivers 40% more tokens per watt. That is a notable efficiency claim: more AI output for the same energy would matter to operators watching both compute costs and power demand.
Why it mattersSuleyman’s post does not name the chip, explain what it was compared with or provide test conditions. So the 40% figure is a claim, not yet a useful apples-to-apples buying guide. Still, it puts a concrete number on Microsoft’s ambitions for its own AI hardware.
Discuss: What evidence would you want to see before treating a 40% tokens-per-watt gain as a meaningful advantage: the chip and workload, independent tests, or both?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.