Custom silicon and GitHub anchor Microsoft's bull case
The bull case argues that Microsoft's custom Maya 200 chips and the high ROI of GitHub Copilot provide strong margin defense and developer lock-in.
The argument
The speakers highlighted that shifting 20% to 30% of AI workloads to Maya 200 chips could significantly reduce reliance on expensive NVIDIA GPUs, while GitHub Copilot acts as a primary funnel for Azure cloud adoption.
The thesis, stress-tested
✓ What validates it
- ✓Microsoft reporting material margin improvements from custom silicon deployment
- ✓Azure growth remaining above 35% driven by GitHub developer conversions
▸ Risks discussed
- ▸Failure of Maya 200 to match NVIDIA's performance at scale
- ▸Slowing developer adoption of GitHub Copilot due to competing coding assistants
Hear it yourself
"of, you know, hundreds of millions of people and companies worldwide. So there's a reason we use Excel repeatedly in our Constellation and Topic episodes to kind of explain how exactly horizontal market software works. And that mostly is that it dominates so many different industries, pretty much exactly the opposite of vertical market software. And the same is true for PowerPoint and Word, and they are even the standard now when you work with AI. So I usually work in Google Sheets, but when I work with Claude, for example, it kind of defaults to Excel."
11:30
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE