Subscribe
Sign in
Home
Chat
Archive
About
Latest
Top
Discussions
The Changelog Q2.26
What changed in our technology theses, models, and conviction over the last six months
22 hrs ago
•
Ben Bajarin
3
Breaking the Memory Wall With CXL
Why disaggregated memory is moving closer to commercial deployment and who benefits
Aug 6
•
Ben Bajarin
9
4
2
July 2026
The Behind-the-Meter AI Buildout
Why Bloom Energy’s Q2 shows that onsite power may be starting to scale
Jul 30
•
Ben Bajarin
10
2
The Hyperscaler Capacity Partner Hierarchy
Why the preferred external AI infrastructure partner may own the physical bottleneck while leaving compute control to the hyperscaler
Jul 28
•
Ben Bajarin
5
In 2023-2024 We Weren't Bullish Enough
What the 2023–2024 models saw, what they missed, and why today's market requires a different unit of analysis
Jul 24
•
Ben Bajarin
4
Cybersecurity and the Enterprise AI Control Layer
Why Cybersecurity Becomes the Budget That Lets AI Scale
Jul 23
•
Ben Bajarin
4
2
China’s AI Arms Race Runs Through Its Chips
A guide to China’s chips, data-center compute, and open models
Jul 21
•
Ben Bajarin
6
1
1
Gigawattonomics
The price of a watt, the price of a FLOP, and the return on an AI factory
Jul 16
•
Ben Bajarin
6
Meta Is an Ads Company Building a Token Factory
A growth thesis, compute monetization framework and 2026-2030 financial model
Jul 14
•
Ben Bajarin
5
The GPU Tsunami: TSMC, Intel, and Samsung Foundry
How AI silicon is turning foundry from wafer capacity into platform capacity
Jul 7
•
Ben Bajarin
11
1
Where AI Constraints Become Pricing Leverage
Mapping Margin Quality Beneath the AI Infrastructure Constraint Stack
Jul 2
•
Ben Bajarin
5
June 2026
Counting Real AI Capacity
Counting AI infrastructure by deployable capacity rather than by announced megawatts. Deployment is paced by the slowest synchronized layer of the…
Jun 30
•
Ben Bajarin
5
2
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
