Subscribe
Sign in
Home
Data Infra
Data Center
Security
Archive
Latest
Top
Discussions
How AI Tokens Are Made
Inside the five scheduling, memory, and routing problems that turn GPU compute into output tokens — and why neoclouds keep buying the companies that…
Jul 30
•
Chris Zeoli
33
1
What a Gigawatt Costs
$37.6 billion, six and a half years, and 57 cents of every dollar that never reaches a chip company. What the AI buildout is physically made of, and…
Jul 27
•
Chris Zeoli
40
3
1
The InfiniBand and Ethernet Wars
The fabric war is over and Ethernet won it on economics. The pattern the ending exposed is about to run again one tier up, and it points at where AI…
Jul 24
•
Chris Zeoli
25
2
1
Kimi: When Frontier Weights Become Open
On July 27, anyone with 1.4 terabytes of storage, 5TB of memory and a rack of accelerators will be able to run the world's third-most-intelligent AI…
Jul 18
•
Chris Zeoli
45
2
Who Captures Value in AI Infrastructure?
A value capture framework for the AI infrastructure stack: where profit accrues on over a trillion of spend
Jul 13
•
Chris Zeoli
46
2
5
The AI Networking Stack
A GPU is only as valuable as the fabric that feeds it — how copper, light, and silicon move trillion-parameter workloads, and where the durable margins…
Jul 5
•
Chris Zeoli
33
4
How an AI Token Travels Through a Data Center
Trace one prompt from gateway to scheduler, KV cache, GPU, and network — every component that powers AI inference.
Jul 1
•
Chris Zeoli
66
4
4
June 2026
How Claude Code Cracked Coding
Software fell first because it is the only knowledge-work domain that ships with its own grader. Inside the verifiable-reward loop that took Claude Code…
Jun 28
•
Chris Zeoli
18
7
China's Open-Weight Takeover
A field map of Chinese AI labs and platforms: Qwen, DeepSeek, Kimi, ByteDance, and Alibaba. How they came to dominate open models, work around compute…
Jun 24
•
Chris Zeoli
22
2
Software Ate the World. Now Hardware Is Eating Software.
Gravity, all the way down: AI value is rotating back into semiconductors, data, and models — and thinning to a sliver at the application layer.
Jun 23
•
Chris Zeoli
25
3
2
What is CXL?
The interconnect rewiring AI memory: turning stranded DRAM into a shared resource as memory prices spike and KV cache becomes the bottleneck
Jun 22
•
Chris Zeoli
22
2
3
What Does Broadcom Do?
Broadcom booked $30 billion in AI chip orders in a single quarter — against $10.8 billion shipped. This is what they're actually building, and why the…
Jun 15
•
Chris Zeoli
16
1
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
