Akash's Substack

My personal Substack

By Akash Borate

· Launched 5 months ago

Understanding How Agentic AI Is Rewriting the CPU:GPU Ratio

Agentic AI is changing the economics and architecture of AI infrastructure. As tool calls, retrieval, validation, and orchestration become dominant…

Jun 21•Akash Borate - Factryze

6

Understanding how to bring up NVIDIA GB300 NVL72: From Facility to First Run

Having last touched the GB300 during my tenure at Meta, I can tell you: the number of things that need to align for this machine to land and provision…

May 10•Akash Borate - Factryze

11

1

Understanding Multi-Node Multi-GPU Fault Monitoring

Most teams obsess over performance metrics. Few monitor fault behavior. That blind spot is quietly burning millions in wasted GPU hours. The smartest…

Apr 4•Akash Borate - Factryze

2

Understanding GPU stragglers - Infra deep dive

Not all GPU failures crash your job. Fail-slow events are worse silent, expensive, hard to find. A practical infra deep dive into detecting GPU…

Mar 3•Akash Borate - Factryze

3

Understanding GPU Utilization Metrics

GPU utilization is one of the most important metrics to know if you are renting or plan to use GPUs. At first glance, this sounds obvious: Higher…

Feb 2•Akash Borate - Factryze

5