What I Learned Building a Custom AI Chip

Published 2024-03-26 · Updated 2026-05-23 · 5 min read · AI Hardware and Infrastructure · By Sahin Boydas

After spending years working in Silicon Valley, I've learned how important the right AI hardware is for a company’s success. I’m sharing the lessons I wish I knew when I started—from dealing with GPU shortages to designing custom chips.

I almost gave up on what i learned building a custom ai chip entirely. Then something clicked that changed my whole approach.

After spending years working in Silicon Valley, I've learned how important the right AI hardware is for a company’s success. I’m sharing the lessons I wish I knew when I started—from dealing with GPU shortages to designing custom chips.

The Counterintuitive Truth

Here's what surprised me most about what i learned building a custom ai chip: the best practitioners do less, not more.

When I was building MovieLaLa, we tried to do everything at once. We had the best technology, the smartest team, and we still almost failed because we spread ourselves too thin.

The lesson I took from that experience, and from watching hundreds of other companies, is that simplicity beats complexity every time. It sounds simple. It's incredibly hard to execute.

Why Most Approaches Fail

Let me be direct: about 70% of the approaches I see to what i learned building a custom ai chip are fundamentally flawed. Not slightly off. Fundamentally flawed.

The root cause is usually one of three things:

  • Copying what big companies do without understanding why they do it. What works for Google doesn't work for a 10-person startup.
  • Over-engineering the solution when a simple approach would work better. I've seen teams spend six months building something that could have been done in two weeks.
  • Ignoring the human element. Technology is the easy part. Getting people to actually use it is where the real challenge lives.

The Framework That Actually Works

I'm going to share the exact framework I use when evaluating what i learned building a custom ai chip. It's not complicated, but it requires discipline.

Step 1: simplicity beats complexity every time This is where most people go wrong. They skip this step entirely and jump straight to execution. Don't do that.

Step 2: you need to move fast and break things Once you have the foundation right, this becomes much easier. I've watched founders struggle with this for months when the answer was staring them in the face.

Step 3: Iterate relentlessly Nothing works perfectly the first time. The companies in my portfolio that nail what i learned building a custom ai chip are the ones that treat it as an ongoing process, not a one-time project.

The AI Angle

I can't talk about what i learned building a custom ai chip in 2026 without mentioning AI. As someone who's invested in Anthropic, OpenAI, Scale AI, and Hugging Face, I have a front-row seat to how AI is transforming this space.

The short version: AI makes good practitioners better and bad practitioners worse. It's an amplifier, not a replacement.

I've seen companies use AI to 10x their what i learned building a custom ai chip capabilities. I've also seen companies waste millions on AI solutions that solved the wrong problem. The difference comes down to understanding what you're actually trying to achieve.

This connects to broader themes around NVIDIA, AI data centers, edge AI, AI cloud, GPU shortage that I've been thinking about a lot lately.

Wrapping Up

I've shared a lot here, and I know it can feel overwhelming. But here's the thing about what i learned building a custom ai chip: you don't need to get everything right on day one. You just need to get started and keep improving.

The founders in my portfolio who excel at what i learned building a custom ai chip share one trait: they're relentlessly practical. They don't chase perfection. They chase progress.

That's the mindset I'd encourage you to adopt. Start where you are. Use what you have. Do what you can. And keep pushing forward.

As always, I'm rooting for you.

Frequently Asked Questions

What would you do differently looking back?

I'd move faster on the things that were working and cut the things that weren't sooner. Most founders, myself included, hold onto failing strategies too long because of sunk cost. Speed of learning is everything.

What was the biggest challenge in this case?

Almost always, the biggest challenge is people and alignment, not technology or strategy. Getting the right team focused on the right problem is harder than any technical challenge I've encountered.

Can these results be replicated?

The specific numbers will vary, but the underlying patterns and principles are transferable. The key is understanding the context behind the results, not just copying the tactics. Every company has unique constraints that shape what works.

More in AI Hardware and Infrastructure

  • From TPU to Your Own Custom Silicon: A Founder's Journey — After years in the trenches of Silicon Valley, I've seen firsthand how the right AI hardware can make or break a company. I'm sharing the hard-won lessons and contrarian insights I wish I had when I started, from navigating the GPU shortage to building our own custom silicon.
  • Surviving the GPU Apocalypse: A Founder's Guide to the Shortage — After years in the trenches of Silicon Valley, I've seen firsthand how the right AI hardware can make or break a company. I'm sharing the hard-won lessons and contrarian insights I wish I had when I started, from navigating the GPU shortage to building our own custom silicon.
  • The 6 AI Infrastructure Mistakes That Are Secretly Killing Your Startup — After years in the trenches of Silicon Valley, I've seen firsthand how the right AI hardware can make or break a company. I'm sharing the hard-won lessons and contrarian insights I wish I had when I started, from navigating the GPU shortage to building our own custom silicon.
  • Cerebras Systems — Portfolio Company | Angel Investment by Sahin Boydas — Building the world's largest AI chips for training and inference at unprecedented scale.
  • Why the Future of AI Is Not in the Cloud 338 — After years in the trenches of Silicon Valley, I've seen firsthand how the right AI hardware can make or break a company. I'm sharing the hard-won lessons and contrarian insights I wish I had when I started, from navigating the GPU shortage to building our own custom silicon.
  • The Counterintuitive Truth About AI Chip Design 869 — After years in the trenches of Silicon Valley, I've seen firsthand how the right AI hardware can make or break a company. I'm sharing the hard-won lessons and contrarian insights I wish I had when I started, from navigating the GPU shortage to building our own custom silicon.

All AI Hardware and Infrastructure articles · Sahin's angel investments · Startups he founded