The GPU Gold Rush: How Vast.ai Scaled With AI Demand

Mar 27, 2026

Travis Cannell, CEO and early employee at Vast.ai, runs a two-sided marketplace for on-demand GPUs. He explores why inference is driving massive GPU demand. He explains marketplace dynamics, pricing strategies, and how software and support defend a low-cost model. He also discusses rapid growth triggers and how AI reshapes hiring and org design.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Inference Is Fueling GPU Demand

Inference is the primary driver of current GPU demand as models must run on GPUs to respond to prompts.
Travis explains training produces a model file, but inference loads that model into GPU RAM whenever a user interacts with it, making ongoing compute essential.

ANECDOTE

When Vast.ai Hit Hypergrowth

Vast.ai's hypergrowth began around November and accelerated in early 2026 tied to Opus 4.6 and CloudCode releases.
Travis compares the moment to a bigger ChatGPT 3.0 moment and notes worldwide pickup across APAC and Europe.

ADVICE

Balance Low Price With Strong Verification

Compete on price plus platform experience: keep marketplace bare-bones, verify hosts, and let suppliers set prices to maintain low cost.
Vast.ai doesn't own hardware; verification and bandwidth tests preserve trust while enabling low GPU prices.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

In this episode of the ProductLed Podcast, Wes Bush and Esben Friis-Jensen sit down with Travis Cannell, CEO and first employee at Vast.ai, the marketplace for on-demand, low-cost GPUs powering AI workloads around the world.

Travis breaks down one of the biggest shifts happening in AI right now: the rise of inference. He explains why inference demand is exploding, how that shift is fueling Vast.ai’s rapid growth, and why more teams are looking for flexible, affordable GPU access outside of traditional cloud platforms.

The conversation also gets into how Vast.ai built a two-sided GPU marketplace with 20,000 GPUs, why its pricing model creates powerful marketplace dynamics, and what makes its software-first approach difficult to replicate. Travis shares how the company thinks about competition, customer support, GPU hosting economics, and why winning in a fast-growing marketplace depends on much more than just low prices.

They also explore how AI is changing org design inside high-growth companies. Travis talks candidly about pausing hiring, using AI to accelerate engineering work, and why Vast.ai has leaned into an in-office culture while staying extremely lean.

If you want a clearer picture of where AI infrastructure is heading, and how one company is scaling quickly with a software-first model, this episode is packed with insight.

Key Highlights:

02:34 - Why Inference Is Fueling the Next AI Boom
04:15 - Inference Explained in Plain English
06:15 - The Moment Vast.ai Hit Hypergrowth
10:04 - Why Teams Choose Vast Over AWS
12:34 - Building a Two-Sided GPU Marketplace
17:03 - Competing on More Than Just Price
18:52 - The Real Economics of Hosting GPUs
25:21 - The Network Effects Behind Vast.ai
31:31 - Building a Lean Team During Hypergrowth
36:09 - Why AI Changed Their Hiring Strategy

Resources:

🚀 Vast.ai: Marketplace for low-cost GPUs: https://vast.ai
💼 Connect with Travis Cannell on LinkedIn: https://www.linkedin.com/in/traviscannell/
💼 Connect with Wes Bush on LinkedIn: https://www.linkedin.com/in/wesbush/
💼 Connect with Esben Friis-Jensen on LinkedIn: https://www.linkedin.com/in/esbenfriisjensen/
🧠 Sign up for the ProductLed Newsletter: https://www.productled.com/newsletter