Thunder Compute × Lyra: From Founding Engineer to Series A

About Thunder Compute

Thunder Compute is a YC-backed, developer-first GPU cloud. The promise is simple: the cheapest way to run GPU workloads, with a clean UX, fast onboarding, and hyperscaler-grade features. It is built for developers who want cloud GPUs without the cost or the complexity of the incumbents.

When founder Carl Peterson came to Lyra in August 2025, the product was early. What he needed was the hardest hire a startup ever makes: a founding engineer who could take the whole product surface, the frontend, the CLI, the billing system, and the infrastructure, and own it without hand-holding.

The hardest hire, made simple

A founding engineer is a bet. It usually means a months-long search, a senior salary or a real equity grant, and a lot of faith placed in one person. Carl's early concern was exactly that. He wanted seniority, especially for the UI work.

Lyra made a different case: hire for slope and ownership over years on a resume, and embed someone who plugs in like a founding engineer from day one, with none of the search, the equity, or the single-point-of-failure risk. After a transparent conversation, Carl took the bet.

That bet became a ten-month partnership. Rather than a rotating bench, Thunder Compute got one engineer who went deep and stayed. Carl worked with the engineer peer to peer, sharing GPU infrastructure details, CLI docs, and architecture decisions directly in a shared channel, the way a founder works with a founding engineer. The engineer shipped daily, owned entire systems end to end, and became the person Carl trusted with the platform. When the new live pricing experience went out, Carl's response said it all: "it's all live and people love it."

What a founding engineer owns

A founding engineer owns the product, not a ticket queue. Over ten months, that meant the core surfaces of the business:

• Product and frontend. A full frontend rebuild, ongoing console and UX work, and a live pricing experience for customers.
• Developer surface. A new command-line tool, SDKs, and integrations that let developers build on the platform the way they already work.
• Platform and infrastructure. Continuous delivery set up from scratch, a complete billing overhaul, and a cleaner, better-tested architecture underneath.
• Scale and trust. Fraud prevention, safer pricing and admin tooling, and data-driven experimentation across the signup and payment flows.

The engineer became the person trusted to keep the platform running and improving, the role a founding engineer grows into. The trust ran both ways. When Carl needed a marketing billboard turned around overnight, Lyra's design team delivered the same day.

The result

• A product Carl describes as completely transformed: faster, more reliable, and built to scale.
• A founding engineer's worth of product shipped, across the frontend, CLI, billing, SDKs, fraud, and infrastructure, with no founding-engineer search and no equity grant.
• With the product transformed and built to scale, Thunder Compute announced a $13M Series A in August 2026, led by Matrix Partners, with Y Combinator and CEAS Investments joining alongside angels from Cognition, Anthropic, CoreWeave and Together AI.

"Lyra completely transformed our product, rebuilding the frontend, launching a new Go CLI, setting up CI/CD, and improving our workflows. Everything now runs faster, feels more reliable, and is built to scale."

Carl Peterson, Co-founder and CEO, Thunder Compute

‍

What the raise was built on

The round is for the layer underneath the cloud. Thunder Compute's argument is that the GPU shortage is a software problem: most GPUs sit idle most of the time, and virtualization is what lets many workloads share the same hardware. The company spent four years building that layer, and the proving ground was its own cloud, the product that carried more than 10,000 users. That cloud is the product a Lyra engineer spent ten months inside.

On the day it was announced, Carl put the milestone plainly: "Today, we raised a $13M Series A to solve the GPU shortage."

‍

The takeaway

Every early-stage company needs a founding engineer, and almost none of them can hire one quickly. The search takes months, the comp is steep, the equity is real, and the whole bet rides on one person working out.

Thunder Compute skipped the gamble. Through Lyra it got the founding-engineer outcome, one trusted engineer owning the entire product surface, embedded in days, scaling at the founder's pace, with no lock-in. That is the model: not a contractor, not an agency, a founding engineer who happens to come through Lyra. It is how an early YC product became something built to scale, all the way to a Series A.

Ready to build with us?

Whether you're filling an engineering gap or scaling alongside your senior team, Lyra embeds with you from day one.

Let's talk →

Visit Thunder Compute →

‍