Built By Berkeley

Deal of the Week:

We have been banging the drums that Berkeley is ground zero in the AI revolution for a while, and the last couple of weeks have confirmed that Berkeley is the undisputed leader in open source. A couple of weeks ago it was LMarena’s $150M round at a $1.7B valuation, and this week we have a double whammy:

• Inferact – Raised a $150M seed round at an $800M valuation to commercialise the vLLM open-source project. The idea is to bridge the gap between infrastructure (different types of compute hardware) and models, enabling models to run efficiently across different types of hardware.

• RadixArk - Raised an undisclosed round at a $400M valuation led by Accel. RadixArk is commercialising SGLang, which is also focused on improving inference workloads.

Inference is moving into the mainstream. Nvidia paid up with the $20B for Groq’s inference chips at the end of last year, and now VCs are funding the next wave of software to help optimise this. The first big use of compute was training models; now you actually need to use the models (inference). And if you believe Jensen Huang, this will go up a billion-fold! 

What do all of these—and the $134B valued Databricks—have in common? They all started as open-source Berkeley projects, and they all have Ion Stoica at the heart of it. Just another day in the life: $137B of valuation across four startups in a few weeks for Ion. He’s on a serious heater. 🔥

Brex Hit Gravity. AI Hasn’t

Brex: $12B → $5B

Failure? No. Gravity.

When a business model is understood, valuation follows. Interchange has known physics. Brex got what interchange businesses get.

AI has no physics yet. Which means today's valuations could be brilliant or absurd. We'll find out when the physics are discovered.

I decided this full article was to long for the newsletter so read here.

Velociraptors come to life?

A year ago I made a prediction that we were at the ‘The Dawn of the Age of the Velociraptor’ 🦖🦖🦖 

The idea was that in the bygone era of Unicorns, startup success was measured by capital raised and paper gains. But with the end of the ZIRP (Zero Interest Rate Period) era and the rise of AI-native companies, we are entering the Age of the Velociraptor. (Yes, I’m using the Jurassic Park version—more cinematic, less scientific.)

What is a Velociraptor?

These are AI-native companies growing at insane speed with extremely small teams. Think of them as lean, AI-supercharged, and relentlessly aggressive.

Definition: A company with fewer than 50 employees generating $100M+ ARR ($2M+ per FTE).

We are in the early innings of this epoch, but the signs are there. Full post here (including risks and why now) from early 2025.

But how is the prediction looking? It seems to be on track. a16z recently released a state of the markets r report, and ARR/FTE per employee at startups in the 90th percentile exploded in 2025. In 2024, it was $400K per FTE, and in 2025 it jumped to $700K for the top-performing startups. This is insane acceleration! 🚀

I now think I was wrong about the number of employees being as low as 50 (you still need people 🙂), but I think I was right about the high ARR/FTE, and we will continue to see this through AI-driven efficiencies.

Quick Takes

• Lovely story of some Berkeley hustle (and bureaucrats taking them down…boo). Someone fund these guys (me?!)

• A big debate going on if we are seeing the death of SaaS with AI, the team at Avenir see it more as a fork in the road. Nice presentation.

• Great marketing by the founder of GPTZero but if true AI slop continues to gain traction. Lets end AI slop now!

Summary by the #️⃣ & 💰:

• 5 Berkeley-founded companies funded

• $205M of capital raised from the 19th Jan to 25th Jan

💡 Got any ideas or feedback on how to improve this weekly digest? Just hit reply.

Acquisition

🤖 Diligent Robotics. Acquisition 🇺🇸 Healthcare robot assistants. 💰 Serve Robotics

🐻 Vivian Chu, Co-Founder & CIO. BS EECS Article

Closed Rounds

🖥️ Inferact. $150M Seed 🇺🇸 LLM inference optimization platform. 💰 Andreessen Horowitz (a16z), Lightspeed Venture Partners, UC Berkeley Chancellor’s Fund

🐻 Woosuk Kwon, Co-Founder. PhD Computer Science Article

🧬 Think Bioscience. $54.8M Series A 🇺🇸 Microbial drug-discovery platform. 💰 Regeneron Ventures, Wireframe Ventures

🐻 Jerome Fox, Co-Founder & CEO. PhD Chemical Engineering Article

🔬 Spheryx. $0.5M Seed 🇺🇸 Particle suspension analytics technology. 💰 NCATS, U.S. Department of Health and Human Services, National Science Foundation

🐻 Laura Philips, Founder & CEO. PhD Chemistry 

📊 ARTI Analytics. $0.5M Seed 🇺🇸 AI decision-support software. 💰Undisclosed

🐻 Joerg Ferchau, Founder & CEO. MEng Material Science Article

🧠 RadixArk. Seed 🇺🇸 Frontier AI inference infrastructure. 💰Accel 

🐻 Banghua Zhu, Co-Founder & CTO. PhD EECS Article

Our goal is to document the startup ecosystem of Berkeley-founded companies. Please share this newsletter with any Cal Bears in your network so we can crowdsource information about all investment rounds and job opportunities.

• Did we miss a company or want to announce a round? Add it here, and we will post it next week.

• Do you have a job you want to post from a Berkeley Company? Add here.

• Is there someone in the Berkeley ecosystem that you would want us to do a profile on? A founder, funder, or general startup person? Add here.

Built By Berkeley is not affiliated with UC Berkeley, but maybe we will be one day if we get enough subscribers….

Built By Berkeley, where we announce all the funding rounds by Berkeley-founded companies. This is a community effort, so please let us know if we missed a company here. 🐻