Deal of the Week

Warp Sciences  just came out of stealth and announced a $215M round. It’s a CRO (contract research organisation), providing the infrastructure to help bring drugs to market. The bet is that AI can supercharge that process. Designing the molecule is one thing. You still need to test it, prod it and get it into trials. These guys will bring AI to bear on the problem! 

Interestingly, a second-order effect of all this interest in biosciences is that monkey prices have gone vertical, doubling in a year as China pushes into biotech and AI exploration. China currently dominates CRO services, with approx. 74% of the market. Warp wants to challenge that.

Congrats to founding team member Brian Spears, a Berkeley PhD, on this mega round.

There’s a fascinating divide opening up between biotech VCs and AI VCs + AI Frontier Labs. Many biotech VCs won’t touch “AI for biosciences”, leaving generalist and AI investors to fund a lot of these deals. The people who know drug development best seem much less convinced than the people who know AI best. It will be fascinating to see who’s right.

In the meantime, I’ll leave you with this from Matt Gline (a well known biotech operator), CEO of Roivant:

“Biotech is a woodchipper that is built to ingest souls and shred them.”

You can see a version of this divide across the whole AI trade. On the West Coast, the AI-pilled see the crazy advancements and limitless potential, then breathlessly write $10–20M seed cheques at $50–100M valuations on a conversation. The Wall Streeters look on in disbelief: “What terminal valuation are you underwriting here?”

Well, Anthropic’s numbers are getting hard to believe, and the IPO could put that hyper-rational side to the test. More on that later.

Berkeley Brutalist Architecture & Personal Agents Ramifications

I briefly touched on the impact of personal AI agents like Muse and Instinct on the economy a few weeks ago. I then went back and forth on what the killer use case was for these agents, and my usage dropped off a little bit. However, in recent weeks, it feels like more and more of my actions are going through Muse. It just feels like the future.

It is, I admit, fairly trivial stuff (The end of busywork is coming!), but if you add up the time savings, it starts to become consequential. Importantly, more and more things are going through one UI screen: Muse for personal use, and both OpenAI and Anthropic for work.

If this trend continues, the impact is just huge. Patrick Collison (founder of Stripe) captured some of these thoughts, which I summarise below:

I am not sure we will see the best product always win. Capitalism has a way of making money through discrimination.

He also called out Berkeley’s campus’ most ugly building. This is I agree with

Are we entering the era of Technohedgonomy?

Every era seems to find a new way for humans to project power over other humans.

Colonialism → Neocolonialism → Technohegemony (a word I have invented but quite like).

Famous historian Niall Ferguson, Berkeley grad Eric Schimdt, and Lisa Klassen wrote this fascinating article arguing that we are entering a new era. Empires of AI. Or, as I would call it, technohegemony.

The idea is that the US and China, by controlling the infrastructure of intelligence, from chips and models to data centres and capital, will be able to project an entirely new kind of power over the rest of the world.

I am not sure I fully agree with the premise as AI starts to diffuse. But perhaps it doesn't matter. Even the fear of technohegemony will accelerate the deglobalisation of data as nations and economic blocs scramble to maintain their data and tech independence.

But if this era does manifest, who wins between China and the US?

At the moment, China is leading on the power buildout.

The US, meanwhile, is leading on the actual data centres being delivered and promised. Although, of course, those future data centres will require power. A sort of chicken-and-egg problem that may end up being the ultimate bottleneck in the AI race.

A prediction from my side is that we will see the end of Chinese open-source models in the US, especially in large US enterprises. I think US open-source models will catch up (see Muse, Reflection, etc.) over time, or the cost of using frontier labs will drop enough for most tasks.

Meanwhile, US enterprises will increasingly be unwilling to risk data going back to China, especially if this recent Cohere post is true about a model sending data back to China (or my old post about UK equipment sending data to China). Of course, running an open-source model on your own infrastructure is very different from sending data to back to China, but I suspect the perception of risk alone will be enough to change behaviour.

I would be very surprised if Chinese AI systems (and US ones, for that matter!) didn't have some mechanisms built in that could allow information to be sent back to their creators. Whether that is actually happening is another question entirely. A new era of realpolitik!

And while nations compete for control of AI infrastructure, some of the individuals financing this revolution are accumulating extraordinary wealth.

Eric Schmidt is on a heater.

It was recently estimated that he owns ~1% of Anthropic, having invested a large cheque when the company was valued at just $461M. With Anthropic potentially IPOing at $1.5–2T, that stake could be worth $15–20B. Not bad!

Also, a lesson to everyone. Be nice to your exes. According to reports, Eric got into the deal through his ex-girlfriend, who is now married to Anthropic founder Dario Amodei, and recommended the company to him. Eric seems to have a pretty long list of girlfriends. Unique sourcing channel!?

Speaking of Anthropic, they released their S-1 today for what might be the most anticipated IPO of the year (and what a year, with SpaceX's earlier).

The numbers are mind-boggling:

Perhaps the craziest thing is the speed at which Anthropic has reached this scale, started in 2021 they have spent less than $10B to hit a $16B of revenue in the first 6 months of 2026. The speed and efficiency of this company are something else.

Interestingly Anthropic would probably disagree with the article on Imperial AI as the S1 talks of the global benefits of AI to raise economic growth across the world.

Although there is a slight cloud on the horizon. AI cybersecurity company Cathedral announced that it had managed to get into Anthropic's internal systems in just 12 hours. Way to piss on someones Birthday parade on the day they announce their IPO…

It really does seem we are entering a watershed era for cyber.

It reminds me of the start of the First World War, when both sides still had cavalry regiments, not fully internalising that the rules of engagement had fundamentally changed.

The same may be true for cybersecurity in the AI era. We are still fighting with cavalry while the other side has machine guns.

Quick Takes

We wrote about the wonders of the American Financial System and it seems the economist also picked it up with this nice bit of prose:

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

🤝 Sangyo Sosei Advisory. Acquisition  🇯🇵 Technology-focused M&A advisory. 💰 Natixis CIB (acquirer).

🐻 Gary WB Chan, Co-Founder & Managing Partner. BA Sociology. Article

🏭 Zetwerk. $522.9M IPO 🇮🇳 Digital industrial manufacturing platform. 💰 Public-market investors.

🐻 Amrit Acharya, Co-Founder & CEO. MBA (Haas). Article

🧪 Warp Sciences. $215.0M Series A 🇺🇸 AI-powered robotic chemistry labs. 💰 Lightspeed Venture Partners, IVP, Premji Invest.

🐻 Brian Spears, Founding Team Member. PhD Mechanical Engineering. Article

🤖 EKA. $150.0M Series A 🇺🇸 Dexterous robotic manipulation models. 💰 Andreessen Horowitz, F-Prime Capital, E14 Fund.

🐻 Pulkit Agrawal and Tuomas Haarnoja, Co-Founders. Both PhD EECS. Article

🧬 Estuary Biotherapeutics. $93.0M Series A 🇺🇸 Gene therapies for kidney disease. 💰 RA Capital Management, 4BIO Capital, GV.

🐻 Wayne Sackett, Co-Founder & CEO. BS Chemical Engineering. Article

💻 AheadComputing. $63.5M Series A1 🇺🇸 High-performance RISC-V processor cores. 💰 Eclipse, Maverick Capital, Toyota Ventures.

🐻 Debbie Marr, Co-Founder, President & CEO. BS EECS. SEC Filing

👁️ Maxinsights. $30.0M Seed 🇺🇸 Training data for physical AI. 💰 Peak XV Partners, South Park Commons, Alumni Ventures.

🐻 Yuguang (YG) Yuan, Co-Founder & CEO. MBA (Haas). Article

❄️ Pascal. $8.5M Seed Extension 🇺🇸 Energy-efficient solid refrigerant technology. 💰 Sazze Partners, LB Investment, KDB Silicon Valley.

🐻 Jarad Mason, Co-Founder & Chief Science Officer. PhD Chemistry. Article

🧫 Epicrispr Biotechnologies. $8.0M Grant 🇺🇸 Epigenetic therapies for genetic diseases. 💰 California Institute for Regenerative Medicine.

🐻 Lei Stanley Qi, Scientific Founder. PhD Bioengineering; MA Physics. 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.

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. 🐻