Benchling CEO on bringing AI to biotech cloud workflows after a brutal industry downturn

Up next, we have the co-founder and CEO of Benchling calling in from the JP Morgan Healthc Care Conference. We are very excited to welcome [music] Saji to the show. Uh

welcome to the show.

Welcome to the show. How are you doing?

I'm doing well. Thanks for having me guys.

Thanks for hopping on the show. Uh first, uh I mean kick us off with an introduction on yourself and Ben Schling. I think most people are familiar, but we'd love to get an update on the shape of the business today. And then I want to go into the JP Morgan Healthcare Conference as well.

Sounds sounds great. Yeah, I'm the co-founder and the CEO of of Benchling. Uh we make software that helps scientists uh discover and develop new medicines. We spent about 10 years helping bring science to the cloud. It was pretty much trapped on premise and in spreadsheets before that. And then today we're very focused on bringing AI to scientists everywhere by putting models and simulation directly in the workflow of scientists and building agents that automate all the toil and drudgery that's that drug discovery is filled with.

Yeah. Uh, someone was pitching me cursor for bio and I said that can't happen because there's no VS code for bio and [laughter] you said love open source or hate it like it's just it seems like it's just a a fact that if you don't can't fork something quickly uh how are you going to spin up and grow as quickly as cursor has grown? Uh has that proven true? Do you feel like you're the cursor for bio and that uh your setup structurally

they're the benchling for benching.

Yeah. Yeah. Really? I mean, I shouldn't even try. But walk me through that.

Yeah. Uh, but walk me through what you're actually building. Yeah.

Yeah. I I think AI in bio is kind of like it's kind of like uh GPT without the chat. So, GPT came out and like it didn't explode until someone put the right interface on it. That's how I feel a lot of AI in bio where you have these amazing new models and capabilities for doing different tasks. But making a medicine is so hard and complex that putting it all together in the right interface and making it super easy for the scientists in the labs to actually use is a really difficult problem and that's what we're focused on at at Benchling. And what's the uh we've seen in coding which I think people are most familiar with uh you know autocomplete gets better and better then you get uh sort of the back and forth the copy paste from the LLMs then you get the fully agentic prompt systems clawed code and uh is there something similar happening where there's an autocomplete for bio before there's an a an agent for bio like how what is the shape of the interaction that the scientist is having with the various ways you can interact with AI models. I

I think there's even lowerhanging fruit than that. Uh we have customers where they're using our AI agents to ingest all this data that's trapped in spreadsheets and PDFs and sharepoints that's you know been you know it's 6 10 15 years old. The people who made the data have like long moved on and they're able to understand what's been done before and sometimes not even rerun experience. We had a we had a customer who, you know, brought 10,000 different experiments into benchling that previously hadn't been touched and a scientist was about to run some studies that would have taken about 8 months and they realized that someone else had already done the work many years ago and they were able to cut a bunch of them out and like life science is just full of examples like that that's going to help uh compress the time it takes to make drugs and get them to market.

Uh where do you

We're not even at cursor yet.

Yeah. Where do where do you think AI is overhyped in uh life sciences and where do you think it's underhyped?

Oh man. Um I think it's I think it's there's a lot of excitement. I don't want to say overhyped. There's a lot of excitement about these models that help you uh create a new hypothesis and that's really important but drug discovery is 9,999 steps after that. many of which are incredibly hard and and full of toil. And there's a lot of opportunity especially when you get to things that are regulated or in animals or you know closer to manufacturing to make huge differences uh in the lives of scientists everywhere to get get medicines to patients faster.

How do you think about speed of token generation? We were talking to the uh the CEO of Cerebras and he was very optimistic about uh applied AI speeding up in bio specifically because more and more of the questions that scientists are asking look like deep research reports whereas consumers might just ask oh you know what's the capital of I don't know Illinois and it's a quick you know LLM it just fills the next word in it takes one token uh but a lot of scientists might say for every question that I'm asking throughout the day, every 15 minutes, I'm effectively firing off a deep research report that might take a half an hour and so speeding that up makes sense. How are you feeling about the different uh different options on the silicon side? What partnerships have you done? How are you thinking about accelerating the actual workflows?

Yeah, we've got partnerships with both uh Nvidia and Anthropic, which I can talk about in just a second. But I think life science is actually not even at the point where you know token speed is the rate rate limiter. Um obviously you know in in France San Francisco AI is everywhere your billboards are selling you GPUs and and whatever but if you go to Boston other life science hub or I spent time in Europe in the last couple months I was just in Asia in December visiting biotech and pharma companies. AI adoption is not going to just happen spontaneously in all these organizations. Most of these scientific organizations actually are just getting like co-pilots and whatnot turned on. There's a lot of legal, safety, regulatory, security concerns that they have. And scientists workflows are so so complicated. They're already jumping between a bunch of different tools with really complex data that AI is not necessarily purpose-built for. And so there's a lot of like last mile work to do before token speed even becomes a problem. And that's that's what we've been trying to focus on at Benchling.

Is the decline of Boston as a biotech hub overstated? We read a piece by Will Manitis uh declaring Boston dead almost. He was very dramatic about it. Uh how are things going in Boston? Is it still competitive place? Uh there's been a number of different reports around hiring maybe, you know, uh weakening and softening. Uh but what are you seeing in Boston? Uh, I think there's still a lot of like really good scientific talent in Boston and they can come out to San Francisco to raise money if they need. They're just a they're just a plane flight away. I think maybe zooming out what you're probably hearing is like biotech went through a like it's it's a pretty good vibe right now at JPM, but biotech went through a really tough couple years. Like they the bioarm industry went through like the.com bust equivalent for a few years. You know

what was the driver of that? because it feels like during COVID there was a lot of biotech energy and then through the GLP1 boom we're seeing more but uh is this crisper overhang like what what is the what is the foundational technology of the of the.com if the if the internet was the thing that ultimately caused the bubble of

there were a lot of great new scientific technologies but it takes a lot of time to make a drug it takes seven 10 years the industry how many how many drugs do you do you guys think get approved a year. Can you guess?

Under hundreds. I don't know.

Under 20.

What is it?

Uh there's for the last 10 years, we've gotten about 50 new medicines approved per year. Just 50. Uh and how much do you think the industry spends on R&D?

Is it like$10 billion?

100.

Uh the industry spends almost $200 billion a year on R&D. And you get off and you get 50 you get 50 new drugs a year. And it takes 10 years to figure out if you're right. Okay, so like

uh and by the way though, like drugs, drugs are amazing. Like we want more drugs. Uh and

uh drugs get cheaper over time. They become generic. Hundreds of millions of people take statins and all kinds of other drugs that they don't even think about how cheap they are and how impactful they are. So we get to get to add 50 to our stockpile every single year of new life-changing medicines. We should want more of them. We should want them faster. We want them to be better. And I think because it takes so long though and it's such a difficult artisal process like I think everyone got excited with mRNA and co and I think a lot of money went in and uh I think science takes time to become durable and turned into businesses and to have medicines get commercialized and I think some people a couple years after co also got impatient and there's other things to invest in and so the money kind of went back out and so it was a really difficult couple years for the the industry and you know interest rates geopolitical things if there's any industry that doesn't like uncertainty. It's biotech. You know, you're you're underwriting something that's going to take 10 years to see the results of.

Yeah. How should people think about uh JP Morgan Healthcare, like the conference? Is it uh is it an opportunity to raise money? Is it an opportunity for companies that are thinking about going public? I know the the financial profile of a lot of drug companies is very different than startups at the at the same time. There's a lot of companies that are now taking a more traditional Silicon Valley path uh in in growing their business and raising money. Uh what's the purpose? Why do people go to JP Morgan Healthcare? What do they get out of it?

It's actually really funny. A lot of people are here for JPM, but they don't actually go to JPM. Uh it's like every every conference kind of has their ratio of at the conference versus ecosystem around it. JPM might be the most lopsided where like 90% of people here are not going to the actual conference. It is a banker conference for CEOs and CFOs to raise money but it's it's like this barometer for the industry and most people are actually here for all of the ecosystem around all the events like we're having a event at our HQ tonight that's like a AI and bio panel with you know Eli Liy and Isomorphic Labs and ProFluent and Anthropic um but you know people are here to to get customers to fund raise to see their friends and to their pharma companies that have alumni reunion parties here uh and so it's like it's just the gathering to kick the you're off and and the mood is good, by the way. Like the XBI has been trading well. There's been a bunch of M&A in biotech which has sent a bunch of money back to investors. I think some of the tariff and MFN stuff has settled a little bit for now. Eli Liy, as you said, becoming the first trillion dollar pharma company. I think that's raised everyone's ambitions a bunch. And then, you know,

yeah,

there's there's a benching next.

Yeah. Benchling next. It's easier now since, you know, it's always the fast follower.

Yeah. So, Exactly. Have

you back on when uh when you hit the one T. But uh I'm so glad to hear things are going well. Thank you so much for stopping by.

Yeah. Great to meet you on the show. We've been wanting to make this happen for a long time.

Yeah, this is great.

Thanks. Thanks for having me. You guys need more biotech on the show. TBP,

you're new you're our new correspondent expert. So tell us the the companies that we should have on.

Yeah, we'd love that recommendations.

Enjoyed this.

All right. I'm here for a vibe check anytime you need it.

Fantastic. We'll talk to you soon.

Cheers. Goodbye. [applause] Um, well, we have a very special guest in the TBP and Ultra Realm live in