Table of Contents
Jon Brennan-Badal founded Opentrons 8 years ago. The idea: robots that could do the practical science in the lab. Now, they have suddenly emerged as the leaders in the field.
Their mission: democratize access. For now, that means make lab automation as easy and affordable as possible. During the pandemic, it took little over 40 Opentrons robots to provide the majority of COVID-19 testing for the city of New York - 15 million people counted on them. Top-tier investors have piled in, will their bets pay off in 2025?
How did you found Opentrons?
Sure, I can tell you from the start. Long before I co-founded Opentrons, I graduated as a History major from Columbia. Straight out of university, I co-founded Comixology, a digital comic reader platform, which was acquired by Amazon. I learned three things from that experience. First, learning how to actually build a start-up. Second, how to compete in very competitive markets. And third, how to build a platform company. I took those key learnings and applied them to Opentrons. Our fundamental mission has stayed the same since the beginning: democratize access to life science technology.
Technology tends to be too esoteric, and too difficult to use, whereas it should be simple and easy to incorporate into operations. So we started with automated liquid handling, which we knew would save scientists a lot of time in the lab. To do this, we brought in the key members of the founding team that had developed the first-ever affordable 3D printer - MakerBot. That was our first step, to get the most qualified people on board.
How does the Flex Prep differ from the Flex?
With the Flex, as well as the Flex Prep, you get a fully featured liquid handler that's highly capable and highly affordable. From a usability perspective, our machines require very little technical expertise.
In the case of the Flex Prep, you walk up to that robot, you program a new sample prep protocol - it literally takes a minute or two using the touchscreen - and you press run. That’s it. And the Flex Prep starts running through whatever combination of lab activities you need it to do.
The Flex Prep totally changes the game - it has extra capabilities for sample preparation. So I think, and as far as I know for the first time in history, with the Flex Prep you get a full-blown liquid handler that is just as easy for any scientist to use as it would be to manually pipette, but it does everything for them, so they can spend their time elsewhere in the lab.
The Flex, of course, also provides all these no-code capabilities - if you have one, you just need to run a software update. And the Flex can automate more complex workflows, like NGS library prep.
Lots of companies in this industry build great technologies, but too often they're very expensive technologies, so only a very small percentage of scientists can access them. Our underlying philosophy is that we think a lot more high-level scientific discovery can be unlocked by making simple solutions that automate key tasks. But they have to be ones that most labs can afford and invest in, and crucially, they need to be easy to use.
What is the main innovation behind your technology that might make you succeed where others have failed?
So we talked about focusing on ease of use. The other key dimension for accessibility is affordability. And we, as a company, have made significant investments over the years in reducing the fundamental cost of building these types of robotic systems. And so that has caused us to have very extensive vertical integration.
We spend our time not just designing the robotic system but also designing and manufacturing all the key components that go into that robotic platform. And those are the kinds of activities that no one else in this space does.
As a result, we do not leverage any of the traditional supply chains that other competitors do. We built all of that up from scratch, from first principles, and by doing that, we're able to very significantly reduce the fundamental cost of these systems, and ensure that these systems are designed for mass manufacturing. With our Flex platform, we brought that to market, and within the first year of introduction, it is now the top-selling robotic platform in that middle-market segment. Even just scaling up that manufacturing at that pace and at the volumes that we're doing is fairly unprecedented in this space.
How might a traditional lab scientist make use of the Flex?
It depends on the specific kind of application they're looking to use it for, but the experience end-to-end is fundamentally easier than other solutions in the market. So first, installation of the robot—we can provide services to install that robot in your lab, but it's an easy enough process and many end-users choose to do the installation themselves because they can take it out of the box and have it up and running in an hour or two.
Second, the development of the applications that can be run on the robot. There have been thousands of applications built for our platform, and many of them can be used by end customers because our partners and customers have contributed back to the platform. So oftentimes, customers can use or modify existing applications for their own needs, and of course, use them in no-code environments. That also accelerates and simplifies the experience.
And then, of course, with the day-to-day running of the robot where you might just need to make some minor tweaks to the protocol—you have a different sample count or whatever—it's all centered around highly intuitive experiences on the touchscreen. You don't need to have the training to take advantage of some of these advanced settings because we spend time on these usability studies and all of these sets of activities so that scientists can make use of the experience without needing any kind of extensive training. It works and the idea is it should be easy to use, just like someone's iPhone is easy to use.
How do your robots compare with humans in terms of time and accuracy?
From an accuracy perspective, our type of system has comparable accuracy to systems that are far more expensive. And of course, the error rates relative to human pipetting are much lower. It depends on the experience and accuracy of the individual bench scientist, but if you think about the activity of, let's say, doing a full plate of NGS library prep—which involves a countless number of steps over what can be an eight or 12-hour time period by someone manually pipetting—it's not uncommon for there to be mistakes or multiple mistakes across that type of assay, which by automating, you eliminate.
From a throughput perspective, if you configure the Flex with a 96-channel, you can be doing orders of magnitude—10x more—the throughput relative to what an individual can do in the same timeframe.
What is your vision for Opentrons in 2025?
Our main story is bringing automation to labs for the first time and expanding that accessibility. And so right now, the majority of our customers are new to automation. We see 2025 having two major themes. One, continuing to drive significant additional market adoption of these types of tools to labs that previously couldn't afford it (or for those whom it was previously too technically challenging to adopt).
Second is continuing to bring to market some amazing AI-based products where you can use natural language to program and optimize your protocols. Those kinds of investments will further reduce the friction associated with using lab automation. We think that using a graphical interface makes these systems much more accessible, but being able to just give a simple natural language command to program a new application is fundamentally easier than even using a graphical experience to program a protocol. So we see 2025 being a story of changing just how easy it is to use automation by using NLP.
Looking to the future, what are the trends in the wider life sciences industry do you expect to see over the next few years?
I would bet that the industry will double down on AI. We see AI continuing to be transformative in programming the Flex. Increasingly, there's an ecosystem of AI startups that have been building applications on top of our platform. One of the significant trends that we've seen—there are several VC-backed startups building on our platform—has been building AI co-scientists that can program and develop new experiments or protocols autonomously.
A very simple application can be, say, you want to incorporate a new workflow in your lab, so you choose your reagent kit that might be instrumental for that. Using our platform and some of these AI co-scientist-based LLMs, even right now you can have that robot run autonomously overnight and optimize the yield of that kit or whatever specific things you're looking to optimize in an autonomous fashion. I think exceptionally so next year, we're going to see a growing set of actually useful applications of using an AI co-scientist coupled with our platform to enable autonomous optimization of protocols, as well as a growing catalog of even more complex applications and experiments.
