Inside a $300m bet on AI for materials discovery

Summary

---

Podcast Summary: Catalyst with Shayle Kann

Episode: "Inside a $300m Bet on AI for Materials Discovery"

Date: November 6, 2025
Host: Shayle Kann (Energy Impact Partners)
Guest: Doge Chubuk (Co-founder of Periodic Labs, former Google DeepMind)

---

Overview of the Episode

This episode delves into the intersection of artificial intelligence (AI) and materials discovery with a focus on the ambitious $300 million seed-funded startup, Periodic Labs. Shayle Kann invites guest Doge Chubuk to discuss how advancements in large language models (LLMs) and hardware automation might enable breakthrough discoveries in materials science—potentially even the holy grail: room temperature superconductors. The conversation explores what's changed in AI over the past year, the challenges of scientific reasoning, the laboratory-in-the-loop approach, and the prospects and business models for AI-driven R&D.

---

Key Discussion Points & Insights

1. Catalyst for Starting Periodic Labs

[04:22 - 06:47]

• Advances in LLMs and Reasoning Abilities:
  Doge notes that since their last conversation (a year prior), LLMs—especially those with advanced reasoning (e.g. OpenAI's "01" model)—showed dramatic improvements, particularly in complex mathematical and scientific reasoning. This was a crucial factor inspiring him to launch Periodic Labs.

  "The LLMs have improved even further... The reasoning models have gotten good enough that they are able to sort of get around that [training data limitation] challenge via reasoning."
  — Doge Chubuk [06:47]

• Out-of-Distribution Generalization:
  The AI breakthroughs offer hope in moving beyond simply interpolating known data to reasoning more like scientists and making educated "jumps" to new discoveries.

2. Hybrid Lab Approach: Digital + Physical

[09:19 - 11:33]

• Combining Physical and Digital Experiments:
  Periodic Labs aims to merge frontier AI research with an advanced physical laboratory—feedback loops where LLMs propose experiments, run simulations, and digest results to inform further research.

• Automating the Workflow:

  "The LLM can propose, for example, synthesis recipes, or simulations to run... The LLMs are pretty good at tool use, it can actually do it itself..."
  — Doge Chubuk [09:56]

• Bottlenecks in Automation:
  While high-throughput experiments are becoming commoditized, full automation of characterization (assessing the results of experiments) remains a challenge, but one they hope to soon overcome.

3. Targeting Superconductors and ‘Breakthroughs’

[12:08 - 16:37]

• Why Superconductors?
  The team prioritizes superconductors because a genuinely novel discovery—even prior to real-world deployment—would be immediately impactful for science and tech (e.g. quantum computing, fusion).

  "If somebody discovers a room temperature superconductor today, even before it makes it into a product, it has huge impact. First of all, it changes how we think about the universe."
  — Doge Chubuk [13:47]

• Path to Breakthroughs vs. Incremental Progress:
  Doge cautions that LLMs (and humans!) are still limited by their training data; thus, true breakthroughs remain difficult but the automation and scale at which they operate increases the odds of stumbling upon big discoveries.

  "I actually would guess that there's a law out there that we haven't discovered yet that says that you can't just look at your training set that's different than what you're trying to discover and just predict it."
  — Doge Chubuk [12:58]

4. Reward Functions and AI-driven Discovery

[20:45 - 21:34]

• Reward Design for LLMs:
  Instead of chasing only room temperature superconductivity, they optimize for various valuable traits (e.g. higher critical temperature, ductility, critical magnetic field).

  "For real life experimental measurement of Tc, it's much harder to reward hack, which we love... it's a nicer, unhackable reward."
  — Doge Chubuk [20:51]

5. The Human in the Loop & Hypothesis Generation

[22:18 - 23:25]

• Humans vs. AI:
  While automation is a goal, Doge is pragmatic—human scientists remain crucial in hypothesis generation, an area where current AI is still lacking:

  "There are things that ML/AI is better than humans, but one of those things is not hypothesis generation."
  — Doge Chubuk [22:18]

6. Cost Structure & the $300M Seed Round

[24:07 - 24:55]

• Huge Compute & Hardware Costs:
  Training LLMs and running experiments both drive costs. GPUs for LLMs/simulations are surprisingly among the biggest expenses, even greater than physical labs in some scenarios.

7. Generalist vs. Specialist Models

[25:20 - 27:51]

• Why Cross-disciplinary AI Matters:
  The complexity of modern science means no human can master all relevant fields, but a well-trained LLM could possess deep, cross-domain knowledge—potentially making the most exciting discoveries "at the fringes between disciplines".

  "A lot of the exciting discoveries happen to lie in between fields... we're excited about an LLM that can basically do that at a scale that humans couldn't yet."
  — Doge Chubuk [27:51]

8. Synthetic vs. Physical Data

[27:51 - 30:10]

• Emerging Value of Lab-generated Data:
  Chubuk reflects on how the unique experimental data generated in their lab could drive foundational scientific insights unavailable anywhere else.

  "There are certain experiments that told us so much about how we understand ... the universe."
  — Doge Chubuk [28:22]

9. Business Models: Licensing, Productization, and R&D-as-a-Service

[30:10 - 34:02]

• R&D as a Scalable Service:
  Short-term, providing unique, science-aware LLMs to other R&D organizations could be lucrative. Longer-term, if Periodic Labs can repeatedly pioneer materials, it could redefine the industry (akin to Genentech in drug discovery).

  "We already see this big need and a big potential for impact by providing these LLMs to do physical R&D."
  — Doge Chubuk [32:56]

• Analogy to Cloud Infrastructure:
  Shayle likens their offering to AWS for R&D—leveraging scale and unique infrastructure (AI + lab) to serve many clients.

10. Looking Forward: LLMs and Future Breakthroughs

[34:46 - 36:48]

• Riding the Tide of Platform LLM Advancement:
  Ongoing improvements from the big LLM developers (like OpenAI) help Periodic Labs, especially as LLMs get better at code, simulation, and scientific reasoning—a rising tide that could lead to the next great scientific leap.

  "We actually basically rise with the tide, right? Like, as LLMs get better, there's so many advantages of that to other applications."
  — Doge Chubuk [35:27]

---

Notable Quotes & Memorable Moments

• On the Limitations of AI Reasoning in Science:
  "There's a difference between winning gold medals in Math Olympiads and scientific discovery... You can't really practice how to discover the next big theory."
  — Doge Chubuk [00:11 & 07:49]

• On Choosing Superconductors as a Focus:
  "Superconductivity is a bit like that. To discover an exciting superconductor, we probably have to develop so many capabilities on the way there that's by themselves very useful."
  — Doge Chubuk [15:48]

• On the Value of Interdisciplinary Science:
  "Science is kind of like a fractal in the way it's hierarchically organized and there's so much surface area that humans have exploited, of course, but then there's probably so much left to exploit, and we're excited about an LLM that can basically do that at a scale that humans couldn't yet."
  — Doge Chubuk [27:51]

• On Business Excitement and Company Culture:
  "We are hosting weekly seminars where the physicists will teach the computer scientists about the physics and the computer scientists will teach the physicists about LLMs. And of course there are a lot of people in between. It's actually again, a fractal."
  — Doge Chubuk [34:14]

• On What Could Change in AI Next:
  "One of them could be things like hypothesis generation or more auto domain generalization... because that's what you're kind of trying to get at with your reward."
  — Doge Chubuk [36:11]

---

Important Timestamps for Segments

• Start of Episode / Introduction: [00:07–02:29]
• Background: Why Periodic Labs, What's Changed: [04:22–06:47]
• LLMs + Physical Labs, Closing the AI/Experiment Loop: [09:19–11:33]
• Superconductivity as Breakthrough Target: [12:08–16:37]
• Discussion of Scientific Reasoning Limitations: [12:58–13:56]
• Reward Functions & Incremental vs. Breakthrough Progress: [20:45–21:34]
• Role of Humans vs. AI in Research: [22:18–23:25]
• Funding and Resource Allocation: [24:07–24:55]
• Generalist vs. Specialist Scientific AI: [25:20–27:51]
• Synthetic vs. Physical Data Value: [27:51–30:10]
• Business Models for AI-driven Science: [30:10–34:02]
• Looking Ahead for AI in Science: [34:46–36:48]

---

Summary Tone & Takeaways

The episode is both optimistic and sober, reflecting the cutting-edge yet challenging state of AI-driven scientific discovery. Doge and Shayle are pragmatic about the barriers ahead—especially the difference between automated incremental progress and true paradigm-shifting breakthroughs—and yet excited by the rapidly increasing capabilities of modern AI when paired with new experimental approaches. The vision is huge: creating the infrastructure and know-how for AI to make real-world, not just digital, scientific discoveries at a pace and scale previously unimaginable.

---

For anyone interested in the fusion of AI, scientific discovery, and climate technology, this episode offers both a visionary roadmap and a clear-eyed view of the challenges that lie ahead.

Transcript

[Latitude Media Announcer]

Latitude Media, covering the new frontiers of the energy transition.

[Shail Khan]

I'm Shail Khan and this is Catalyst.

[Doge Chubuk]

I have to say there's a difference between winning gold medals in Math Olympiads and scientific discovery. Like you can practice for Math Olympiads by studying previous year's problems. You can't really practice how to discover the next big theory, but they were getting better at reasoning on complex problems.

[Shail Khan]

Coming up can AI discover a room temperature superconductor? Volume 2.

[Latitude Media Announcer]

The AI boom is here, but the grid wasn't built for it. But Bloom Energy is helping the AI industry take charge. Bloom Energy delivers affordable, always on, ultra reliable on site power. That's why chip makers, hyperscalers and data center leaders are already using Bloom to power their operations with electricity that scales from megawatts to gigawatts. This isn't yesterday's fuel cell. This is on site power built to deliver at AI speed. To learn more about how fuel cells are powering the AI era, visit bloomenergy.com or click the link in the show notes.

[Energy Hub Announcer]

Surging electricity demand is testing the limits of the grid, but Energy Hub is helping utilities stay ahead. Energy Hub's platform transforms millions of connected devices like thermostats, EVs, batteries and more into flexible energy resources. That means more reliability, lower costs and cleaner power without new infrastructure. Energy Hub partners with over 120 utilities nationwide to build virtual power plants that scale. Learn how the industry's leading flexibility provider is shaping the future of the grid. Visit energyhub.com Clean energy is under attack.

[Latitude Media Announcer]

And it's more important than ever to understand why projects fail and how to get them back on track. The center left think tank thirdway surveyed over 200 clean energy professionals to answer that exact question. Their newest study identifies the top non cost barriers to getting projects over the finish line, from permitting challenges to NIMBYism. To read Third Way's full report and to learn more about their PACE initiative, visit thirdway.org pace or click the link in the show Notes.

[Shail Khan]

I'm Sheil Khan. I lead the early stage venture strategy at Energy Impact Partners. Welcome. So a year ago, a little over a year ago, I had Doge Chubuk on this podcast to talk about using AI for materials discovery, which has all sorts of interesting applications in the spaces that we talk about here. At that time. Doge had been leading efforts in that area for Google DeepMind for some time and I thought of him as being both very knowledgeable in the space, obviously also pretty sober about it. Fast forward a year, Doge left Google DeepMind earlier this year and along with Liam Fettis, who was one of the co creators of ChatGPT, started a company called Periodic Labs which raised, wait for it, a $300 million seed round like led by Andreessen Horowitz, Periodic is doing AI for materials discovery. And not just that, also physics and chemistry. And they're also very much hardware in the loop. The way I like to frame it is that they are building two kinds of frontier lab at once. There's a frontier AI lab and a frontier scientific lab, the type of lab that we used to talk about and then they're trying to make those two things work together to make breakthrough discoveries. Notably, one thing we talked about last time was how the, the AI materials discovery companies at the time tended to start by going after often discovery of something like metal organic frameworks or MOFs for carbon capture, which I think of as less of a breakthrough opportunity really from a global scale. Whereas the big, perhaps biggest breakthrough to prove would be the discovery of a room temperature superconductor. Well, Periodic makes no promises, but they're very publicly working on high temperature and maybe room temperature superconductors. Based on that last conversation, to be honest, I wouldn't have predicted this. So it was time to have Doge back on and hear what changed and how. Here's Doge. Doge, welcome back.

[Doge Chubuk]

It's great to be back.

[Shail Khan]

It's great to see you. A lot has changed since the last time we talked, so I looked back. So the last conversation that we had was just over a year ago. It was September 2024 and I was having you explain to me the wild world of AI for materials discovery in particular, and the work that you'd been doing at Google DeepMind, but also just like the broader landscape. And I'll tell you my takeaway from that conversation, which you could tell me if I had the wrong takeaway at the time, but my takeaway was promising field, pretty unclear if and when this new wave of LLMs and all the reinforcement learning, all the things that have shown up in the past few years, pretty unclear if and when that would generate a real meaningful breakthrough discovery in materials. And we, we talked through a bunch of the reasons why it's challenging. Training data maybe chief amongst them, but, but I came away with a pretty, I think like a sober view of the, the path there. Okay, so fast forward a year and you left Google DeepMind, started a company to do that amongst other things. So I guess the first question that I have for you is what changed in the last 12 months to give you conviction that now is the time.

[Doge Chubuk]

Great question. So when we talked, I was doing research in the field of computational material science and machine learning specifically. We were using graph neural networks, we were using density functional theory, and we were trying to discover materials. One thing that changed since our discussion was the LLMs have improved even further. So at the time I wasn't using LLMs much at all. But I think right around when we were talking, the O1 came out right, the reasoning models started showing up and that was a huge update for me because you might remember that one of my big concerns is machine learning works best on the training set distribution. But in science and technology we almost only care about auto domain generalization. So what Ovan showed is if you spend test time compute, you can get better results. So that was very exciting to me because there was one way of investing resources that was beyond the training set.

[Shail Khan]

So, okay, if I can try to translate that into layperson terms. The reasoning model, like OpenAI's 01 model introduced, unlocked a door kind of that maybe allows you to break this challenge of the limited training data set that you have in materials discovery. That was what we spent a lot of time talking about a year ago was like, you know, you can compare the corpus of data that an LLM trains on to, to do language, which is enormous, it trains on the Internet basically versus the corpus of data that you were dealing with in trying to discover novel materials. And it was thousands of data points, not tens of billions or whatever. And so that presumably hasn't changed at least yet. But you're saying that the reasoning models have gotten good enough that they are able to sort of get around that challenge via reasoning or possibly generating their own synthetic data. Like, what is it that allows them to, to break that?

[Doge Chubuk]

So I'm not saying that they're good enough already, but that was one step in the positive direction. And another thing we've seen is they've gotten really good at math. So since last time we talked, they started winning gold medals in math Olympiads and they're doing similarly well on coding, really well on physics Olympiads. And, and I have to say there's a difference between those things and scientific discovery. Like you can practice for math Olympiads by studying previous year's problems. You can't really practice how to discover the next big theory. But it did show you that they were getting better at reasoning on complex problems. So then what else do we need? So I think the biggest thing we need is to have our own lab because once you have a very intelligent reasoning LLM, you still can't discover things unless you make trials, right? Just like humans, the LLMs will be wrong often when they try to predict things outside of their training set. But you try many things and then at some point you get a really cool discovery. And this is, as we talked about history last time, this is quite common in solid state chemistry, solid state physics, but a lot of discoveries happen somewhat by accident, but of course with a lot of background understanding of the physical system and a lot of trial and error.

[Shail Khan]

So, okay, so this is what you're doing at periodic, right? You're sort of combining the digital domain with the physical domain. You have a lab in both sense. It's a frontier lab and an LLM sense, and a frontier lab in a laboratory sense, in the traditional sense of the word. And you're sort of merging the two. I'm curious in practice how you imagine that feedback loop working. So is it a traditional. You develop a theory, you run an experiment, you generate data from that experiment, but in this case you feed the experiment back into your customized LLM as an additional set of training data and then that's the way that the loop works, or is it more complicated than that?

[Doge Chubuk]

Yeah, exactly. I mean, it's pretty simple, I think, as you said. So the LLM can propose, for example, synthesis recipes, or it can propose simulations to run. And because the LLMs are pretty good at tool use, it can actually do it itself and then you get some results back. So the results from experiment could be some characterization data. Results from the simulation can be some trace of some simulation you did. Now the LLM can go through it with the context of its previous training, maybe the context of relevant papers, textbooks. But also now the results that it just got that no one else have ever seen, then now it can kind of tweak the experiment, tweak the simulation for the next step.

[Shail Khan]

Right. You said one thing in there that I guess is worth pointing out, like you're trying to automate this as much as possible. The LLM might run the experiment.

[Doge Chubuk]

Yeah, absolutely. I mean, one of the other advances that's been happening recently that I think made periodic possible is the high throughput experiments have been getting better. You know, there are many examples of this now across academia, industry, where these robots, they became quite commoditized, actually, just mixing powders or mixing liquids and then sending it to characterization. I think one thing that isn't as advanced right now, but we feel like we can do pretty soon, is automated characterization itself. So you mix powders, you put it in some characterization tool, you get the result out. What is the actual output? I think that's pretty difficult right now for AI tools, but we feel like we can improve that pretty quickly.

[Shail Khan]

I want to talk about one specific application that I know you're going after that we actually did talk about a year ago. But also I want to talk about it as a way to see whether one of the other things you described as the fundamental challenges has changed, which was, as I understood it, AI being pretty good at the next incremental discovery, but not necessarily good at the breakthrough discoveries. You said through history, usually that's done accidentally, or often it's done accidentally because it's not. You can't like, reason your way to this massive breakthrough discovery. So let's talk about superconductive materials.

[Doge Chubuk]

So.

[Shail Khan]

Right. We talked about this last time where all these companies that existed at the time that were doing AI for materials discovery were starting on things like discovering a novel MOF for carbon capture or whatever. But we said, like, the thing that would be the real breakthrough, the big thing, would be a room temperature superconductor. And you guys have, since launching, been very public about, like superconductive materials, maybe not room temperature. I'm curious for you to tell me how likely you think that is. But like high temperature superconductors is on the roadmap. So why, first of all, and then, and then, second of all, like this question of, do you think you have a path to the truly breakthrough. What would the path be to a truly breakthrough discovery as opposed to finding something that is a material that is superconductive at a ever so slightly higher temperature than the best that we've got today.

[Doge Chubuk]

Yeah. So to answer your first question, I think it's still true that it will be difficult to just reason your way into a much better superconductor. I actually would guess that there's a law out there that we haven't discovered yet that says that you can't just look at your training set that's different than what you're trying to discover and just predict it. There's been rules that we discovered from 1800s on where you connect energy to work. So thermodynamics is the first example. There's more recently Landauer's limit, which shows that you have to spend a certain amount of energy to delete information, which can be used to describe Maxwell's demon contradiction. I bet there's something similar for how hard it is to discover things that's outside of your Training domain. Okay, So I don't think that's been fixed since last time we talked. But because we have a lab internally, we can just try things and try them at large scale and often and hopefully as intelligently as possible. So even though we won't reason our way into a much better superconductor, we'll be able to push our trials in the direction that's most promising or most promising for us, given our training set at the time. I think that hasn't changed and I think there's reason to be hopeful because in the big scheme of things, it's a pretty new field. I mean, if you look at coupe rates, they were from 1985. There's been a lot of advances more recently. So yeah, we're very excited. One reason we chose superconductivity is if you find a good superconductor that's impactful immediately. Right. Like last time we talked about how long it can take to translate materials improvements into products. One nice thing about superconductors is if somebody discovers a room temperature superconductor today, even before it makes it into a product, it has huge impact. Right. First of all, it changes how we think about the universe. Second, it helps us do physics experiments that wasn't possible before. And whenever you think about a sci fi ish technology like quantum computing, fusion, superconductors come up because that's kind of what we need. It's kind of like the one of the most exciting macro scale quantum phenomena. So that's one reason we picked it, because it kind of is exciting as soon as we succeed. The other reason is it requires all sorts of improvements to get there. When we think about OpenAI and DeepMind, I remember back in 2016 people used to make fun of these institutions for prioritizing AGI so much because they were saying we're going to do AGI, but what happened is they developed so many other tools on the way to AGI, they were useful in themselves. But today they have these LLMs that you might consider AGI or something really impressive. Superconductivity is a bit like that. To discover an exciting superconductor, we probably have to develop so many capabilities on the way there that's by themselves very useful. For example, automated synthesis, automatic characterization, being able to model or predict high temperatures per conductivity because we don't have a theory for it yet. So it's kind of like a nice goal that unites people and requires a lot of other useful things to happen on the way. And it's one of those things that physicists find really exciting. So the physicists in our company are really excited by this mission, but also computer scientists find it very exciting. This is one of those things that I think both sides can really appreciate. So those are some of the reasons that you picked it.

[Latitude Media Announcer]

You heard the phrase speed to power a lot lately, but here's what it really means. AI data centers are being told that it will take years to get grid power. They can't afford to wait, so they're turning to on site power solutions like Bloom Energy. Bloom can deliver clean, ultra reliable, affordable power that's always on in as little as 90 days. Bloom's fuel cells offer data centers other important advantages. They adapt to volatile AI workloads. They have an ultra low emissions profile that usually allows for faster and simpler permitting. And they're cost effective too. That's why leaders from across the industry trust Bloom to power their data centers ready to power your AI future? Visit bloomenergy.com to learn more.

[Energy Hub Announcer]

The grid wasn't built for what's coming next. Electricity demand is surging from data centers to EVs and utilities. They need reliable, affordable solutions that don't require building expensive new infrastructure. Energy Hub helps by turning everyday devices like smart thermostats, EVs, home batteries and more into virtual power plants. These flexible energy resources respond in near real time to grid needs, balancing supply and demand. Plus they can be deployed in under a year and at 40 to 60% lower cost than traditional infrastructure. That means more reliability, lower costs, cleaner energy. You can't get much better than that. And that is why over 120 utilities across North America trust Energy Hub and to manage more than 1.8 million devices. Learn more at energyhub.com or click the link in the show Notes do you.

[Latitude Media Announcer]

Ever wonder why it takes so long to get clean energy projects up and running? Do you have permitting reform on the brain? Are you NEPA reform Curious? The new PACE study from Third Way pinpoints the non cost barriers that stand in the way of clean energy deployment and keep new solar and transmission projects in limbo. They surveyed more than 200 industry professionals to understand what's slowing down deployment and offer practical solutions on how to fix it. To read Third Way's full report and to learn more about the PACE initiative, visit thirdway.org PACE.

[Shail Khan]

I guess back to this question of how do you distinguish between the incremental innovation which to be clear, if you develop or discover superconductor at a higher temperature than anything that we've got Today that's meaningful, but it's not, you know, it's probably orders of magnitude less meaningful than if you discover a room temperature superconductor. And I presume that the, the scientific challenge is commensurately distinct between those two. And you know, the way that alums work, as I understand it, at least in part is on these reward functions. And so are you, are you setting your AI system a goal of, you know, find a room temperature superconductor and then everything flows back from there? Here are the steps and all the things we have to fix to get to room temper, or do you say improve this characteristic such that we can incrementally, you know, build our way there. In other words, are you going to find 10 super. I think of it as sort of a different thing from, but like the alternative version of this is what happens in fusion, nuclear fusion, where everybody is sort of chasing this same goal of Q is greater than one, right? Like energy, energy break even and everybody is getting incrementally closer and closer and closer and eventually NIF breaks it or somebody breaks. Is it going to look like that or is it going to look like we've discovered nothing until we discover the room temperature superconductor?

[Doge Chubuk]

Well, so as you said, I think there are many different ways of improving superconductors without getting a room temperature superconductor. So one of them could be having a significant increase in tc, but another one could be a really high critical magnetic field, which turns out might even be more important for fusion applications than TC itself. Another one can be more mundane, like some mechanical properties like a superconductor that also is ductile and we can make it into devices. So we wouldn't rule out all of these very exciting developments just for like a room temperature tc.

[Shail Khan]

But how do you set the reward function for your model? What are you optimizing it for then?

[Doge Chubuk]

I mean, I think that's an empirical question. I think one thing I should say is it's quite nice because it's hard to reward hack. You know, one of these issues with RL and training LLMs is you might worry about reward hacking. And in simulations, again, reward hacking can be a problem even in dft, but for real life experimental measurement of tc, it's much harder to reward hack, which we love. So if our reward was increasing tc, that just seems like a nicer, unhackable reward. But in terms of what specifically will get us there, you know, we're not sure yet. I mean, it's an empirical question. We can probably try all of them. I'll list the things you propose and we'll try all of them.

[Shail Khan]

I guess that gets to the other question, which is like, what, what is the human in the loop look like here? Right. And again, as you said, like, if we haven't solved the sort of AI is good at incremental innovation and not orthogonal breakthrough innovation thing, but humans are historically at least better at it. Is it like folks on your team developing a theory of something and that gets fed through the model and you get the results and you feed it back in, you see whether it's a promising category. Like, is the germ of the original idea of what to look for coming from a human or is it coming predominantly from the model and then the humans have to interpret and send it off in various directions?

[Doge Chubuk]

Yeah, I mean, that's a great question. You know, we, we're not really prioritizing full automation anyway, so if we get better results with humans doing part of it, that's great. This is also actually a question for our lab, right? Do we want to automate every single aspect of the lab? At some point you end up needing humanoids for that. And I think that's not like Liam, my co founder and I, we are trying to be very pragmatic about it. Our goal is to get the best result possible on the things we care about. And how much of the automation comes from the ML, how much of it comes from more traditional tools and how much of it gets done by humans? I think that's kind of again, an empirical question. So yeah, we're not like, I think, as you said, it does seem like today there are things that ML AI is better than humans, but one of those things is not hypothesis generation. So I mean, there are two options. We either have to improve these hypothesis generation, which is possible, or the other option is we have humans providing some of the hypotheses and then AI doing the execution.

[Shail Khan]

I guess the other question here is cost. I mean, you guys raised a $300 million seed round, so that implies on the outside that your cost structure will look similar to other companies that basically are going to use just an enormous amount of compute. And so like a lot of that cost compute in your context, I could imagine maybe that being true, but also maybe that not being true because again, you just don't have the same corpus of data. You can't build a 10 billion parameter model right now because the data isn't there to do it. And so instead that cost is going to go more toward the robotic Lab and all that kind of stuff. How should I be thinking about how much compute you'll use and where that costs come from?

[Doge Chubuk]

Yeah, so honestly, computer is very expensive and we are going to train LLMs, we are going to use GPUs to run simulations. So that does end up being a large part of the cost. Yeah, it's funny, if you asked me this question 10 years ago, I would have thought that the biggest part of the cost must come from the labs because physical is real. You're building this lab, you're buying instruments. But turns out the GPUs are so expensive and training LLMs is so expensive. So when we were thinking about how much to raise, we kind of laid it out in terms of the GPU cost, the lab cost, and this was kind of minimum number we felt like was viable. And yeah, we'll see. The GPUs have been getting more expensive recently. I guess we'll see how the market dynamics continue.

[Shail Khan]

To what extent do you end up building generalized model or models versus models designed to a specific domain, even a specific scientific domain? Right. Like you guys are doing your material discovery obviously, but physics and chemistry and these things all intertwined. But like, is the same model going to be equally capable across all domains? Is that the intent or is that just not how they supposed to be architected?

[Doge Chubuk]

That's right. And that's actually something we were very excited about. You know, one thing I've been kind of noticing is like in the past, say three, four years, I had a chance to collaborate with very, you know, world class best in their field scientists. And even when you work with them, you realize that while their expertise on a few domains is incredible, maybe best in history, there's just so much more to know in chemistry and physics that they may not know all the other aspects of it. So this is why I brought up superconductivity, because you might actually need to be really good at solid state chemistry and synthesis or difficult novel materials, just because you don't know which chemistry the superconductor is going to come from. Some of these ideas you might have may not be as stable thermodynamically, so you need to be intelligent about how to kinetically force it into that phase you want. But at the same time, in addition to solid state chemistry, you need to be incredible at condensed matter physics because there are so many different kinds of superconductivity, we don't understand most of them very well. And there's nobody in the world who knows both of those equally well, or like sufficiently well. And turns out this is true for many different aspects. If you need to use robots for high throughput synthesis. Again, there are only so many people who understand robots and how to use them for synthesis. So I think this was different in 1800s, probably. There was probably a time when a physicist could contribute and be one of the best in the world on many fields of physics, but it's definitely not true today. And this is one of the reasons, I think we are very excited about LLMs, because when you talk to them, they seem like they have a pretty good understanding of solid state chemistry and solid state physics at the same time already. And we're trying to improve them further in the physical sciences specifically, because that's where we are really interested in. Then we're hoping that they'll be good at multiples of these. Then a really exciting prospect with that is a lot of the exciting discoveries happen to lie in between fields. That's why it's sometimes easier to be interdisciplinary. And there's so many of these surface areas between these different fields. It's like, I guess science is kind of like a fractal in the way it's hierarchically organized and there's so much surface area that humans have exploited, of course, but then there's probably so much left to exploit, and we're excited about an LLM that can basically do that at a scale that humans couldn't yet.

[Shail Khan]

How good are the LLMs today? Or the best in class of what you guys have at generating synthetic data in this domain? Another way to ask this question is, you know, if you fast forward three years, you're fully up and running and you're operating, how much of the of the valuable insight you will generate do you think will come from the physical data coming out of your lab versus the synthetic data that the LLMs create on top of that?

[Doge Chubuk]

Yeah, that's a great question. I obviously don't know the answer, but it's great to brainstorm about that because on the one hand, the lab data will be kind of our additional data that other LLMs may not have. And you might think then the lab data will only be as valuable as the results in it. But on the other hand, what's interesting about scientific data is it's not just a few bits or numbers, right? Like, for example, there are certain experiments you can run where the result you get from it is just say three floating point numbers. But the implications of those could be tremendous, right? It's not just going to be like a few bytes it will actually be potentially an incredible amount of understanding just from a few experiments. And this has been how it is in human history. Right. Like, there are certain experiments that told us so much about how we understand about the universe. And the way to do this with synthetic data can of course, be you run simulations that relate to that experiment. And when you get the experimental result, that actually validates or refutes so much of the simulations you ran. And then that is a lot of information in itself. So it's a very interesting question. And I think there are some actually differences about how you think about synthetic data when it comes to an LLM that's good at science and. Exactly. I mean, this is one of the reasons I really want to work on this, because this opens up questions for LLMs and LLM training. This may be different than what the Frontier Labs are thinking about right now. If they're only thinking about math and logic and what's on the Internet, like accounting tasks, that's a bit different than if you're trying to do experimental physics, experimental chemistry. It just seems like a very exciting question to explore.

[Shail Khan]

I want to talk a little bit about how you build a business out of this. I mean, you mentioned the superconductor example, and you said, like, there's a lot of value in this long before this novel superconductor falls, goes into a product, but ultimately kind of has to go to a product of some sort for you guys. And I think we talked about this a little bit last time, too. There's this question, okay, so if your job, your core job, periodic, is to discover new things that are going to be valuable in the world, say you do it, to my mind, there's sort of a binary decision you have to make at that point. Do you try to sell the discovery, license, the technology, license, the IP to somebody else who's going to go produce it and turn it into a product, or do you produce it? Do you sell the product? Do you sell the tech, the discovery, or do you sell the product? Do you have a prior on which direction you want to go here?

[Doge Chubuk]

Great question. And, you know, I think the two options can be. Be correct depending on the context, depending on the timelines. But honestly, it also depends on where we are in the company. So at the very beginning, you can imagine our LLMs will be very impactful for other companies doing physical R and D already. Yeah, exactly. A lot of interest in being able to use these LLMs. Sometimes the data restrictions don't allow it because you don't necessarily want to put your data on an LLM on the web sometimes. The other issue is you haven't trained LLM on your data, so it's actually not as good as it could have been. That kind of improvement could be really impactful because we've seen how impactful LLMs can be in other fields where they have access to the data. So there's a lot of, I think, headroom for impact there. But in the longer run you can also imagine a case where we as a field get really good at designing materials intentionally. That hasn't been the case. But if you look at drug design, there was a time when designing drugs wasn't very profitable. And I think people will look at it and say this is not a good business. But what happened with Genentech is the field got so good at designing drugs that at some point it became very valuable itself. The machine learning field has been making huge improvements in material science that was kind of hard to predict. So it'll be interesting to see how far that goes and whether materials discovery by itself becomes a very exciting business similar to drug discovery. But for us, we already see this big need and a big potential for impact by providing these LLMs to do.

[Shail Khan]

Physical R and D. Yeah, almost like this is going to be the wrong analogy, but it's partially right. Like, almost like an aws, like you're going to have the infrastructure. In this case, the infrastructure is your custom designed LLM that is smart about physics and chemistry and all these domains and also your physical lab and them being interconnected with each other. And, and so you have all this infrastructure and scale in that infrastructure that you can use to go convince whatever large company that's doing R and D that they should just be outsourcing it to you rather than building, rebuilding the whole same thing in house, which is not exactly what the cloud providers are, but you know, there's enough of a relationship there so that that feels right. But it is, I suspect, yeah, I guess is what you're saying. This is, I suspect, a smaller ultimate opportunity than the. You proactively discover a bunch of novel materials that change the world and then however you monetize them, you prove you're able to do so repeatedly and then you're Genentech and you know, it's a whole different category.

[Doge Chubuk]

Yeah, and you know, one other thing you see is people are so excited about this. You know, they want to see LLMs not just conquer the digital world, but also, you know, really impact the physical world and impact the atoms, basically. So I feel like this has to be done. And the team has been very excited. It's really amazing. We are hosting weekly seminars where the physicists will teach the computer scientists about the physics and the computer scientists will teach the physicists about LLMs. And of course there are a lot of people in between. It's actually again, a fractal. So yeah, I think there's been a lot of excitement about seeing if these technologies can be used not just for the digital world, but also for constructing the atoms around us.

[Shail Khan]

I guess maybe the last question you said at the beginning, the thing that changed between a year ago and now in part was advancements in the big LLMs, right? The O1 model and so on. Is there a next? Like what could change in what could OpenAI release in a year or two years from now? That would be a big leapfrog for you. Are you branching off now from what the big LLMs are going to do and you know, everything that all advancements are going to come from periodic, or is there something else that they could offer that is a step, function change in your capability to discover new materials or whatever?

[Doge Chubuk]

Yeah, great question. I think we actually basically rise with the tide, right? Like as alums get better, there's so many advantages of that to other applications. For example, the LLMs are getting very good at coding and that's not surprising because programming is a kind of closed environment. You can just simulate in your computer and get valuable feedback and then quickly improve. But as computers get better at coding, that's huge for science because then you can run simulations more efficiently. The simulations themselves can improve similarly with tool use experiments. So I think as LLMs improve in general, that's going to help a lot with science applications. There are maybe longer term things that can happen. One of them could be things like hypothesis generation or more auto domain generalization. But then a question there is will that come from status quo, like how LLMs are being trained now, or will it come from actually labs that try to improve scientific reasoning for these outlines? Because then maybe hypothesis generation emerges naturally or auto domain generalization emerges naturally, because that's what you're kind of trying to get at with your reward. So I think that'll be a very exciting question to see maybe next time we chat.

[Shail Khan]

Love it. All right, thank you so much for taking some time. Again, congrats on periodic. Super excited to see what you guys discover. And, and for, for when you're room temperature, superconductor is shooting electricity all around the world around me.

[Doge Chubuk]

Okay, well, thanks a lot. It was great chat.

[Shail Khan]

Doge Chubuk is the co founder of Periodic Labs and a former researcher at Google DeepMind. This show is a production of Latitude Media. You can head over to latitudemedia.com for links to today's topics. Latitude is supported by Prelude Ventures. This episode was produced by Daniel Waldorf. Mixing and theme song by Sean Marquan. Stephen Lacy is our Executive editor. I'm Shayl Khan and this is Catalyst.