Jan 23, 2024 · 31m · green-blueprint
How AI is rapidly advancing new materials for clean energy
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The Carbon Copy, host Stephen Lacey interviews Citrine Informatics CEO Greg Mulholland to examine how physics-informed artificial intelligence is revolutionizing materials science to overcome critical bottlenecks in clean energy and industrial decarbonization.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The hosts as informed peer 4.3 Guest teaching 4.3 Guest disagreement 0.1 The hosts pushing back 0.0
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 28:14 Mulholland dismisses LK-99 hype
Mulholland rejects sensationalized claims surrounding room-temperature superconductors, pointing out that LK-99 was scientifically flawed and that true breakthroughs will require fundamentally new physics.
Hardest push from the hosts ▶ 28:11 Lacey presses on room-temperature superconductor timelineLacey puts Mulholland on the spot with a direct challenge on whether Citrine or AI can deliver a room-temperature superconductor.
Biggest teaching moment ▶ 13:00 Contrasting scientific data with LLM web scrapingMulholland educates listeners on the core architectural difference between LLMs scraping text for free and scientific AI needing to extract value from costly negative results excluded from published papers.
The host holds their own ▶ 25:23 Lacey cites Microsoft and Google DeepMind materials initiativesLacey demonstrates strong industry command by citing Microsoft's solid-state electrolyte discovery and Google DeepMind's GNoME materials stability exploration to frame the competitive landscape.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Origin Story: From Electrical Engineering to Materials Informatics | 4 | 3 | 0 | 0 | Lacey introduces Greg Mulholland's background with detailed technical narrative (nanowires, gallium nitride). Mulholland elaborates on his transition from electrical engineering to materials informatics in business school. | |
| Understanding the Massive Scale of the Materials Industry | 3 | 4 | 0 | 0 | Lacey asks foundational framing questions about the scale of the materials industry and discovery cadence. Mulholland provides educational context, comparing market sizes and distinguishing between incremental alloy refinements and fundamental Nobel-worthy discoveries. | |
| Materials as the Bedrock of the Clean Energy Transition | 4 | 4 | 0 | 0 | Lacey prompts specific clean energy examples. Mulholland details the critical material dependencies in solar cells, wind turbine lightweighting, and battery chemistry supply chains. | |
| Traditional Lab Constraints and the High Cost of Experimentation | 4 | 5 | 0 | 0 | Lacey asks about the practical constraints and economics inside research labs. Mulholland explains the high cost per experiment, trial-and-error bottlenecks, and the reliance on human intuition. | |
| The Contrast Between Large Language Models and Scientific AI | 5 | 5 | 0 | 0 | Lacey draws parallels between mainstream generative AI applications and specialized materials discovery. Mulholland contrasts internet-scraping LLMs with scientific AI, highlighting how negative experimental results are trapped in corporate silos. | |
| Citrine's Approach: Blending Physics Knowledge with Sparse Data | 4 | 5 | 0 | 0 | Lacey asks for a walkthrough of Citrine's workflow. Mulholland explains how domain physics and expert prior knowledge are encoded before small-volume experimental datasets are ingested. | |
| Real-World Impact: Batteries, Fuel Cells, and Toxicity Reduction | 4 | 4 | 0 | 0 | Lacey asks for tangible platform breakthroughs in sustainable materials. Mulholland highlights fuel cell polarizability discovery, fire-retardant battery gels, and PFAS toxicity reduction. | |
| Industry Horizons: Big Tech Initiatives and Novel Physics Frontiers | 6 | 4 | 1 | 0 | Lacey demonstrates domain familiarity by citing Big Tech benchmarks including Microsoft's solid-state electrolyte and Google DeepMind's GNoME database, then playfully probes on room-temperature superconductors. Mulholland clarifies why fundamental stability datasets complement rather than compete with applied informatics. |
Statements from this episode (11)
Mulholland: Citrine founders used AI to invent new thermoelectric materials
“We actually invented a new class of thermoelectric materials using sort of some combination of machine learning, artificial intelligence, depending on how you define these things.”
Mulholland: China pushes LFP batteries while US relies on high-risk cobalt chemistries
“In China, there's a big push for lithium iron phosphate. In the U.S., we have, you know, sort of standard lithium ion batteries that have a lot of nickel, manganese, and cobalt in the materials, and those things are hard to source, and rising in cost, and some…”
Mulholland: Truly new materials typically take 5 to 10 years to develop
“Typically new materials, like truly new materials are developed in sort of the five to 10 year time, time scale. And then they take a long time to roll out because you have to prove that they work and work for long periods of time. And then individual refineme…”
Mulholland: High lab experiment costs prevent large data volumes in materials
“One of the things that we don't have in the materials industry is huge volumes of data because everything costs money, right? You know, there is no free experiment.”
Mulholland: Scientific AI gains most value from unpublished negative experiment results
“Unfortunately, from a scientific standpoint, the most valuable thing to AI is the two examples. But you know that those scientists did not just do two experiments. You know that those scientists did months and months and months of work and refinement and faile…”
Mulholland: Citrine projects often start with 25 to 100 data points
“So very often, a citrine materials development project starts with 25 or 50, maybe a hundred data points.”
Battery maker used Citrine's AI to develop fire-quenching gel in weeks
“They brought their data in, and it was just a couple weeks later, they had a perfectly functioning quenching gel in their product that they would not have been able to find otherwise. It would have been a many, many months long overall project.”
Optimal battery performance requires co-optimizing cathode, anode, and electrolyte
“There is no best cathode. There is no best anode. There is no best electrolyte. You need to build the system that they are all compatible with each other to make the best battery.”
Mulholland: Materials industry lags pharma by 5-7 years in tech adoption
“It also means that the chemicals and materials industry tends to be maybe five or seven years behind the pharmaceutical industry and its adoption of new technologies.”
Mulholland: Room-temperature superconductors will require new physics
“My prediction is that it will require at least a variant on known physics, if not entirely new physics, to make happen, just because that's sort of the nature of those massive breakthroughs, especially, I mean, room temperature superconductor is, A guaranteed …”