CNTs, Assembly Joints and Solving Maintenance: A conversation with Baratunde Cola, CEO of Carbice

This is not a story about thermal interface materials. It's a broader, richer story about maintenance costs, credit card fraud, and why the entire global economy is paying a hidden tax on bad joints.

Baratunde Cola has been working with carbon nanotubes for 25 years. PhD at Purdue. Award-winning professor at Georgia Tech. Vanderbilt fullback. Winner of the 2017 Alan T. Waterman Award, given to the top scientist or engineer under 35 in the United States. Somewhere along the way he invented a catalyst that grows vertically aligned carbon nanotubes directly bonded to plain aluminium foil, founded Carbice in 2011, and built the largest vertically aligned CNT production facility in the world: 23,000 square feet, AS9100D and ISO 9000 certified, in Atlanta, Georgia.

Bara doesn't pitch Carbice as a thermal interface company. He pitches it as an assembly joint platform. And once you hear that framing, it's hard to unhear it.

What I Learned

1. It's not about the material. It's about the joint.

I walked in thinking about thermal conductivity. Bara reframed the entire conversation around mechanical assembly joints. Every adhesive, every washer, every bolt connection in every industrial product fails in exactly two ways: compression set and shear delamination. Carbice eliminates both. Thermal performance is almost a side effect. The real product is a joint that never fails. Once you hear that framing, the market size explodes from "thermal interface pads" to "every assembly joint in every product ever built." That's a different business entirely.

2. Maintenance is a hidden GDP-level tax. Like fraud.

Bara threw out some staggering numbers. $900 billion in electronics contract manufacturing last year, with perhaps 30% tied to rework and assembly issues. A $50 billion insurance industry exists purely to monitor whether switchgear bolts are coming loose. Grid outages trace back to distribution joints. Data centre fires trace back to joints. This isn't a niche problem. It's a systemic tax baked into the cost structure of modern industry, exactly like credit card fraud is baked into transaction fees. You don't see it, but you're paying for it.

3. The moat is underestimation.

Every major materials company has already tried and failed to manufacture aligned carbon nanotubes at scale. Their pride has been crushed. They're not coming back. That's a moat you can't replicate with money. Meanwhile, Bara spent 15 years building manufacturing nobody believed was possible to scale and make money, deliberately keeping quiet about the fact that Carbice even uses carbon nanotubes until the product was already shipping.

I. Why It Took 25 Years

Lawrence: So you've been doing this for 25 years. Are we near making an industrial impact?

Bara: There are a lot of myths about carbon nanotubes. I think cost has always been a myth. And like with all technology, getting new stuff to happen has more to do with people and societies than science. The science of carbon nanotubes is actually not that difficult. What excited me, and the reason I've stayed in this for so long, is that the science and the scalability were clearly solvable problems.

But the way innovation happens... a lot of it came out of universities. And there's not a lot of patience for things that aren't aligned with the incentive system. The incentive system is to get papers out, to excite people. Not to get into the details of what's practical in manufacturing.

With carbon nanotube thermal interfaces, everybody thought that the thermal conductivity of the nanotube was the key driving parameter. We had all these cartoon PowerPoints of vertical rods sticking up. Part of my PhD dissertation was to say that's actually not true. The nanotubes are there to make contact, and they actually look like curly fries. They never make a point contact. People still ask me about phonon transport. None of that actually matters. What matters is that they bend over like a toothbrush on your tooth and make a lot of contact.

Lawrence: As a dumb VC, AI for materials science is a huge thing now. If you were to discover carbon nanotubes today, people would think this is magic. There'd be a new startup and it would raise $100 million. But it's been 25 years. Is this a heuristic for all materials?

Bara: It's the same for everything. Most things are good enough. "Good enough" delays all innovation. The world is filled with redundancy. In the chip space, I tell people all the time: there's a lot of talk about cooling, but nobody buys cooling. Every system has cooling margin. When they talk about cooling, what they're really talking about is reliability in application at scale.

II. The Assembly Joint Platform

Lawrence: You have a product for thermal management. But you don't call it that.

Bara: We actually call it assembly joints. Our product is a thermal interface, but it's actually a mechanical interface. The breakage between cold plate and heatsink, which is pervasive across the industry, voiding and all that... it never happens with our product. Because it never happens, that's a game changer.

Lawrence: How would the average person describe this?

Bara: We make assembly joint solutions. A mechanical assembly joint is a washer and a bolt. It's the adhesive that holds your shoe on. It's paint drying on a wall. All of these things fail fundamentally in only two ways: either they lose thickness through compression set, or they have shear delamination — they peel off. The key platform at Carbice is that we eliminated the only two failure modes of an assembly joint.

When you take two surfaces and fasten them together, there's a stress load. Carbon nanotubes are orientated in our design like a bunch of squiggly trees in a forest. They absorb that stress like springs. Nano springs. They extend and compress — no other joint material can do this. No matter how much vibration, thermal expansion, temperature spiking... these springs forever absorb and relax.

Lawrence: Like a Tempur-Pedic mattress.

Bara: Exactly. And typically when you squeeze a sponge, it gets wider somewhere else. That's the Poisson's ratio. That causes stress. Our material presses down without doing that.

Lawrence: When I think about nanomaterials, there's the bulk commodity end and then the premium end. Where does Carbice sit?

Bara: This is what makes us special. We sit in between. We're the only ones in between. Instead of random bulk powder or lithography premium, we've aligned the nanotubes. So they're semi-premium. They're directional. And because of that directionality, we can take advantage of things like capillary pumping — where the free tip ends of our curly fries gets sucked into the interface through a capillary force. That wouldn't work with bulk powder nanotubes.

Lawrence: Tell me about Dow. When I first saw the partnership, I thought, surely Dow could do this themselves?

Bara: Our biggest moat is underestimation. I learned this playing college football.

Here's what most people don't know: every single material company in the world has already tried to do what we're doing in cooling directly or supporting university labs. And they all considered their efforts a failure, which means someone took the fall and organizational pride has been crushed. I was on the scientific advisory board for some of these companies. They'd fly me around as the world expert. I realised they'd failed, which means pride has been crushed. They're never going to invest in it again. That's the moat.

So I invested in it. We solved the manufacturing. This first piece of Carbice was made in February 2011. I did about 200 recipes because the secret is catalysis. I invented a catalyst to grow these things on plain aluminium foil.

Five years later, Dow comes to me and says, let's partner. I said no. Everyone came, I said no. Because I know they'll go inside their organisation with a "not invented here" mentality.

The partnership works because they're a liquids company, not a pads company. They supply us with polymers and we can help add to their product offerings. Our product can be tailored with an endless number of features by coating the carbon nanotubes. Dow has some of the best polymer chemists in the world and they made two new polymers for us that have produced new features in two new products that customers like. One is the first semi-structural adhesive thermal and electrical pad and it is already shipping to a large semiconductor company.

Lawrence: Walk me through the platform. What can you build on it?

Bara: We eliminate compression set, shear delamination, and stress concentrations.

So you've got thermal interfaces, electrical contacts, gaskets, bolt washers, EMI shields, anti-condensation surfaces, and radiators. For electrical contacts, this is the only material going through a million insertions without losing performance. For gaskets: my father worked in the chemical industry. Six to twelve times a year, you change graphite gaskets on heat exchangers because they leak. Put Carbice in, it'll never leak. For bolt washers: in data centres, with AI workloads, things vibrate so much that bolts come loose. Put a Carbice washer on, done.

Lawrence: If you could somehow measure all maintenance costs globally...

Bara: I've done it. Nobody believes my number because I'm the only one who has. Last year, the electronics contract manufacturing industry spent something like $900 billion. Maybe 30% of that is tied to rework and assembly. Then there's a $50 billion insurance industry just for bus and switchgear joints. The insurance companies require optical pyrometer monitoring because they know the bolts come loose, electrical resistance goes up, and it causes fire.

When the grid goes out, it goes out because of distribution. And distribution goes out because of the joints.

Lawrence: It reminds me of fraud. The costs baked into credit card transactions. If you removed fraud, everyone would pay less. Maintenance is the same hidden tax.

Bara: It's like taking out a country's worth of economic waste. The modern world needs a different way. Just like with fraud, if you find a different way, you save countries of income.

And here's where the economics get interesting. The bill of materials for assembly materials is the lowest thing in the bill of materials. So even if you pay 10x what you currently pay from a material standpoint, you still save money. One of the aerospace primes did a TCO analysis and determined it was over a 60% savings on the cost to build a satellite using us. 60% on the entire cost to build a satellite. Because they were gluing, reworking, throwing away 50% of the boxes.

III. The Data Centre Moment

Lawrence: A hyperscaler comes to you wanting to outfit a 100 MW data centre campus. Can you do it?

Bara: When we fill our current real estate with more equipment, we'll do 20 football fields a year. European football. Soccer. So if you've got 20 football fields of GPUs, we've got you covered. The facility today, with what's installed, can do $30 million of revenue per year. When we get to 20 football fields a year, we'll be anywhere between $500 million and $750 million. Out of this same real estate. No more real estate.

Lawrence: What about yield?

Bara: It's the highest yield you can have for making pads and interface materials. Because it's made from gas, solid structure built from the bottom up one atom at a time. Yield is determined by temperature control. Every time we go bigger, yield gets higher. We're on our fifth generation reactors. Five sigma quality with plus or minus one micron across substrates.

We have seven of the top ten global aerospace and defence primes as customers. They always audit your facility. We usually have the fastest audits because people are surprised at how good it is.

Lawrence: Do you see a world in 20 years where nanotubes are in so many products, like plastics, where everybody knows the name but nobody thinks about which flavour they're using?

Bara: Yeah. We're elevating it. We've been shipping on Intel and AMD processors with CyberPower since last summer. We were featured at their booth at CES. I have a vision. When I was a professor, I was an award-winning public speaker. I used to talk to tens of thousands of students. I'm the ex-football guy who's going to talk about nano. I want to make it famous.

Lawrence: Last question. Financing.

Bara: We're going out for more money. It's not going to be a huge amount because we'll be break-even pretty soon and can generate a lot of free cash in our operations. But this is the time of a generation. We need to scale fast. And if you want the US Foxconn play, that requires money too.

Thanks for reading the CNT Dispatch. This was our first long-form interview. More to come. If you have questions for Bara or want to dig deeper into any of this, reply to this email. And if you're a data centre builder who hasn't visited Carbice's Atlanta facility yet... you should probably fix that.