Thintronics®

Here is a fact that sounds made up but isn't: one of the most important materials in the modern computer is manufactured, overwhelmingly, by a Japanese company that is far more famous for inventing MSG. The company is Ajinomoto. The material is dielectric build-up film - the insulator that goes between the copper layers of a chip package - and Ajinomoto has held something like 90% of that market for roughly 30 years. If you are reading this on a laptop, there is a good chance the insulator inside it is theirs.

This is the kind of fact that, once you know it, is hard to un-know. And it is roughly the fact that Stefan Pastine, a chemist, built a company around. The company is called Thintronics, it was founded in 2019, it is based in the San Francisco Bay Area, and its entire premise is that the insulator - the layer nobody thinks about, the layer that is definitionally supposed to do nothing - is actually where a lot of the performance of next-generation computing is going to be won or lost.

To appreciate why, you have to accept a slightly counterintuitive idea. Everyone is spending enormous sums making transistors smaller and chips faster. But once you have a lot of very fast chips, the bottleneck moves. Now the question is how cleanly the signal gets from one chip to another - across the package, across the board, across the rack in an AI data center. That is a materials question. Specifically it is a question about the insulator, because the electrical properties of the stuff surrounding the copper determine how much of your signal survives the trip and how much energy you burn getting it there.

Thintronics' pitch is that the incumbent films were designed for an earlier era, and that you cannot patch your way to the terabit age by tweaking a 30-year-old formulation. So instead of buying an existing material and adjusting it, the company designs new molecules and builds the insulator from the bottom up - what it calls a molecular-design approach, combining computational, molecular, electrical, and process engineering. The claim is that this produces materials that are, in a technical sense, unlike anything currently on the market: lower dielectric constant, lower loss, tunable between flexible and rigid, and roughly 50% thinner than conventional options while keeping the signal clean.

What it is actually for

The concrete applications are the least glamorous and most consequential parts of the digital economy. AI data centers, where the interconnect speed target has climbed to 224 gigabits per second and beyond, and legacy fiberglass-based insulators struggle to keep up. High-speed networking - the switches and integrators that move data around. And RF and millimeter-wave systems, the stuff underneath 5G and eventually 6G wireless, where the material properties of the substrate directly shape how well the antenna and the radio work.

The other thing worth knowing about Thintronics is where its founder came from. Pastine previously built and sold a company focused on recycling hard plastics before turning, in 2019, to chip materials - which is a slightly unusual résumé for someone trying to break into the semiconductor supply chain, and which shows up in the company's interest in recyclable thermoset chemistry. It is a materials-science story that happens to also be a sustainability story and a reshoring story, because the pitch is explicitly a "made in the USA" alternative in a market currently dominated by a single overseas supplier.

Now, the honest caveat. Breaking a 30-year, 90%-share incumbent in a materials market is extraordinarily hard. Materials businesses are slow: qualification cycles are long, customers are conservative, and "our insulator is better on paper" is a long way from "your insulator is now designed into a shipping product." Thintronics is a roughly 29-person company that has raised on the order of $23 million. Ajinomoto is a multibillion-dollar giant that has been iterating on this exact product since before most AI startups existed. So this is not a done deal. It is a bet.

But it is an interesting bet, and it is the kind investors seem to like right now. The $23 million Series A was led by Maverick Capital and Translink Capital, with Tallwood Venture Capital participating and a grant from the US National Science Foundation. Then the company added a Series A extension that brought in two strategically pointed investors: M Ventures, the corporate venture arm of Merck KGaA, and TGVP, the US venture arm of Toppan Holdings - a Japanese printing and electronics giant that knows an enormous amount about semiconductor packaging. When a food-and-materials conglomerate and a printing company both write you a check, it is usually because they understand your problem better than a generalist would.

The strategic logic of attacking a market like this is straightforward, even if the execution is brutal. A near-monopoly means one supplier, one roadmap, one point of failure - and a lot of large customers who would, all else equal, prefer a second source. It also means the incumbent's product roadmap is the industry's roadmap, so if you believe the next generation of computing needs materials the incumbent hasn't prioritized, there is an opening. Thintronics is aiming at exactly that gap: the very fast, very thin, very low-loss end of the market that AI and next-gen wireless are dragging into existence.