His Supervisor Can't Afford a House in London Either
Once, Heng asked the supervisor sitting across from him: “Can you buy a house in the centre of London?” “No, I can’t.” His supervisor was the deputy head of the department. Heng said that in that moment, something clicked.
Once, for an experiment, he needed a piece of steel. In the UK, three months went by and he couldn’t find one. He posted a request in China, and quotes from suppliers poured in.
“Explain clearly what you want to do. Convince the people around you first, then convince the investors.”
Once, Heng asked the supervisor sitting across from him:
“Can you buy a house in the centre of London?”
“No, I can’t.”
His supervisor was the deputy head of the department.
Heng said that in that moment, something clicked.
Heng works on a new kind of power generation technology. Put simply: fuel goes in, electricity comes out.
The competitor is the gas turbine. There, chemical energy is first burned into heat, heat is converted into mechanical energy, and mechanical energy is converted into electricity, at roughly 30% efficiency. A fuel cell relies on an electrochemical reaction and jumps from chemical energy to electricity in a single step, reaching efficiencies of up to 50%. The spent fuel becomes carbon dioxide and water. The technology originally came out of the Apollo moon landings.
Is it cheap? Not yet. A gas turbine costs about three thousand dollars per kilowatt; an American company working on similar technology charges five to twelve thousand.
Heng believes cost gets driven down by scale.
In China, more than a dozen companies and teams are already working in this direction, and some have built demonstration systems in the tens to hundreds of kilowatts. But very few have truly solved lifetime, cost, batch consistency, and large-scale manufacturing all at once. The barrier is high: materials, cells, interconnects; every step needs someone who has actually done it.
Fortunately, these happen to be exactly what Heng spent his PhD and postdoc on. He did his PhD at the University of Surrey, the university that built the world’s first modern reprogrammable small satellite, and it was there that he put this entire process together from scratch.
Heng had originally planned on an academic career, until he saw how the academic game is played here: what gets judged is mainly not your papers, but whether you can bring in money. And bringing in money depends mainly on networking. The UK pie is only so big, competition is fierce, and not many chances would come his way.
One of his classmates was also doing a postdoc. During a break between experiments, the two of them were chatting.
Heng said, “I don’t think there’s much hope in doing this.”
His classmate replied, “Right. That’s why I started a company.”
Heng is preparing to start one too.
A startup in energy manufacturing has to be done back in China.
Once, for an experiment, he needed a piece of steel. In the UK, three months went by and he couldn’t find one. He posted a request in China, and quotes from suppliers poured in. In the end his supervisor required him to buy it in the UK, and he had to lower his specifications.
A lab at Oxford built a device the size of a small room that can produce aviation fuel. The technology is excellent. The equipment cost 100 million RMB, and it was bought from China.
Then there was a PhD student working on thermal catalysts. He had published quite a few papers on his catalyst, including some in top-tier journals by the Chinese Academy of Sciences ranking. When he talked with the chemical giant BASF, he discovered that what they needed was a catalyst that could run under high pressure, and what he had been studying had never been tested under high pressure at all.
Heng says the PhD training here is still quite good: students build their equipment from zero, learn the principles, learn how to do research. But because of funding constraints, the lack of industry-oriented equipment and the support that goes with it is a real weakness.
I asked him: does AI help with your experiments now?
He said, “Having AI design experiments on its own right now is most likely unreliable. If you have no relevant foundation and follow its design completely, and something goes wrong, things explode.”
He added, “Having it extract data and experimental protocols from a paper, that works. Having it design directly, that most likely doesn’t.”
I asked him how he built his team for the startup.
He said, “Explain clearly what you want to do. Convince the people around you first, then convince the investors.”
Now the team is in place, backed by the university, and he is heading back to China shortly.
One more thing. The most elegant part of this technology is that the whole process can run in reverse. When there is surplus electricity, apply power, and carbon dioxide and water turn back into syngas, then into liquid fuel, which can be canned, stored, and shipped anywhere it is needed.
Compare that with lithium batteries for energy storage: expensive, and when a large bank of lithium batteries goes wrong, it is a fire that cannot be put out. A few years ago an energy storage station in California caught fire and burned for a full week.