The world that awaits us

When futurists talk about transformative technologies, nuclear fusion, hydrogen fuel cells, quantum computing, and AI are always on the agenda. The first two promise clean, limitless energy, while the third and fourth taken together could solve many currently intractable problems.

AI has seen extraordinary progress and is already transformative technology. More and more tasks are being taken over by algorithms that outperform humans. But the other three technologies may be years away from commercial viability. Big technical problems need to be solved before nuclear fusion, fuels cells at scale, and quantum computing under normal conditions, get off the ground.

Untold billions have been invested in research over decades on these. But some people argue the engineering challenges are so formidable that they may be impossible to solve at scale. Fusion research started in the 1950s, quantum computing in the 1970s, and the first hydrogen fuel cells date back to the 19th century. The length of the timelines indicates the technical difficulties.

Stable, efficient fuel cells exist and run cars, trucks, and trains. It is hydrogen storage and transportation that raise huge question marks. Hydrogen is the lightest element and it is very hard to store in quantity because it’s so light. 

Liquefaction of hydrogen (to make it easier to store) requires chilling the gas down to near absolute zero (minus 273 degree Celsius), and that means a lot of smart engineering and lots of energy expenditure. Storage tanks that can safely contain this potentially explosive fuel are also expensive and hard to design.

While hydrogen may need cooling, nuclear fusion requires heating. The fuel (usually plasma — that is, an electrically charged gas) must be heated to insane levels of 20 million Celsius or multiples of that. That requires a lot of energy. Moreover, nothing solid survives at those temperatures. So the containment of the fuel is done by placing it within very strong magnetic fields. Alternatively, every reaction involves destroying the chamber where it occurs.

Both methods have been demonstrated successfully with temperatures reaching 100 million degrees, and impressive fusion reactions have occurred. But the energy required to generate and control a fusion reaction far exceeds the energy generated by the fusion reaction. So this isn’t much use in practical terms. In addition, the equipment required is, to say the least, very expensive. Incidentally AI can probably reduce the costs of running fusion experiments, because AI can control magnetic fluctuations and laser arrays more efficiently. This too has been demonstrated.

Quantum computing demands ultra-cool, ultra-stable conditions. Quantum computers usually operate at near absolute zero and some are suspended carefully in mid-air to ensure vibrations from passing trucks don’t cause errors. However, quantum machines also need to be connected to “normal machines” to do meaningful work, and normal equipment (including chips which handle bits, not qubits) doesn’t work at minus 272 degree Celsius. In addition, software to deal with qubits with their quantum peculiarities is very different from that used in conventional programming. But this is a problem less difficult to solve.

There are several promising developments in quantum computing. Researchers are trying to develop qubits that work at near room temperatures with some success. Other researchers are trying to develop normal equipment that works in cold conditions, again with some success. Some of this R&D has already reached demonstration, or proof of concept stage.

It’s likely the next five years will see quantum machines being built to run in less exotic conditions. Once that occurs, making the finances work will be the next big challenge. As and when AI can be run on quantum computers, the combination might come closer to making fusion reactors financially viable, or help to find cheap storage solutions for hydrogen.

Ideally circa 2050, our houses will be powered by fusion reactors generating clean, cheap, limitless energy, and everyone will carry quantum mobiles running AI and travel in vehicles powered by fuel cells. It’s not impossible and R&D in these areas has acquired extra urgency due to the imperatives of Climate Change. But it will require big conceptual breakthroughs. Smart people have spent decades without finding those. Is there a Plan B?