Quantum Industry Urged to Scale on Compute-per-Watt, Not Qubit Count

The article says the tech industry was not ready for what came after ChatGPT. Power grids, cooling systems and data centers are being rebuilt because none were designed for the demand that arrived. UN researchers expect AI to double the power and water that data centers use by 2030, according to the analysis. Had the industry known that demand was coming, it would have designed for it, and quantum computing is early enough to do that.

AI hyperscalers care about how much useful computation they get for each dollar invested, the article says. That total cost of ownership is dominated by depreciation and power usage. In the current infrastructure buildout, hyperscalers have sufficient access to capital to fund capital expenditure but insufficient access to power. The constraint is acute enough that data centers in space are becoming an economic possibility if growth continues at this pace. Even if AI matures and growth slows, total cost of ownership will dominate, translating into capital requirements and making compute-per-watt one of the most important metrics.

Quantum computers, however, are still sold on qubit count, and qubit count alone says nothing about economic returns on a system, according to the article. What will matter at scale is how much compute-per-watt the end user gets, and that is not driven by qubit counts alone.

The article uses superconducting qubit processors as an example. They have been stuck at around 100 qubits for almost a decade because, on today's chips, more than 90% of the surface is taken up by wiring that controls the qubits and reads their answers, rather than by the qubits themselves. The usual workaround is to network many small processors together. Networking is lossy and gives sparse connections, which is bad in classical chips and exponentially bad in quantum chips. The work of holding the system together grows faster than the qubits it adds, and more power goes into running the machine than into computing with it. That gives higher qubit counts, but those are not equivalent to systems built with less networking because it costs compute-per-watt.

The article argues the quantum industry should care more about experience curves. An experience curve means that every time the total number made doubles, the cost of each one falls by a fixed amount. It drove down the price of solar panels and batteries, and quantum will not be an exception, but only if the industry builds at volumes that let the curve work. Optimistic quantum roadmaps never talk about price reductions, the article says, even though it is one of the most important problems the industry needs to solve. At today's price per qubit, a million-qubit machine would cost somewhere between $100 billion and $1 trillion. The cost per qubit has to fall at least a hundredfold in the coming years for quantum computers to be economically viable, which the article calls doable but not solved by more laboratory proof-of-concepts.

The transistor is the clearest case in computing, according to the article. A single transistor once cost around a dollar. Today a chip carries billions of them, and each one costs a fraction of a cent. That fall came from decades of making more of them, driving down the cost year after year. Costs fall the way they fell in classical computing: volume manufacturing of standardized parts combined with compounding, but not stepwise, technical progress. This requires an open architecture in which specialist companies each build one layer of the machine. The article calls this Quantum Open Architecture: processors from one company, cryogenics from another, control electronics from a third. It is what will drive specialization and in turn the required economies of scale. With the field transitioning from science to engineering, there are plenty of cool one-off demos at startups today, but very few companies actually building the supply chain.

The article says quantum computers will change drug discovery, material development and machine learning. It also ties quantum back to AI: the same workloads straining the grid today are the kind of heavy computation a quantum machine could one day take on at a fraction of the power, once the economics are there. Getting there as fast as possible will create a massive amount of value, the article says. The technology is progressing, but the industry has not focused on the economic metrics that matter. If it does not do so soon, the article warns, it risks building an industry that makes expensive demos rather than actually useful computers. Qubit counts alone do not matter, and building supply chains takes years, if not decades.