This paper argues that fusion is scaling the way the semiconductor industry did: build machines, sell them to paying customers, and use the revenue and operating experience to build better machines. No single breakthrough drove the chip industry’s cost curve, and the paper makes the case that fusion’s progress comes from the same steady, compounding engineering improvements, comparing the two industries’ cost curves side by side.
It also points to a key difference. Semiconductors had to wait for each new market to be invented, while fusion’s markets, including industrial testing, medical isotopes, spent fuel recycling and eventually grid-scale energy, already exist. The paper covers the main risks, such as materials durability, tritium handling and capital intensity, and explains why staged commercialization is the primary way to manage them.