The AI revolution has ushered in an era of smart machines, and smart society. Powered by ever advancing software and circuits, and abundant energy.

But there’s a physical constraint to the exponential growth of intelligence. AI models can be quantized and optimized, and processors can be made leaner, meaner (more efficient) and faster. But you still need more chips, and lots of them.

Hundreds of millions. Billions. For the cars, robots, satellites, drones, phones, and AI factories. And every other smart device you can think of. They all want the latest ‘super chip’.

Which is why Elon Musk’s crazy Terafab idea is paying off so handsomely. He recognized that the only way to meet the demand and bring down the cost is massive scale. So they reinvented the chip factory. From first principles. And the cheaper the chips, the bigger the demand – it’s a virtuous vortex of value.

Not that China hasn’t also entered the fray with a two-pronged approach. Their pilot 2D semiconductor nanofab is scaling up radical new technology, while their mass market silicon wafer plants are churning out chips by the millions.

With AI powered designs and optimized automated factories, there’s no advantage in superior knowledge or patents. Everyone’s chip is the best chip. It’s all down to execution, and who has the skill to produce consistently high quality with massive quantity.

As the price of intelligent chips falls, it’s a two-horse race to the bottom: Terafab versus China. No one else is even in the running.