Science fiction novelist Arthur C. Clarke postulated in 1973 that “any sufficiently advanced technology is indistinguishable from magic.” Known now as Clarke’s “third law,” it was by no means fiction. In 1971, Intel sold the 4004, the world’s first microprocessor, with 2,300 transistors. It would launch today’s digital economy. When Apple introduced its Macintosh computer in 1984, it had the Motorola 68000, named for its 68,000 transistors. The Intel Pentium, with 3 million transistors, was the leader at the dawn of the web age in 1994. And now? The global champ is Nvidia’s GV100 Volta, with more than 21 billion transistors.
This stunning progression can run another five or six years. Processors in 2025 will have 300 billion transistors. In a global economy beset by turbulence, Moore’s Law is one sure thing. As we survey the 2019 China Rich List, take note: digital industries are now one of the greatest sources of wealth creation in China, rivalling traditional fields such as property and manufacturing. A surprise? Not if you factor in the awesome exponential gains of computing power.
Coupled with more computing power is the rise of artificial intelligence. AI in some form has been around since the 1950s—but only now is computing power sufficient to make AI practical. If AI is the next route to big wealth, where does China stand versus the other AI superpower, the U.S.? If the U.S.–China trade war continues, can China produce a domestic chip industry to power its AI ambitions and keep pace with the likes of Intel, Nvidia and Qualcomm in the U.S., and Samsung and TSMC in Asia?
First the chip question. Computing power’s exponential gains are due to shrinking transistors. Nvidia’s GV100 Volta has transistors 7 nanometers (7 billionths of a meter) in length. A sheet of paper is 100,000 nanometers thick, and anything below 400 nanometers is invisible to the naked eye. By 2021, the top chip makers will be down to a 5-nanometer standard, only twice the size of a single DNA strand.
Although Huawei is a world-class chip designer, no chip company in China is yet capable of the 7-nanometer standard. Thus Huawei, as it parts ways with Qualcomm, still needs to manufacture chips with Samsung and TSMC. Overall, China manufactures less than 20% of the chips it consumes. Can China catch up? That is one goal of China’s 2025 technology plan—but it will undoubtedly take China five to 10 years to do so. Meanwhile, China’s booming tech sector must rely on imports.
As for AI, the race between China and the U.S. is tightening. China may already be ahead of the U.S. in areas such as computer vision and face recognition. Yet most of China’s top AI researchers were U.S. trained and, for now, the majority of the world’s AI development tools and open-source platforms are U.S.-based (Baidu’s impressive PaddlePaddle platform notwithstanding). The big risk for the U.S. is political error—driving away AI talent by restricting immigration and visas.
Rich Karlgaard is editor at large at Forbes. As an author and global futurist, he has published several books, the latest of which is Late Bloomers, a groundbreaking exploration of what it means to be a late bloomer in a culture obsessed with SAT scores and early success. For his past columns and blogs visit our website at www.forbes.com/sites/richkarlgaard