Responding to the DeepSeek Moment

DeepSeek’s whale logo is a perfect metaphor for a startup that seemingly leaped out of nowhere and landed with a colossal splash that left the AI world surprised and soaking wet. The Chinese company’s R1 model appears to perform nearly as well as OpenAI’s o1 model, and was produced for a fraction of the cost. It’s a startling moment, and no doubt many American companies and political leaders are concerned with how to reassert dominance in innovative AI. They would be wise to consider basic research as a response.

Basic research is scientific or technical research that is conducted with no immediate application intended. In a sense, it’s learning for the sake of learning. Scientists, researchers and scholars are funded to take the time to think broadly and ask questions. They are not constrained to focus on commercial applications. Instead, smart people are given the latitude to explore and see where innovation takes them.

While this may seem like an unusual tactic for addressing the particular objective of how to lead in global AI, it has historically been shown to support specific goals.         

Consider the Cold War. America and the Soviet Union were locked in a desperate race to decide who would rule the Atomic Age. The Soviets were first into space with Sputnik, a moment that feels similar to the DeepSeek breakthrough. But America’s investment in basic research soon led to a variety of discoveries across multiple disciplines that propelled U.S. innovation.

Unexpected and disparate findings were put to military and commercial use: GPS, the internet and touchscreen devices. Had scientists been limited to thinking only about particular technology uses from the outset, they may have missed the winding path to unexpected breakthroughs. The same logic applies here. Give smart people more tools to conduct broad research and some of the findings will be applicable to advancing our thinking around artificial intelligence.        

The U.S. currently spends a considerable amount on basic research: approximately 3-4 percent of our GDP. But R&D spending increases have been modest.? Meanwhile, other countries are making more muscular investments to catch up with American innovation. China has increased their national R&D spending by 18x since 2000. Clearly, they’re going all-in on this type of research.

Increased research spending also builds on America’s inherent advantages: a diversity of ideas, freedom of thought, and an iconoclast ethos that can’t help pushing boundaries. These traits are hard to imitate and offer the keys to future innovation in AI and beyond.