July 20, 1969, is a day no 13-year-old (at the time) boy in America can forget. The Eagle landed and there was one small step for man and one giant leap for mankind. America had won the race to put a man on the moon. The competition was between the American capitalists and the Soviet communists. The Soviets jumped ahead with Sputnik in 1957 and Yuri Gagarin in 1961. On September 12, 1962, President John F. Kennedy declared America would make it to the moon by the end of the decade, a claim that seemed nearly impossible at the time.
Less than seven years later on July 13, 1969, the Soviets launched the Luna 15 robotic probe to the moon. Three days later the Americans launched Apollo 11, carrying a crew of three astronauts. The Soviet’s Luna intended to perform a soft landing on the Moon, drill and collect a sample, and then launch itself back to Earth. The American lunar module Eagle landed on July 20 with Neil Armstrong and Buzz Aldrin inside. The following day, about 350 miles away from the American astronauts, the Soviet’s Luna attempted a landing but instead crashed into the surface of the Moon going 298 miles per hour and was destroyed. The space race ended on that day.
What gave America the edge? The Soviets had brilliant engineers, but they didn’t have the depth and talent of the American team, which included German scientists like Wernher von Braun who brought advanced rocket technology to the American space program. The US also had more economic resources to dedicate toward space exploration and thousands of firms that were intensely competing to provide NASA with new ideas and technologies.
The Soviets lost the Space Race but continued to challenge the Americans with an expansionist military. By 1980 the US appeared to have fallen behind the USSR as a military power. Then Ronald Reagan showed up with a vision to defeat the Soviets through moral leadership and the superior economic strength of free markets. America also outspent the Soviet Union in the arms race and new innovations like the Strategic Defense Initiative, otherwise known as Star Wars. Reagan delivered his “Evil Empire Speech” on March 8, 1983, to the National Association of Evangelicals. Meanwhile, the Soviets’ central planning, economic stagnation, and political repression doomed them to failure. Less than seven years after Reagan’s speech, the Cold War ended on November 9, 1989, when the Berlin Wall came down. Capitalism wins again.
A New Cold War?
Many suggest we face a new Cold War with China over Artificial Intelligence. Signaling the seriousness of this issue, President Trump, and three AI promoters held a press conference on January 21, 2025, to announce Stargate, a joint venture to fund a $500 billion artificial intelligence program. Then, over the course of the next 10 days, three announcements shook the AI world: DeepSeek, TinyZero, and DeepAgent.
As a result of these developments, it appears that Stargate may fail to launch at all.
- DeepSeek: OpenAI and Google Gemini have been spending between $100 million and $200 million to train their models. On January 28th, DeepSeek, an effort funded by a Chinese hedge-funded company, reported developing a new AI system that matches OpenAI’s systems for $5 million. DeepSeek’s model uses 75 percent less memory, reading whole phrases instead of single words, and instead of one massive AGI (Artificial General Intelligence) trying to know everything, they built a system of specialists. Traditional models rely on 1.8 trillion parameters being active all of the time. DeepSeek, by contrast, relies on 671 billion in total, but only 37 billion are active at once, 97.9 percent fewer than OpenAI.
DeepSeek uses 2,000 GPUs instead of 100,000 and runs on gaming GPUs instead of specialized hardware. The DeepSeek system is open source, which means anyone can verify, build upon, and implement these innovations. You can download the new app on your iPhone. Bonus: AI now has a counterpoint to the environmentalists who say AI uses too much electricity. DeepSeek just brought down the cost of inference by almost 98 percent.
- TinyZero: But you don’t need to buy a warehouse of Nvidia cards to do AI. You can rent them by the hour. On January 31, PhD candidate Jaiya Pan and his research team at the University of California at Berkeley announced TinyZero. This new model achieved DeepSeek-level performance by renting two H200 Nvidia chips for under five hours at just $6.40 per hour, for a total of $30. TinyZero is also open source and freely available for download on GitHub.
- DeepAgent: Then on February 3, Liang Chen and a team of PhD computer science students at Peking University and University of Hong Kong introduced Deep-Agent/R1-V, a new AI model trained on 8 A100 GPUs for 30 minutes, costing $2.62 in total. Their code, models, datasets, more details, and all open-source resources are also available on Github.com.
Disruptive innovation is disrupting disruptive innovation. Clayton Christensen, the originator of the disruptive innovation theory, would be pleased.
After the death of Mao, the Chinese Communist Party publicly recognized the power of free-market capitalism to lift people out of economic poverty. From 1978 to 1989 Deng Xiaoping served as the leader of China and is credited with transforming China’s economy into a “socialist market” economy.
China made astonishing economic progress, moving from 95 percent of their population living on under $5 a day in 1990, to less than 5 percent living on that amount in 2020. Many said that while China could catch up, that it couldn’t innovate. Chinese AI development is challenging this notion.
The stunningly low cost of these new models underscores a growing trend: while tech giants pour vast sums into proprietary AI development, open-source and independent researchers are proving that high-performance AI can be built at a fraction of the cost. (Some have suggested that China is stealing intellectual property from OpenAI and others. This may be true. If this were the case, however, DeepSeek, TinyZero, and DeepAgent would probably not have published all of their code, models, datasets, more details, and all open-source resources on Github.com.* )
What Would Hayek Do?
The Stargate Project looks very much like a replay of FDR and the central planners of Keynes. The effort in the US to create a cartel to regulate and control AI development through Stargate appears more like the construction of a new DMV to protect BigAI from competition. Hayek explained why a system of free markets with distributed knowledge and know-how was far superior to central planning. DeepSeek, TinyZero, and DeepAgent’s approach looks Hayekian. Capitalism recognizes the potential worth of every individual to create value in a free market. In America, we’re not surprised when a small team of brilliant thinkers can create an entirely new approach to solving a problem.
Hayek would argue that our policy should focus on activating billions of people to climb AI learning curves. Rather than competing under the false ideology of scarcity — fighting over a finite supply of atoms — we should be cooperating in an open system of discovery, where knowledge is abundant and hyper-exponentially scalable.
Who Can Learn the Fastest?
AI will accelerate our ability to uncover valuable knowledge, but we’ll also realize that the only true intelligence behind artificial intelligence is human intelligence. When people have the freedom to innovate, the capacity to create resources is limitless.
This isn’t a battle between China and the US — it’s a contest between abundance and scarcity, between entrepreneurial capitalism and bureaucratic centralization.
The real question is: Will the future of AI be Silicon Valley or Shanghai?
*Editor’s Note: This article was updated to address allegations that China steals AI technology.
