Until this week, little was known about DeepSeek—the Chinese AI-powered chatbot that has suddenly taken the U.S. market by storm, rattling tech stocks and stunning investors who never saw it coming. Its founder, Liang Wenfeng, remains an enigmatic figure in the U.S. market, lacking ties to Silicon Valley or American tech. Yet his unexpected foray into AI, which began in December 2023, has quickly become the talk of the town. But what’s behind the buzz?
Why is Deep Seek taking the U.S. AI market by storm?
Since the rise of Open AI’s ChatGPT, Chinese tech firms have been racing against the clock to develop something equally impressive, if not far superior. While several models have emerged in China, none have come close to offering the global competition that DeepSeek now vies for. High-Flyer, the hedge fund behind DeepSeek, has a long history in AI research, specializing in data financial analytics. According to its website, the hedge fund boasts a team of distinguished mathematicians and physicists, holding advanced qualifications and medals in statistics, research and cybernetics.
DeepSeek-R1, its latest reasoning model released last week, operates similarly to popular AI chatbots like Google’s Gemini or OpenAI’s ChatGPT, offering a familiar interface and range of functions. The model has caught the attention of developers, investors and everyday AI users alike for one standout reason: It could be 20 to 50 times cheaper to use for some tasks than OpenAI’s o1 model while delivering comparable performance. According to DeepSeek, only $5.6 million was spent to power its latest base model, not including research and prior experiments. Still, it’s only a fraction of the billion-dollar investments U.S. competitors have poured in over recent years.
This cutting-edge AI service is at the heart of China’s push to dominate the global AI industry. On the day DeepSeek-R1 went public, Wenfeng was invited to a symposium of tech experts hosted by Chinese Premier Li Qiang. He and other AI pioneers are being celebrated as key figures in China’s technological future, tasked with safeguarding the nation’s self-sufficiency amid Washington’s push for total control over AI’s future. If its explosive AI expansion continues, the nation of over 1.4 billion people is on the brink of an unprecedented investment boom. Asset management firm Artisan Partners says China’s AI investments could boost its economy by an additional $7 trillion by 2030.
DeepSeek launch erases billions from Nvidia: the biggest 1-day wipeout ever
DeepSeek’s rise in the global AI market has been swift and staggering. Just days after its launch, the app surpassed ChatGPT as the most downloaded on Apple’s Top Free Apps chart. Meanwhile, Nvidia, the leader in AI chips, saw its stock plunge 17%. Nvidia experienced a historic loss of around $593 billion in market capitalization, marking a record one-day value wipeout for a company. It’s a significant setback for Nvidia, which has profited from the chip boom, driven by the belief that innovation would be even more dependent on its technology.
DeepSeek has reported that training DeepSeek-V3 required less than $6 million and took just months to build. It’s not just cost that has enabled DeepSeek to suddenly outperform its American competitors; the Chinese platform has also stuck to its promise of providing a true open-source service. This means researchers, businesses and developers can freely access and modify its code for personal projects or analysis at no additional cost.
OpenAI was initially founded as a nonprofit with a mission to provide global access to AI for a positive impact. It has since transitioned into a for-profit model, shifting away from its open-source beginnings. Its CEO, Sam Altman described DeepSeek’s model as “impressive” but emphasized his commitment to delivering superior models through OpenAI. Altman will be the first to ask where his model now needs an urgent shake-up.
Silicon Valley executives are facing a moment of reckoning as they come to terms with the rapid progress of overseas competitors. The purported cost-efficiency of DeepSeek serves as a stark reminder that the U.S. can no longer aspire to monopolize AI as it once did. Disruption is inevitable as bold new players enter the scene, but where they’ll emerge from remains uncertain. Meanwhile, Britain has this month re-declared itself as a future “world leader” in AI, boasting the highest number of generative AI startups in Europe and is also in the stages of planning an ambitious “homegrown rival” to ChatGPT.
Cyberattack forces DeepSeek to halt sign-ups as U.S. Navy bans platform
Just like TikTok and its national security concerns, DeepSeek is now facing scrutiny over its content moderation policies, data protection practices and potentially troubling political connections. DeepSeek’s privacy policy reveals that user data is kept “in secure servers located in the People’s Republic of China.” The company also collects data on “device model, operating system, keystroke patterns or rhythms, IP address and system language.” In compliance with Chinese cybersecurity regulations, all China-based companies are obligated to share data with the government upon request. Given DeepSeek’s current operations, it could be sending more U.S. user data back to China than TikTok, which moved to U.S. cloud services at Oracle to ease security worries in 2022. Warnings over DeepSeek’s links to the communist regime have already prompted lawmakers to propose stricter controls and limitations on data use this week. On Tuesday, the U.S. Navy also raised concerns over the “potential security and ethical concerns” of DeepSeek, leading to a ban on the use of the tool. Navy personnel have been instructed not to use DeepSeek for either work-related and personal purposes.
Additionally, DeepSeek halted registrations Monday after claiming “large-scale malicious attacks” were affecting their services. For American users anxious about their data security, this news offers little comfort. The motive behind the attacks is still unknown, but operations have resumed without fail, with DeepSeek indicating “a fix is being implemented.”
The chatbot strictly follows government censorship rules, steering clear of sensitive topics and avoiding criticism or obvious diversions from the One China Policy. It’s said to align closely with the official stance on sensitive issues, much like other platforms such as RedNote that must maintain conformity with the state’s approved content.
Silicon Valley forced to confront AI’s real costs and open-source future
For years, U.S. AI firms have focused on hyperscaling, building huge data centers and driving up energy demands. But that approach now appears flawed as DeepSeek finds ways to sidestep the financial and logistical barriers slowing others down. Realizing that more agile and cost-effective rivals can overtake them is a challenge the U.S. tech industry must address.
Maintaining dominance in AI was always destined to be short-lived. Washington’s crackdown on foreign tech may slow China and DeepSeek’s rise, but it doesn’t address the root of past domestic setbacks. Overlooking lesser-known global startups is a costly mistake, one that investors and developers are only now coming to terms with. And the trend isn’t slowing down: Chinese tech giant Alibaba has also just made its move, unveiling the Qwen 2.5 AI model, which it claims surpasses DeepSeek-V3. Tech giants like OpenAI, Google and Meta must now step up—proving they have the financial muscle and innovation to compete with China.
To stay competitive in Silicon Valley, companies must justify to DeepSeek and others that they remain powerful players—capable of learning from missteps and leading future innovation. Their top priority will be to cut prices and chart a smarter course to success, stepping away from the chip-collecting frenzy that’s distorted what was once a practical vision. Silicon Valley’s next AI breakthrough will hinge on a return to ingenuity, not just excess resources.
Photo by Mojahid Mottakin/Shutterstock
