On October 4, 1957, the Soviet Union launched Sputnik 1, the world's first artificial satellite. The event triggered a significant response from the United States, sparking fears of Soviet technological superiority and initiating the space race during the Cold War. The launch led to increased investment in science and technology education and the establishment of NASA the following year. “Deepseek R1 is AI’s Sputnik moment,” says venture capitalist Marc Andreessen.
DeepSeek is an AI development firm based in Hangzhou, China that specializes in open-source large language models (LLMs). Founded in 2023, the company’s aim is to improve AI development by reducing inefficiencies and costs. Its first model was released in November 2023, and after several iterations, it released its R1 LLM last month. The model, which is available under an open-source license, grants users free access and has been compared to other top performing reasoning models including OpenAI’s ChatGPT. Most significantly, however, DeepSeek was able to create and train R1 at a substantially lower cost—20 to 50 times lower—with only a few months of development. As a result of the company’s emphasis on cost, efficiency, and open-source collaboration, it has positioned itself as a disruptive force in the AI industry. The implication is that DeepSeek will have a cascading effect on AI valuations. Already, developers are testing R1 as a cheaper alternative. Indeed, one of DeepSeek’s previous iterations, DeepSeek-V2, sparked an AI price-war in China last year. Chinese tech giants ByteDance, Tencent, Baidu, and Alibaba each reduced their LLM prices to stay competitive. After R1 abruptly became the number one downloaded free app on Apple’s App store, it stands to reason that a similar adjustment is plausible in the United States. If Sputnik serves as a guide, perhaps DeepSeek’s success and innovation will act as a catalyst and springboard to spur American AI advances.
DeepSeek Dilemma: The Legal Risks
DeepSeek’s status as a China-based company presents unique legal and policy challenges for investors, particularly regarding copyright infringement, data privacy compliance, contractual liabilities, and evolving geopolitical dynamics. As an open-source model, DeepSeek allows developers to leverage its data and architecture to create their own generative AI systems. However, allegations of unauthorized training on proprietary data have raised legal uncertainties that could impact venture capitalists and startups relying on its technology.
Shortly after R1 emerged in the U.S. market as a competitor to ChatGPT, Claude, and Gemini, OpenAI alleged that DeepSeek had trained its model on OpenAI’s proprietary data without authorization. While OpenAI itself has faced similar claims from entities such as Intercept Media, Author’s Guild, and the New York Times, it has largely defended its practices under the doctrine of fair use. These lawsuits remain unresolved, and no clear common law precedent yet exists on whether using copyrighted material to train LLMs constitutes infringement.
Should courts rule against AI developers in these cases, the implications for DeepSeek—and by extension, its downstream users—could be significant. If DeepSeek is potentially liable for infringement, other LLMs incorporating its open-source code may face derivative liability. This scenario could pose financial and operational risks for venture capitalists investing in generative AI startups, potentially reducing returns if companies are forced to retroactively license training data or overhaul their models to comply with legal mandates.
In addition to intellectual property risks, data privacy compliance is another critical concern for startups utilizing DeepSeek’s technology. International data protection regulations and security considerations have already led to restrictions on DeepSeek’s use in Taiwan, Italy, and Australia. Crucially, DeepSeek stores private user data on servers located in China, raising concerns about compliance with regulations such as the General Data Protection Regulation (GDPR) and other national data security laws. Reliance on DeepSeek’s LLM infrastructure could expose AI startups and their investors to enforcement actions or compliance burdens, particularly in jurisdictions with strict cross-border data transfer rules. Given the growing regulatory scrutiny of AI models worldwide, investors must assess whether DeepSeek’s data practices align with, not only regulatory requirements, but their own privacy risk tolerances.
The legal uncertainties surrounding DeepSeek underscore the broader risks associated with new investment in open-source AI models. The potential for IP infringement claims, combined with a growing body of data privacy regulation, create a complex landscape for venture capitalists and portfolio companies operating in the generative AI space. As the legal framework around AI continues to evolve, courts and regulators will play a decisive role in shaping the future viability of models like DeepSeek. Until clearer precedent emerges, businesses integrating DeepSeek’s technology should proactively assess their exposure to IP liability, data security risks, and jurisdictional compliance challenges.
Another Sputnik? The United States Response
In response to the rising prominence of DeepSeek, U.S. leaders have implemented significant policy measures. These actions, driven by security concerns and the strategic need to manage China’s influence, set a complex backdrop for understanding the subsequent political and legislative responses.
On January 31, 2025, Texas Governor Greg Abbott announced a ban on the use of DeepSeek on all government-issued devices, a decision mirrored by the United States Navy, which also prohibited DeepSeek—citing security concerns.
Just two days earlier, on January 29, 2025, Republican Senator Josh Hawley proposed a significant legislative measure in response to global AI developments. His bill, Decoupling America’s Artificial Intelligence Capabilities from China Act, seeks to comprehensively restrict the use, importation, and exportation of artificial intelligence technologies developed in China. The bill includes heavy penalties for violations, with individuals facing up to $1 million in fines and 20 years in prison, and corporations facing fines of up to $100 million. The bill also aims to prevent U.S. residents from investing in Chinese AI development. According to Senator Hawley’s website, this legislative move was prompted by China's unveiling of DeepSeek.
Amidst these developments, David Sacks, the White House’s artificial intelligence and crypto czar, has expressed concerns about DeepSeek’s methods. He claims that there is “substantial evidence” that DeepSeek used a process called “distilling” to leverage and replicate technologies from OpenAI, amounting to what he described as “intellectual property theft.”
However, in a departure from his party’s cautious stance, President Donald Trump touted that the emergence of DeepSeek “should be a wakeup call” for America’s tech companies. President Trump views the model’s popularity and low-cost creation as a “positive development” for the artificial intelligence industry. This statement comes on the heels of an executive order undoing Biden administration rules concerning barriers to American artificial intelligence innovation. At this point, the executive branch’s definitive stance on DeepSeek remains undefined. However, it is clear that AI deregulation is meant to remove certain guardrails in order to clear the way for even greater investment. According to EY, AI-driven deals “increased fivefold from Q4 of last year, representing over 60% of all Q4 fundraising” last quarter. Those figures will continue to grow if the ambitious $500 billion Stargate project proceeds and the United States follows through on its bid to corner the AI capital market.
Conclusion
Although DeepSeek seems poised to impact current AI valuations, its users must balance the benefits against significant risks. The company gains broad access to user data, while users face potential technical, legal, and business liabilities. Nevertheless, DeepSeek’s rise coincides with AI deregulation in the United States, which is meant to spur investment in AI infrastructures. Developers will undoubtedly be looking to imitate DeepSeek in order to secure some of that funding while cutting costs.
*The views expressed in this article do not represent the views of Santa Clara University.
Comments