The news that two of the biggest players on the global tech stage are competing to invest in DeepSeek is not just a business headline; it is a confirmation of the tectonic shift occurring in the Artificial Intelligence industry. According to a report from eu.36kr.com, the Chinese startup, which managed to shake the dominance of OpenAI and Nvidia in early 2025, is now at the center of an investment war. While the names of the two companies have not been officially confirmed, analysts are pointing toward Chinese giants like ByteDance and Alibaba, or even Western entities looking for ways to bypass the high cost of computing power.
The Rise of DeepSeek and the Disruption of Scaling Laws
For years, the Silicon Valley dogma was simple: more data and more processors (GPUs) equal smarter AI. DeepSeek shattered this myth. With the release of its DeepSeek-V3 and DeepSeek-R1 models, the company proved that clever architecture and advanced algorithms can deliver GPT-4 level performance at a fraction of the training cost. This "efficiency" is what has triggered the investor frenzy. In a world where Nvidia charges tens of thousands of dollars for every H100 chip, the promise of an AI that requires less energy and less hardware is the "holy grail" of profitability.
DeepSeek, based in Hangzhou, began as a research lab backed by High-Flyer Quant, a hedge fund specializing in quantitative analysis. This background allowed them to approach AI not as a linguistic problem, but as a resource optimization problem. Their success in reaching the top of benchmarks using only about 2,000 Nvidia chips, while competitors use hundreds of thousands, has permanently altered the investment map.
Geopolitical Chess and Strategic Importance
The competition to invest in DeepSeek is about more than just profits. It’s about controlling intellectual property in an era of intense geopolitical rivalry between the US and China. US export restrictions on high-end chips to China forced Chinese companies to become more creative. DeepSeek is the brightest example of this "creativity under pressure." The two prospective investors seem to recognize that whoever controls DeepSeek’s technology will have the edge in the next phase of AI, where efficiency will be more important than raw scale.
- Training Costs: DeepSeek reduced the cost of training large models by nearly 90%.
- Open Weights: Its strategy of publishing model weights has created a massive ecosystem of supporters.
- Resilience: Its ability to operate in resource-constrained environments makes it ideal for edge computing and local applications.
ByteDance, for instance, could use DeepSeek’s technology to optimize TikTok’s recommendation algorithms with minimal additional hardware costs. On the other hand, Alibaba could integrate these models into its cloud services, offering customers the most economical AI on the market. The competition is fierce because the stakes involve survival in a market cannibalized by energy costs.
The Challenge of Independence
Despite the intense interest, DeepSeek faces a dilemma. Accepting a massive investment from a tech giant could limit its freedom to publish its research as open source. So far, the company has gained the trust of the global developer community precisely because it is transparent. If it turns into a "closed" lab for a large corporation, it risks losing the advantage of rapid improvement through community feedback.
"DeepSeek is not just a company; it is proof that intelligence isn't always bought with brute force, but with elegant design," says a market analyst cited by eu.36kr.com.
As we head into the second half of 2026, the outcome of this investment battle will determine whether AI remains a game for the few and the wealthy, or if DeepSeek’s approach will democratize access to top-tier technology. What is certain is that the days of reckless capital spending for compute power are coming to an end, as efficiency emerges as the new king of the industry.
