[News] DeepSeek Disrupts AI Landscape: Potential Impacts and Comments from Tech Giants

As Chinese AI startup DeepSeek shakes up the global market by claiming its R1 model surpasses OpenAI’s latest o1 while costing only a fraction of the price, it is also sending a clear message. As highlighted by Nikkei, DeepSeek is forcing a market-wide rethink—on valuations, competition, and even the effectiveness of U.S. restrictions on high-end chips.

According to CNBC, DeepSeek claimed that its large language model was trained for just $5.6 million. Unlike traditional models, R1 breaks prompts into smaller segments and explores various approaches before responding, designed to tackle complex problems in a way that mirrors human thinking, the report adds.

Market Disruptor in Many Ways

The Nikkei report notes that this isn’t a collapse for NVIDIA’s empire, but instead, it’s a market reset. As indicated by Nikkei, DeepSeek isn’t NVIDIA’s rival; in fact, it still relies on NVIDIA chips and likely will for the foreseeable future.

By optimizing the performance of NVIDIA Hopper downscaled chips, DeepSeek maximizes computational resource utilization, according to TrendForce.

However, DeepSeek’s rise does indicate that NVIDIA’s dominance may not be untouchable, as the Chinese AI firm has proven that cutting-edge performance doesn’t require billion-dollar budgets or a stranglehold on semiconductor supply chains. Meanwhile, AI giants like OpenAI, Google DeepMind, and Meta must rethink their once unquestioned strategies for heavy AI investments, as noted by Nikkei.

TrendForce’s latest investigations have revealed that the recent release of DeepSeek-V3 and DeepSeek-R1 underscores an industry-wide shift toward more cost-effective AI infrastructure. This development is expected to prompt end users to conduct more rigorous evaluations of AI infrastructure investments, focusing on adopting more efficient software computing models to reduce reliance on hardware such as GPUs.

On the other hand, according to Nikkei, while DeepSeek’s edge lies in advanced long-text comprehension and reasoning, this makes it a strong contender in finance, law, and technical research.

A Wake-up Call for U.S. Chip Curbs

Notably, DeepSeek’s emergence may also imply that cutting off access to top-tier chips may have only sped up the push for more efficient AI. As highlighted by the Nikkei report, the AI race is shifting from chip dominance to real-world applications, favoring China’s strength in strategic deployment.

Therefore, the key lesson learned from DeepSeek may be that restricting China’s chip access was meant to curb its AI growth, but this would not work.

As highlighted by Nikkei, AI dominance will hinge on how nations effectively implement and scale AI solutions. Thus, DeepSeek’s rise could be interpreted as a wake-up call to those who still see AI as a battle of raw computational power, the report adds.

More Comments from Tech Giants

Interestingly enough, DeepSeek’s progress has sparked a broader conversation about the impact, with tech giants like ASML, Microsoft, and Intel’s former CEO voicing their opinions.

Even NVIDIA had a word about this. An NVIDIA spokesperson, cited by a CNBC report, described DeepSeek’s R1 model as “an excellent AI advancement,” noting that the company’s work shows how new models can be created using this technique with widely available models and computing that comply with export controls.

According to Reuters, citing ASML’s CEO Christophe Fouquet, the emergence of efficient AI models like those launched by DeepSeek is actually beneficial for the chip market. “If the cost doesn’t go down, you may continue to sell a few very expensive chips and make a few people very happy, but you will not basically bring this technology to the masses and the volume will remain small,” he explained.

Anything that will drive costs down is good news for ASML on the long term, Fouquet said, as per Reuters.

On the other hand, the so-called “hyperscalers” which invest big in AI like Google, Meta, and Microsoft, in ASML’s words, seems to be taking a positive view on DeepSeek. According to Fortune, citing Microsoft CEO Satya Nadella, DeepSeek has made some real innovations, and that DeepSeek’s efforts will eventually become commoditized and widely adopted across AI platforms.

Notably, DeepSeek R1 is now available in the model catalog on Microsoft’s Azure AI Foundry and GitHub, joining a diverse portfolio of over 1,800 models, including frontier, open-source, industry-specific, and task-based AI models, as noted by Microsoft’s press release.

Intel’s former CEO Pat Gelsinger, who has become an angel investor in Fractile.ai, a British AI chip startup, spoke on X that instead of reducing demand, making computing “dramatically cheaper” and more efficient — as DeepSeek seems to have achieved — will actually expand the market.

“Computing obeys the gas law. Making it dramatically cheaper will expand the market for it. The markets are getting it wrong, this will make AI much more broadly deployed,” he noted.

Wisdom is learning the lessons we thought we already knew. DeepSeek reminds us of three important learnings from computing history:
1) Computing obeys the gas law. Making it dramatically cheaper will expand the market for it. The markets are getting it wrong, this will make AI…

— Pat Gelsinger (@PGelsinger) January 27, 2025

(Photo credit: DeepSeek)