[News] Deepseek disturbs the AI ​​landscape: Potential effects and comments from Tech giants

Since the Chinese Ki -Startup Deepseek shakes the global market by claiming that its R1 model surpasses Opena's latest O1 and only costs a fraction of the price, but also a clear message. As highlighted by Nikkei, Deepseek forces a market-wide rethinking in reviews, competition and even the effectiveness of the US restrictions on high-end chips.

According to CNBC, Deepseek claimed that his large voice model was trained for only 5.6 million US dollars. In contrast to conventional models, R1 calls for smaller segments and examines various approaches before they react to tackle complex problems in a way that reflects human thinking, the report adds.

Marketers in many ways

The Nikkei report notes that this is not a breakdown for the Nvidia empire, but is a market reset. As indicated by Nikkei, Deepseek is not Nvidia's rivals; In fact, it still relies on Nvidia chips and is likely to be for the foreseeable future.

By optimizing the performance of Nvidia Hopper -Downscaled -Chips, Deepseek maximizes the computing resource utilization according to Trendforce.

However, the rise of Deepseek shows that the dominance of NVIDIA may not be inviolable, since the Chinese AI company has proven that the high-ranking performance is not necessary with billions in dollars or a strangle area for the supply chains of semiconductor supply. In the meantime, Ki -Giants such as Openaai, Google Deepmind and Meta have to rethink their once undisputed strategies for heavy AI investments, as found by Nikkei.

Trendforce's recent investigations have shown that the recent publication of Deepseek-V3 and Deepseek-R1 underlines an industry-wide shift into more cost-effective AI infrastructure. This development is expected to make end users carrying stricter reviews of AI infrastructure investments, which focuses on the introduction of more efficient software computer models in order to reduce the dependency on hardware such as GPUs.

On the other hand, according to Nikkei, while Deepseek's border lies in an advanced long -term understanding of long -term understanding and rain, a strong contender for finance, law and technical research.

A wake -up call for us chip curbs

In particular, Deepseek's origin can also imply that the performance of access to top animal chips may have only accelerated the advance for a more efficient AI. As the Nikkei report emphasized, the AI ​​race shifts from the dominance of the chip to real applications and favors China's strength in strategic use.

Therefore, the key lesson learned by Deepseek can be that the restriction of the China chip access should contain its AI growth, but this would not work.

As highlighted by Nikkei, the AI ​​dominance will prevent how the nations effectively implement and scale AI solutions. Therefore, Deepseek's climb could be interpreted as a wake -up call to those who still see AI as a battle of raw computing power, the report adds.

More comments from technology giants

Interestingly, Deepseek's progress has triggered a broader conversation about the effects, with Tech giants such as ASML, Microsoft and Intel's former CEO commenting.

Even Nvidia had a word about it. An NVIDIA spokesman who was cited by a CNBC report described the R1 model from Deepseek as “an excellent AI progress” and found that the work of the company shows how new models with this technique creates with far -reaching models can be, and computers that correspond to the export controls.

According to Reuters, the advancement of efficient AI models such as that from Deepseek on the chip market is an advantage if he is advantageous for the chip market from ASML from ASML -CEO, Christiophe Fouquet. “If the costs do not fall, they may continue to sell a few very expensive chips and make a few people very happy, but they will not basically put this technology into the masses and the volume will remain small,” he said.

Everything that the costs reduce is good news for ASML in the long term, said Fouquet, according to Reuters.

On the other hand, the so -called “Hyperscalers”, which invest in AI such as Google, Meta and Microsoft in ASML's words, seem to be a positive view of Deepseek. According to Fortune, Deepseek, CEO of Microsoft CEO Satya Nadella, made some real innovations, and that Deepseek's efforts finally be assumed and widespread for the entire AI platforms.

In particular, Deepseek R1 is now available in the model catalog of Microsoft Azure Ai Foundry and Github and contributes to a diverse portfolio of over 1,800 models, including border, open source models, industry-specific and task-based AI models, as specified by Microsoft.

Pat Gelsinger, former CEO of Intel, who an angel investor in fractile.ai, a British Ki -Chip startup, was spoken to X instead of reducing demand and making the computer “dramatically cheaper” and more efficient -as Deepseek apparently It seems to have actually been reached – Will – actually expand the market.

“Computing follows the gas law. If you make it dramatically cheaper, the market will be expanded. The markets do it wrong, this will make the AI ​​much wider, ”he noticed.

Wisdom learns the lessons we have already known. Deepseek reminds us of three important knowledge from computer history:
1) Calculation of the gas law. If you make it dramatically cheaper, the market will be expanded. The markets misunderstand it, that will do AI …

– Pat Gelsinger (@Pgelsinger) January 27, 2025

(Photo credit: Deepseek)