After the sudden and prolific rise of Chinese AI company DeepSeek, and the accompanying drop in shares of US AI firms, President Donald Trump has called the dramatic shift a “wake-up call” for the US tech industry.

It seemed only a week ago when US tech giants thought themselves untouchable, as they continued to change their internal business values whilst announcing major funding programmes to bolster the infrastructure necessary for them to harness more energy to power their AI models.

Now, as shares in firms fall sharply, including for AI chip giant Nvidia, the scene on the AI market, and for AI in general, has transformed.

“The release of DeepSeek, AI from a Chinese company should be a wakeup call for our industries that we need to be laser-focused on competing to win,” Trump said to reporters.

DeepSeek is now the number one most downloaded free app in the US, just a week after it was launched on the Apple App Store.

The firm claims that its AI model, R1, rivals the Large Language Models of OpenAI for just a fraction of the price, using less compute power, without premium chips.

DeepSeek claims their open source DeepSeek-V3 model was trained for around £4.2m, just a mere fraction of what it cost to train rivals such as the GPT series from OpenAI. However, this claim is being disputed by AI leaders, who have shared doubts on if the model functions as well or cost as little as the company claims.

If proven true, however, then it surely will rock the AI world, as well as the strategy behind technological advancements.

Sanctions forced many Chinese AI company’s hands: without access to advanced AI chips from major firms like Nvidia, Chinese firms collaborated and experimented with new methods, ultimately improving AI’s ability to work with less compute power.

DeepSeek founder Liang Wenfeng stockpiled some of Nvidia’s top performing computer chips which were later restricted for Chinese imports, meaning that after developing much of DeepSeeks technology, Wenfeng’s firm then worked on how to get their AI platform onto less powerful, cheaper chips.

The method seemed to work, though DeepSeek is already facing some issues.

Its surge in popularity caused the platform some technical difficulties, and the firm has reportedly already suffered from a targeted cyber-attack.

If DeepSeek can be seen as a wake-up call, OpenAI seems to have shot out of bed. CEO of OpenAI Sam Altman did give the new competitors credit online, calling DeepSeek’s R1 “an impressive model, particularly around what they’re able to deliver for the price.” He called it “invigorating” to have a new rival in the AI arena, as well, though he claimed OpenAI will deliver better models.

Tech giants are nervous as the shares in Meta, Microsoft, and Nvidia suffered as DeepSeek rose in prominence.

While Trump called the shift a wake-up call, the President – to which top tech giant chiefs have cosied up to in recent weeks – said that making a cheaper AI product may be ultimately good for competing US firms.

“If you could do it cheaper, if you could do it [for] less [and] get to the same end result. I think that’s a good thing for us,” he said aboard Air Force One.

Recommended reading

• Could DeepSeek Topple the US’s AI Market Dominance?
• Is the UK Lagging Behind in Global AI Race?
• Gartner: Global AI Chips Revenue to Total £56BN in 2024

While it remains to be seen if top US AI competitors will begin to collaborate like Chinese firms, the wake-up call from DeepSeek could greatly shift investment priorities and dramatically disrupt the AI supply chain.

The AI race was largely constrained by compute power and the availability of advanced – and expensive – computer chips, hence the US’s sanctions on chip imports to disable China’s AI industry.

If DeepSeek continues to prove its capabilities without this, and without as much power consumption or money for training, the landscape can change dramatically.

While still a significant sum for training, the ability to train an AI that can rival OpenAI’s models in certain benchmarks can do much to level to playing field to smaller AI developers and firms, potentially toppling the dominance of tech giants as their net worth better insures the massive startup investments AI models require.

Further, the development of an advanced LLM on cheaper chips can tumble the strategy of chip developers racing to produce advanced computer chips with more processing power, as well as the push for more data centres to power energy-hungry AI systems.

It is understandable, then, that DeepSeek’s AI claims should be a wake up call for what tech giants are trying to achieve in terms of AI advancements.