A shock to the entire AI world

‍

On January 27, 2025, the world's press reported on the DeepSeek event that shocked the AI world:

‍DeepSeek's AI Bomb: $600 Billion Wiped Off Nvidia's Value Overnight” – Fortune

“China's Sputnik Moment in AI: How DeepSeek Shattered Silicon Valley's Illusions” – The Guardian

“Nightmare on Wall Street: DeepSeek Triggers $1 Trillion Tech Meltdown” – The New York Post

“DeepSeek's AI Revolution: The Day Nvidia Lost Its Crown” – The Wall Street Journal

The tectonic plates of the global tech industry shifted overnight. It wasn't a slow, gradual change; it was a brutal rupture, a lightning-fast recalibration of everything investors, engineers, and political leaders thought they knew about artificial intelligence.  

On January 27, 2025, DeepSeek, a virtually unknown Chinese artificial intelligence company,unveiled a model called R1, which threw Nvidia, the undisputed giant of the AI revolution, into chaos, causing the largest loss of value in financial market history: $600 billion in a single day. The entire AI sector lost over a trillion dollars. This was not just a bad day on the stock market. It was a moment of existential crisis.  

What was DeepSeek's "fault?" It dared to contradict the most fundamental assumption of the AI industry: that advanced artificial intelligence requires billions of dollars, colossal amounts of computing power, and the most expensive chips ever produced.  

The West operated under this belief as if it were a law of nature. DeepSeek shattered it. The R1 model would be comparable to OpenAI's ChatGPT or Google's Gemini, but was developed on a budget of just $5.6 million.  Not billions. Millions. If this is true, not only does DeepSeek become a new competitor, but the entire economic architecture of artificial intelligence could be a mirage.  

The fall of a titan

For years, Nvidia has been the heart of AI's digital gold. Every model, every discovery, every step towards general artificial intelligence needed one thing above all else: more Nvidia chips. The company's value skyrocketed, driven by insatiable demand, a limited supply of chips, and the belief that no one else could keep up.  

Then came DeepSeek

With a single revelation, the company not only hit Nvidia's stock, it shook the very philosophical foundation on which its trillion-dollar empire rested.  

If startups like DeepSeek can achieve the same AI performance without spending billions on Nvidia chips, why should companies pay these exorbitant sums?  In just a few hours, Nvidia suffered the biggest loss of value in the history of the financial markets. Panic set in across the industry. Microsoft, Broadcom, Taiwan Semiconductor Manufacturing Company, all the companies involved in AI fell, losing over a trillion dollars.

It was as if the market realized, in a split second, that the rules of the AI race were being rewritten in real time.  

The geopolitical nightmare

The shock was not just financial, it was geopolitical.  For years, the United States operated under the premise that dominance in AI belonged to them.  The best chips. The most talented engineers. The most powerful companies.  China was making progress, but the prevailing belief was that America still had a comfortable lead.  

Now? This lead seems either fragile or non-existent. But the concerns don't stop at the economy. Multiple sources in the cyber industry have raised red flags about possible data transfers to China through DeepSeek.  

Cybersecurity researchers and US government analysts have warned that the model could collect and send sensitive information to servers under the control of the Chinese state.  If these suspicions are confirmed, DeepSeek could represent not only an economic threat, but also a strategic and national security one.  

At the World Economic Forum in Davos, where global tech leaders met just days before, the tone changed. Alphabet chairman Ruth Porat acknowledged that the US may have a lead of "about a year," but warned that "it's not at all certain" that this will remain the case. President Donald Trump reacted immediately, calling the event "a wake-up call" for the American AI industry.  

"We need to find faster and cheaper ways to develop AI," he said, advocating for the removal of Biden-era restrictions on AI development in the US.  

At the same time, DeepSeek came under a wave of suspicion. A Forbes report revealed that the model refused to answer sensitive questions about the Chinese government (such as the Tiananmen Square protests), but provided extensive details about criticism of Joe Biden and Donald Trump. Is this just a new chapter in the AI race - where control over information is as important as computing power?  

What's next?

DeepSeek has opened a deep wound in the AI industry.  If its method is sustainable, brute-force AI - based on billions of dollars invested in chips - could become an outdated model.  

This raises existential questions:

What will happen to Nvidia, which has built its entire strategy on the idea that progress in AI requires more GPUs?  

What does this mean for OpenAI, Google and Microsoft, which have spent tens of billions believing that computing power is key?  

What does it mean for China and the US, which are now competing not just for AI dominance, but for AI philosophy?  

The AI revolution was never meant to be comfortable. It was always about disruption. But no one would have predicted that the biggest AI earthquake would come not from OpenAI, Google or Meta, but from DeepSeek.  

This event is a game-changer. It challenges the assumptions we've had about AI development and the role of big tech companies. It also highlights the growing competition between the US and China in the field of AI. I'm curious to see how this will play out in the coming years.

And the world of AI will never be the same.  

You can find the original article here 

‍

Alexandru Dan

CEO, TVL Tech

‍