I had originally planned to explore the latest AI certifications launched by Dubai and Saudi Arabia, which aim to legitimise AI companies operating in their regions. But then DeepSeek happened, and the conversation shifted overnight.
by Kasun Illankoon, Editor-in-Chief at Tech Revolt
DeepSeek, a Chinese AI startup, shook global markets with a single announcement, sending tech firms, chipmakers, and investors into a frenzy. The company unveiled a game-changing AI model developed for under US$6 million—a stark contrast to the hundreds of millions typically required for AI development. The fallout was swift: Nvidia lost 17% of its market cap, erasing nearly US$600 billion in value, while the Nasdaq suffered its steepest drop since December 18. Even Microsoft, Alphabet, and AMD saw their stock prices falter as market confidence wavered.
Tech stocks tentatively recovered after President Donald Trump described the launch of a chatbot by China’s DeepSeek as a “wake-up call” for Silicon Valley in the global race to dominate artificial intelligence.
What’s the Fuss About?
DeepSeek’s achievement is not just about cost. The AI model operates on lower-cost chips and uses less data, challenging the long-held belief that bigger investments in GPUs and data centres are the future of AI. This raises serious questions about the billions being poured into AI infrastructure by US tech giants. However, innovation doesn’t stop with one breakthrough. DeepSeek may have disrupted the market, but it hasn’t rewritten the playbook entirely.
Sam Altman of OpenAI described the development as “an impressive achievement, particularly for its cost, but it’s just one step in AI’s evolution.”
Nvidia echoed a similar sentiment, stating: “DeepSeek is an excellent AI advancement and a perfect example of Test Time Scaling. DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant. Inference requires significant numbers of NVIDIA GPUs and high-performance networking. We now have three scaling laws: pre-training and post-training, which continue, and new test-time scaling.”
Nvidia, for example, remains indispensable. DeepSeek’s model relies on “test-time scaling,” a process that optimises pre-trained models during use, which still requires significant GPU power for inference and high-performance networking. This dynamic highlights Jevons’ Paradox: as AI becomes more efficient and affordable, demand for GPUs and infrastructure may actually increase. DeepSeek may accelerate the adoption of AI globally, creating opportunities for both startups and established players alike.
The Bigger Picture
DeepSeek’s rise isn’t just a tech story; it’s a market evolution. By lowering the barrier to entry, the company has set the stage for an influx of competitors, forcing established players to adapt. This disruption, while painful for some, fuels the industry’s overall growth and ensures innovation continues. Silicon Valley’s dominance isn’t ending—it’s being recalibrated.
Of course, the road ahead won’t be smooth. The frenzy surrounding DeepSeek triggered a cyberattack on the company, forcing it to temporarily halt user registrations. Such vulnerabilities underscore the risks that come with rapid disruption in an increasingly competitive landscape.
The Wake-Up Call
DeepSeek’s success is a reminder that competition is the lifeblood of progress. While it’s easy to panic over the immediate market fallout, it’s worth remembering that disruption often leads to stronger, more resilient industries. US tech giants may need to rethink their strategies, but their ability to innovate remains unmatched. Nvidia, Microsoft, and others will likely pivot, adapting their approaches to meet the new challenges and opportunities this shakeup presents.
For now, the question isn’t whether DeepSeek will disrupt—it already has. The real question is: how will the incumbents respond?