Former Intel CEO Pat Gelsinger has publicly challenged Wall Street’s reaction to DeepSeek’s breakthrough AI model, arguing that investors are fundamentally misunderstanding the implications for the semiconductor industry. The comments come after DeepSeek, a Chinese AI startup, triggered a massive sell-off in AI stocks on Monday, January 27, 2025, when it released its R1 model that claims to match OpenAI’s leading capabilities at a fraction of the cost.
The market panic was swift and severe. Nvidia alone lost up to $500 billion in market capitalization in a single day, contributing to the biggest stock market rout in US history. Other top AI companies saw hundreds of billions wiped from their valuations as investors feared that DeepSeek’s efficiency breakthrough would dramatically reduce demand for advanced AI chips.
DeepSeek’s R1 model, released on President Trump’s Inauguration Day, appeared to achieve performance comparable to OpenAI’s latest models while using fewer and less powerful chips than those typically deployed by American AI labs. This development raised immediate concerns about whether companies like Nvidia would see their booming AI chip business contract.
However, Gelsinger took to X (formerly Twitter) to argue that the market has it backwards. He invoked what he called “the gas law” of computing, suggesting that making computing “dramatically cheaper” and more efficient doesn’t shrink the market—it expands it. Rather than reducing chip demand, Gelsinger believes DeepSeek’s efficiency gains will actually increase the total addressable market for AI computing.
The former Intel chief, who retired in December 2024 after struggling to position Intel competitively during the AI boom, also noted that DeepSeek’s engineers “had limited resources, and they had to find creative solutions” to maximize performance. This constraint was largely due to tough US export controls that have prevented Chinese companies from accessing America’s most advanced chips.
Importantly, DeepSeek has not fully disclosed the computing power behind its R1 model, leaving some uncertainty about the true efficiency gains. Gelsinger’s perspective aligns with other AI industry voices, including Wharton professor Ethan Mollick, who questioned why more efficient models would reduce compute value, arguing that efficiency enables companies to “serve more customers and products at lower prices & power impact.”
The debate highlights a fundamental question facing the AI industry: whether efficiency gains represent a threat to chip demand or an opportunity to democratize and expand AI adoption across new markets and use cases.
Key Quotes
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…
Former Intel CEO Pat Gelsinger posted this on X to challenge the market’s negative reaction to DeepSeek. He argues that efficiency gains will grow rather than shrink the AI chip market, drawing on historical computing trends.
had limited resources, and they had to find creative solutions
Gelsinger’s observation about DeepSeek’s Chinese engineers explains how US export controls on advanced chips may have paradoxically driven innovation in AI efficiency, forcing creative problem-solving under resource constraints.
More efficient models mean that those with compute will still be able to use it to serve more customers and products at lower prices & power impact.
Wharton professor Ethan Mollick echoed Gelsinger’s perspective, arguing that efficiency doesn’t eliminate the value of computing power but rather enables broader deployment and accessibility of AI services.
Our Take
Gelsinger’s intervention is particularly noteworthy given Intel’s struggles to capitalize on the AI boom under his leadership. His “gas law” analogy has historical merit—cheaper computing has consistently expanded markets rather than contracted them, from mainframes to PCs to cloud services. However, the semiconductor industry faces a unique challenge: if efficiency gains come primarily from algorithmic innovation rather than hardware advances, the value capture shifts from chipmakers to AI developers. The real question isn’t whether AI demand will grow, but whether that growth translates to chip sales or simply more efficient utilization of existing hardware. DeepSeek’s breakthrough may signal a transition from the “scaling era” of AI—where bigger models and more chips dominated—to an “efficiency era” where algorithmic innovation matters more. This could fundamentally reshape competitive dynamics, potentially benefiting software-focused companies while pressuring pure hardware plays. The market’s violent reaction suggests deep uncertainty about this transition.
Why This Matters
This story represents a critical inflection point for the AI industry and semiconductor sector, with profound implications for investors, tech companies, and the global AI race. The DeepSeek development challenges the prevailing assumption that AI progress requires exponentially increasing computing power and chip investments—a thesis that has driven trillions in market capitalization gains for companies like Nvidia.
If DeepSeek’s efficiency claims prove accurate, it could democratize AI development by lowering barriers to entry, enabling smaller companies and countries to compete with well-funded American AI labs. This has significant geopolitical implications, particularly as the US attempts to maintain technological leadership through export controls on advanced chips.
The debate between Gelsinger’s “expanding market” thesis and Wall Street’s “demand destruction” fears will shape investment strategies and corporate planning across the tech sector. For businesses, more efficient AI could mean lower operational costs and broader adoption. For workers, it could accelerate AI integration across industries. The resolution of this debate will determine whether the AI boom continues its current trajectory or enters a new phase characterized by efficiency over raw computing power, fundamentally reshaping the competitive landscape of the technology industry.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- Jensen Huang: TSMC Helped Fix Design Flaw with Nvidia’s Blackwell AI Chip
- OpenAI’s Valuation Soars as AI Race Heats Up
- Wall Street Asks Big Tech: Will AI Ever Make Money?
- The AI Hype Cycle: Reality Check and Future Expectations
- Legendary Investor Jeremy Grantham Predicts Stock Market Crash and AI Bubble Burst by 2024
Source: https://www.businessinsider.com/intel-pat-gelsinger-deepseek-market-wrong-nvidia-2025-1