The stock market is a theater of extremes. One day, NVIDIA is the darling of Wall Street, the next, it’s a doomed relic of the AI bubble. One day, DeepSeek is a niche open-source model, the next, it’s the harbinger of doom for OpenAI and the entire American AI industry.

As investors, we’re constantly bombarded with narratives that swing between euphoria and despair. But what if the truth lies somewhere in between? What if the market’s pendulum swings are just noise, obscuring the signal of long-term trends?

As an avid AI enthusiast who has extensively used both DeepSeek and ChatGPT, I wanted to share my thoughts with my audience. Let’s take a deep dive into the DeepSeek panic, NVIDIA’s valuation, and the broader implications for AI investors. Along the way, we’ll explore why the future of AI might not be as binary as the headlines suggest.

The DeepSeek Panic: Much Ado About Nothing?

The Narrative

Yesterday, the market was gripped by a sudden panic over DeepSeek, an open-source large language model (LLM) developed in China. Headlines screamed about the end of OpenAI, the collapse of NVIDIA, and the death of American AI dominance. But here’s the thing: DeepSeek has been around for over a month. It’s been discussed on YouTube, CNBC, and other platforms. So why the sudden hysteria?

The Reality

DeepSeek is a capable model, but it’s not a ChatGPT killer. Yes, it’s 38 times cheaper to use than OpenAI’s flagship model, making it an attractive option for startups and cost-conscious users. However, it falls short in areas requiring deep domain expertise or extensive global knowledge. For example, when asked about the naming origin of Mumbai’s Cuffe Parade district, DeepSeek provided an incorrect historical account, while ChatGPT nailed the answer.

This isn’t to say DeepSeek is useless. It excels at processing proprietary datasets and performing specific tasks using that data, making it a valuable tool for certain workflows. But it’s not a one-size-fits-all solution. No large-language model (LLM) is.

The Implications

The panic over DeepSeek reflects a broader misunderstanding of the AI market. Just as Android didn’t kill the iPhone, cheaper models like DeepSeek won’t eliminate premium offerings like ChatGPT. Instead, they’ll carve out their own niches, serving different use cases and customer segments.

For example, a startup using AI in the background to power its product might switch to DeepSeek to save costs. However, enterprise customers served by Microsoft, which has partnered with OpenAI, are unlikely to abandon using a premium, reliable model deeply integrated into Microsoft enterprise products for a cheaper alternative. Enterprise customers care about productivity and reliability, and they’re willing to pay more to ensure they’re getting the best results. Mistakes in critical fields like medicine or engineering could be costly, so premium users will happily pay premium pricing when “good enough” isn’t good enough.

NVIDIA: The GPU King Under Pressure

NVIDIA’s stock has been on a tear, fueled by its dominance in the AI hardware market. But critics (myself included) argue that the company is overvalued, with its valuation disconnected from fundamentals. I wrote about this in Uncertainty & The Icarus Print.

The Rethink

The market panic is about a fundamental rethink: the assumption that companies need to spend billions of dollars on expensive GPUs to build AI models has been shattered. DeepSeek demonstrates that capable models can be built for a fraction of the cost and on lower-end hardware.

This doesn’t mean NVIDIA is doomed. Yes, the company’s growth assumptions are under scrutiny. If cheaper hardware and more efficient models become the norm, NVIDIA’s pricing power and margins could come under pressure at the top. But a new market at the bottom could be unlocked, and they could make it up in volume.

McDonald's sells cheap hamburgers to the masses, yet makes billions of dollars a year.

The Case for NVIDIA

NVIDIA’s GPUs are still the gold standard for AI workloads, particularly in high-performance and enterprise environments. While DeepSeek and other models might reduce the need for expensive GPUs in some applications, there will always be a demand for premium hardware in complex, mission-critical tasks.

Moreover, NVIDIA is constantly innovating. The company isn’t just a hardware provider—it’s a leader in AI software and ecosystem development. If anyone can adapt to a changing market, it’s NVIDIA.

The Auto Industry Analogy: Lessons for AI

The Doomsday Fear in the 1980s

In the 1980s, the American auto industry faced a similar doomsday fear. Japanese car manufacturers, led by companies like Toyota and Honda, started flooding the U.S. market with cheaper, higher-quality vehicles. The American auto industry was bloated, inefficient, and producing inferior products. Many predicted the demise of U.S. automakers.

What Actually Happened

Rather than being destroyed, American automakers were spurred to compete. They learned from their Japanese counterparts, adopting lean manufacturing techniques, improving quality, and streamlining operations. Companies like Ford and General Motors didn’t disappear—they adapted and became stronger for it.

The Parallel to AI

The same dynamic will likely play out in the AI market. DeepSeek and other open-source models are like the Japanese cars of the 1980s—cheaper and good enough for many use cases. But instead of destroying established players like OpenAI and NVIDIA, these competitors are forcing them to innovate and improve.

For example, OpenAI will learn new methods from their competitors that will help accelerate its efforts. ChatGPT will become more efficient or gain expanded capabilities to stay ahead of the curve.

When Less Demand Creates More Demand

The paradox of efficiency, often referred to as Jevons Paradox, will play an interesting role in the AI landscape. Simply put, as technologies become more efficient, people don’t use less of them—they use more.

For example, when cars became more fuel-efficient, people didn’t consume less oil; instead they bought bigger cars or SUVs and drove longer distances. Similarly, as AI models become cheaper to train and run, companies won’t reduce their AI investments—they’ll likely expand them.

Cheaper models like DeepSeek could democratize AI, making it accessible to startups and smaller players, while simultaneously driving demand for more GPUs – if not in the sum but in the aggregate.

Conclusion

The market’s pendulum swings between euphoria and despair, but the truth is rarely so binary. DeepSeek isn’t the end of OpenAI, and NVIDIA isn’t doomed to collapse. Instead, the AI market is evolving, creating new opportunities and challenges for investors.

Key Takeaways

  1. DeepSeek is a capable model, but it’s not a ChatGPT killer. It excels in specific use cases but falls short in others.
  2. NVIDIA is overvalued, but it’s not going away. The company will need to adapt to a changing market, where cheaper hardware solutions become more common.
  3. Competition drives innovation. Just as the auto industry adapted to Japanese competition in the 1980s, AI incumbents will innovate to stay ahead.
  4. The AI market is still growing. New models like DeepSeek will expand the market, rather than replacing existing players.
  5. Premium vs. budget segmentation. Startups may embrace cheaper models, but enterprises will continue to pay a premium for reliability and performance.

So the next time the market panics—or rallies—remember this quote:


When everyone is euphoric – it's usually not as good as it seems. When everyone is screaming doom – it's usually not as bad as it seems.