Why The DeepSeek AI Crash Was a Good Thing & Here’s Why Most Experts Are Wrong About AI — Zak Gardezy, CFP® Explains
The DeepSeek R1 model is undoubtedly impressive, and its development at a significantly lower cost than other major generative AI tools has sent shockwaves through the market. Companies such as Nvidia, Broadcom, and other major AI players rely on developers purchasing their high-priced hardware to build increasingly complex and high-quality AI systems. The revelation that a competitive AI model could be created at a fraction of the cost spent by OpenAI, Meta, and others has led investors to question whether expectations around continued demand for Nvidia’s chips are justified — raising concerns about whether Nvidia’s stock price is sustainable. This uncertainty resulted in a sharp 17% decline in Nvidia’s stock, dragging down other AI-related names.
At first glance, the outlook for AI seemed bleak. However, taking a step back, the launch of DeepSeek R1 is arguably the most significant event in AI since the introduction of ChatGPT — and it’s a development that should excite, rather than worry, long-term investors.
Competition Drives Progress
As an avid user of ChatGPT, I’ll be the first to admit that the tool was starting to feel a bit stale. My requests have become increasingly complex, and as OpenAI’s models have attracted millions of users, I’ve noticed performance inconsistencies. For AI to fully integrate into daily life, models must become faster, more accurate, and more efficient.
DeepSeek R1 is trained on a larger dataset than GPT-4, is multimodal, can continuously learn, and according to Dev.to, it demonstrates significantly higher accuracy. If I were an OpenAI executive, this development would light a fire under me.
Of course, OpenAI is already rumored to be working on GPT-5, which could leapfrog everything else on the market. In that case, DeepSeek’s dominance may be short-lived. However, as I discussed on a Money Minutes episode on April 30, 2024, AI development is rapidly becoming a geopolitical arms race — similar to the Space Race.
World powers are backing AI development, seeking to lead the world in this transformative technology. We’ve already seen this with the Trump administration’s Stargate Project and reports that DeepSeek has financial ties to the Chinese Communist Party (CCP). I expect this trend to intensify, with nations competing aggressively until one establishes undisputed dominance — similar to how the Cold War ended with the U.S. as the sole superpower.
With China emerging as a global power, rivaling (or approaching) the United States’ economic and technological strength, we are witnessing the return of a Cold War-style competition, particularly in AI innovation.
China’s AI Will Struggle to Gain Western Adoption
While consumer adoption of generative AI is increasing, commercial applications will ultimately be more impactful. That’s why I’m not overly concerned about DeepSeek or Chinese AI models dominating globally.
The world remains highly skeptical of the CCP — just look at what happened to TikTok in the U.S.. It is highly unlikely that financial institutions, law firms, medical organizations, intelligence agencies, and military sectors will trust sensitive data to Chinese AI servers, no matter how much cheaper they are.
While China, Russia, and politically aligned nations may use these models extensively, Western capital still dominates global markets. The AI race isn’t just about building the best models — it’s about who controls the capital that funds AI development. The U.S. maintains an edge in data security and trust, which is crucial for retaining leadership.
Wall Street’s Shortsighted View on AI Stocks
Some Wall Street analysts are downgrading top AI stocks following DeepSeek’s launch. In my view, this reaction is shortsighted. Here’s why:
32% of American adults reported using generative AI tools in the past week, with 26% for personal tasks and 24% for work-related tasks.
A survey found that 68% of respondents use AI at work, but only 28% utilize it extensively.
Globally, about 500 million people actively use generative AI — roughly 6.18% of the world population.
93% of Fortune 500 companies have begun integrating AI into their operations.
A Microsoft & LinkedIn study found that 75% of knowledge workers are using AI.
Interpreting the AI Data
Despite growth, AI adoption is far from maturity — 68% of American adults either don’t use AI or haven’t reported it, meaning deployment is still restricted by regulations and corporate policies.
AI is underutilized in commercial settings — only 28% of workplace users employ AI extensively, revealing massive room for growth.
Generative AI penetration is low — roughly 6.18% of the global population uses AI, compared to 37.9% of people actively using Facebook and 70% of the world accessing the internet. This suggests AI is still in the early adoption phase.
Fortune 500 integration is still in its infancy — while 93% of major corporations are exploring AI, real adoption remains limited to pilot programs and specific business segments. Most professionals I speak with say their companies restrict AI tools like ChatGPT from being used for emails, presentations, and research.
Knowledge worker usage is promising but skewed — high AI adoption among knowledge workers is encouraging, but these surveys are likely biased toward those already familiar with AI, underestimating the broader workforce’s usage rates.
The Future of AI and Next-Gen Innovations
Given these trends, we are many years away from reaching peak AI utilization — and by that time, AI will have evolved significantly. The next major wave will likely include:
Autonomous AI Agents & Co-Pilots — AI that executes entire workflows with minimal human oversight.
Multimodal AI — Systems that seamlessly handle text, images, video, and voice simultaneously.
Quantum AI — AI running on quantum computing, exponentially improving processing speeds.
Brain-Computer Interfaces (BCIs) — Technologies like Neuralink, enabling direct brain-to-computer communication.
Analysts downgrading Nvidia, AMD, Intel, Broadcom, and other AI stocks due to one new model release appear to be hedging their bets — not making sound long-term assessments.
I don’t know any rational investment analyst who truly believes that one model’s launch spells doom for Nvidia or that the U.S. will surrender in the AI race. Instead, DeepSeek’s advancements will only push Nvidia, AMD, Intel, Broadcom, and others to accelerate development of faster, cheaper, and more efficient AI chips.
In turn, this will fuel breakthroughs at OpenAI, Meta, Amazon, Alphabet, and Anthropic, reinforcing America’s dominance in AI.
If anything, I’m more bullish on AI today than I was before.
Disclaimer: The content of this article is the opinion of Zak Gardezy, CFP® and should not be misconstrued as investment, legal or tax advice. Please consult with a financial advisor prior to making any investment, tax or legal decisions.