https://www.cwbwealth.com/en/news-and-stories/insights/deepseek-disruption
Tariffs and trade wars may dominate headlines today, but just a few short weeks ago another topic “du jour" rocked markets. Little known artificial intelligence (AI) company DeepSeek, made a groundbreaking announcement that wiped out over a trillion dollars in global market capitalization within the tech industry in a single day, sparking panic among investors trying to discern what’s noise and what’s real.
While the market’s initial reaction might seem extreme, it reflects a broader reality: the AI sector is inherently volatile. It’s driven by rapid change, innovation and shifting narratives that can often be polarizing and politicized. It’s an industry that’s evolving at breakneck speed, with trendy headlines designed to make us click. But we should look beyond the headlines to evaluate the environment that’s shaping it.
In light of these exceptional returns, and after billions spent in investing in AI chips and infrastructure, investors began worrying that we are up for a breather in the “AI party”. So, when the DeepSeek announcement broke at the end of January, the market response was swift and brutal. It rippled across the entire AI value chain – from semiconductors and data centres to power producers and software companies.
DeepSeek, the Chinese tech startup, revealed a transformative AI model capable of solving complex tasks with unprecedented efficiency—notably at a fraction of the cost of existing AI models. Why?
DeepSeek has trained its latest model on lower-tier GPUs (specialized chips known as graphics processing units) due to U.S. export restrictions on advanced NVIDIA chips, which has dramatically reduced costs. While these claims are still under scrutiny, costs of training AI models may be poised for a significant decline.
Interestingly, and somewhat counterintuitively, greater efficiency in AI may in fact increase demand for AI infrastructure (a phenomenon known as Jevon’s Paradox). When technology becomes cheaper and more efficient, its use often expands instead of contracting. In other words, as AI becomes faster and cheaper to deploy, more businesses will adopt it and likely use it more extensively, driving demand for chips, power and infrastructure.
For internet and software companies, the ones building the end-user applications, falling computer costs is welcome news as they should accelerate AI adoption and, ultimately, translate to higher returns.
So how do we deal with it? Just as we need to step back from the mirror to see ourselves clearly, we distance ourselves from the emotions often evoked by noise in the market and focus on what really matters: maintaining a long-term perspective and adhering to a disciplined investment process. This holds true whether we’re evaluating NVIDIA, JPMorgan or Costco.
So, is the AI party over—or is it just getting started? History suggests the latter.
Technological revolutions—like the advent of the internet, mobile internet adoption, and cloud computing—have all followed a pattern of investors grossly underestimating the scale and longevity of the cycle (AKA “the party”). Why does this happen? The answer lies partly in how humans perceive growth.
We naturally perceive growth, as gradual and linear. But technological advancements often grow exponentially, especially in their early stages. And there’s a big difference. Imagine taking 30 linear steps—you’d cover just 30 meters. Now imagine that each step doubles in size, creating exponential growth. By the 30th step, you’d have travelled over a billion meters—enough to circle the Earth 26 times! (See figure 1.)
Once you grasp this concept, it’s easier to understand why analysts have taken so long to revise NVIDIA's 2026 earnings estimates to where they are today—a whopping 642% higher than the original estimate two years ago!
Figure 1: The power of exponential growth
Source: CWB Wealth
Only time will tell how the AI industry will evolve. Who will be the winners and who will be the losers. But as with any revolutionary technology in the past, it won’t be a smooth ride. Brace for both risks and rewards.
Sources: Morgan Stanley, Bernstein, DeepSeek, Chat GPT
CWB Wealth Management Ltd. (CWB Wealth) is a wholly-owned subsidiary of National Bank of Canada. Information presented herein is for discussion and illustrative purposes only and is not a recommendation or an offer or solicitation to buy or sell any securities. Views expressed are as of the date indicated, based on the information available at that time, and may change based on market and other conditions. Unless otherwise noted, the opinions provided are those of the authors and not necessarily those of CWB Wealth Management Ltd. (CWB Wealth) or its affiliates. CWB Wealth does not assume any duty to update any of the information. Investment decisions should be based on an individual’s own goals, time horizon, and tolerance for risk. Information herein is intended to provide qualified or accredited investors, or their advisors, with general information and is not, and under no circumstances is to be construed as, a prospectus or advertisement or a public offering of the securities of pooled funds offered by CWB Wealth. Any such offer or solicitation shall only be made at the time a qualified or accredited investor, in those jurisdictions where permitted by law, receives the Confidential Offering Memorandum, or other offering documents as applicable, relating to the respective pooled fund. Nothing in this content should be considered to be legal or tax advice and you are encouraged to consult your own lawyer, accountant or other advisor before making any financial decision. Quoted yields should not be construed as an amount an investor would receive from the Fund and are subject to change. Investors should consult their financial advisor before making a decision as to whether mutual funds are a suitable investment for them. Commissions, trailing commissions, management fees and expenses all may be associated with mutual fund investments. Please read the prospectus, which contains detailed investment information, before investing. Mutual funds are not guaranteed, their values change frequently, and past performance may not be repeated.
CWB Wealth uses third parties to provide certain data used to produce this report. We believe the data to be accurate, however, cannot guarantee its accuracy. Visit https://www.cwbwealth.com/disclosures for the full disclaimer.
While the market’s initial reaction might seem extreme, it reflects a broader reality: the AI sector is inherently volatile. It’s driven by rapid change, innovation and shifting narratives that can often be polarizing and politicized. It’s an industry that’s evolving at breakneck speed, with trendy headlines designed to make us click. But we should look beyond the headlines to evaluate the environment that’s shaping it.
AI: the belle of the ball
Since OpenAI’s release of ChatGPT in late 2022, AI has become one of the most dominant drivers of U.S. stock market returns. AI-focused stocks have not only outperformed broader markets but also delivered spectacular earnings growth. NVIDIA, widely recognized as the posterchild of the AI revolution, exemplifies this trend. In just two years, its stock has soared more than 1,000%. Other companies that were perceived as AI beneficiaries have also significantly outpaced the market, with investors betting on the AI revolution spurring a new era of productivity gains from enabling machines to simulate human intelligence and learning.In light of these exceptional returns, and after billions spent in investing in AI chips and infrastructure, investors began worrying that we are up for a breather in the “AI party”. So, when the DeepSeek announcement broke at the end of January, the market response was swift and brutal. It rippled across the entire AI value chain – from semiconductors and data centres to power producers and software companies.
DeepSeek, the Chinese tech startup, revealed a transformative AI model capable of solving complex tasks with unprecedented efficiency—notably at a fraction of the cost of existing AI models. Why?
DeepSeek has trained its latest model on lower-tier GPUs (specialized chips known as graphics processing units) due to U.S. export restrictions on advanced NVIDIA chips, which has dramatically reduced costs. While these claims are still under scrutiny, costs of training AI models may be poised for a significant decline.
AI efficiency could lead to more—not less—demand
Efficiency gains of this magnitude seem to have taken the market by surprise, but they’re not without precedent. In the semiconductor industry, the growth in processing power on a single chip has been nothing short of a miracle, consistently doubling every two years for more than five decades as Moore’s Law predicted it. This phenomenal efficiency gain helps explain why today’s smartphones outperform the supercomputers that once guided astronauts to the moon.Interestingly, and somewhat counterintuitively, greater efficiency in AI may in fact increase demand for AI infrastructure (a phenomenon known as Jevon’s Paradox). When technology becomes cheaper and more efficient, its use often expands instead of contracting. In other words, as AI becomes faster and cheaper to deploy, more businesses will adopt it and likely use it more extensively, driving demand for chips, power and infrastructure.
For internet and software companies, the ones building the end-user applications, falling computer costs is welcome news as they should accelerate AI adoption and, ultimately, translate to higher returns.
Market impact: what’s noise and what’s notable
While markets tend to be efficient over the long term, in the short term that’s not always the case. In the short term, fear and greed can temporarily distort the picture. Just like looking into a distorted mirror at an amusement park, they can stretch, shrink and wrap reality and distort investors’ perception of value and risk.So how do we deal with it? Just as we need to step back from the mirror to see ourselves clearly, we distance ourselves from the emotions often evoked by noise in the market and focus on what really matters: maintaining a long-term perspective and adhering to a disciplined investment process. This holds true whether we’re evaluating NVIDIA, JPMorgan or Costco.
So, is the AI party over—or is it just getting started? History suggests the latter.
Technological revolutions—like the advent of the internet, mobile internet adoption, and cloud computing—have all followed a pattern of investors grossly underestimating the scale and longevity of the cycle (AKA “the party”). Why does this happen? The answer lies partly in how humans perceive growth.
We naturally perceive growth, as gradual and linear. But technological advancements often grow exponentially, especially in their early stages. And there’s a big difference. Imagine taking 30 linear steps—you’d cover just 30 meters. Now imagine that each step doubles in size, creating exponential growth. By the 30th step, you’d have travelled over a billion meters—enough to circle the Earth 26 times! (See figure 1.)
Once you grasp this concept, it’s easier to understand why analysts have taken so long to revise NVIDIA's 2026 earnings estimates to where they are today—a whopping 642% higher than the original estimate two years ago!
Figure 1: The power of exponential growth
Only time will tell how the AI industry will evolve. Who will be the winners and who will be the losers. But as with any revolutionary technology in the past, it won’t be a smooth ride. Brace for both risks and rewards.
Sources: Morgan Stanley, Bernstein, DeepSeek, Chat GPT
CWB Wealth uses third parties to provide certain data used to produce this report. We believe the data to be accurate, however, cannot guarantee its accuracy. Visit https://www.cwbwealth.com/disclosures for the full disclaimer.