He also warns that the growing use of debt to finance AI investments could become a concern. While debt is appropriate for infrastructure such as data centres, Lee says breakthrough innovation should be backed by equity capital, as true technological advances require investors willing to take bigger risks rather than settle for incremental improvements.
This is an edited transcript of the interview.Q: South Korea, Taiwan and the US remain tightly correlated because of AI. If investors really want diversification, should they look at India, or is India not that exciting right now?A: If you really want diversification, you have to step out of the one sector tying all these markets together, which is the AI semiconductor world. Unfortunately, Asia is very tied into that sector.
If you look at a diversified market like the United States, companies such as Lilly are making new highs because investors are anticipating new GLP-1 drugs that could address many health problems in the US and globally. So, diversification has to be thought about broadly across sectors and products.
I also want to remind everybody what Warren Buffett said about diversification. He said diversification can worsen portfolio performance because you take money away from what you think is your best idea and spread it across a basket of ideas.
For most investors, the core portfolio should certainly be diversified. But for more speculative investments, where you are looking for outsized returns, you really have to back your best ideas. Right now, the AI trade still captures most of that imagination.
Q: We are seeing sharp moves in AI-linked stocks. South Korea’s market has been under pressure and ASML fell after reports of a Chinese company developing competing chipmaking equipment. How do you see that?
A: I was fascinated because the news that sparked the sell-off was about a Chinese company that may be producing a machine to compete with ASML.
Let me put it this way. If you needed surgery and had the choice between the best surgeon in the world and someone who had just finished residency, who would you choose? Clearly, you would choose the best surgeon if you could. Only if that surgeon wasn’t available would you go with the other option.
Similarly, this company may become a challenge to ASML in the future. I don’t put it past the Chinese to move very quickly because they learn fast and acquire intellectual property at an impressive pace. But I still think they are more of a future challenge than an immediate threat.
More importantly, this is a good lesson for the AI trade. AI has become less of a macro story about improving productivity or being disinflationary and more of a micro story about which companies can actually create products that improve their bottom line.
That’s why many software companies are making new highs. The opportunity is increasingly at the application level, where AI is used to improve productivity or help companies develop the next generation of products, whether that’s new GLP-1 drugs or future cancer treatments.
I believe AI will remain at the fulcrum of future profits. But investment opportunities will become much more diverse and increasingly focused on applications.
Q: Have we already seen the best valuations for AI companies?
A: Yes, I think we have seen the best valuations for this phase because our imagination is still limited when it comes to AI applications.
The next phase of higher valuations will likely come from the areas where Jensen Huang is investing—across the best application ideas for Nvidia’s technology.
That’s where investors should keep their focus. Will the winners come from open models, closed models or both? More importantly, who will be using these technologies?
I have a lot of confidence in companies like Palantir, which work directly with businesses to help them use their data more effectively. Those are the kinds of companies that will make the most of AI technologies.
Q: There are concerns that circular financing deals are helping fuel the AI boom. Does that worry you?
A: Very much so. What worries me is that, in the past, if investments were funded through cash flow and equity, equity investors were willing to swing for the fences. They were prepared to make high-risk, high-reward bets.
But when financing shifts to debt, it’s a different story. During the mergers and acquisitions era of the 1980s, Professor Michael Jensen argued that higher debt burdens forced management teams to become more efficient because they had to service that debt or risk failure.
The problem is that this encourages companies to pursue safer, more reliable technologies rather than take the risks needed for breakthrough innovation.
We are still at a stage in the AI cycle where we need to swing for the fences. Nine out of ten ideas may fail, but the one success could change everything.
Watch the full conversation here
That’s why equity financing is so important. It encourages the kind of research and risk-taking needed for major breakthroughs. Debt financing, on the other hand, worries me because it pushes companies towards mediocrity.
To be clear, much of the debt financing today is going towards infrastructure such as data centres and other proven technologies. That’s perfectly legitimate. But when it comes to funding frontier innovation, I believe companies should be looking for equity investors because those are the people willing to take the big risks.
Catch all the latest updates from the stock market here
