It is difficult to overstate the significance of the rise of artificial intelligence for capital markets. The massive scale of current AI investments and the rapid pace of technological progress raise questions and increase stakes for nearly all investors, companies, and countries as they try to anticipate what the future holds. According to Brent Thill, the Tech Sector Leader of Software and Internet Research at Jefferies, the story of AI is just beginning, although certain long-term trends are becoming apparent.
On November 11, Brent and his team host several hundred senior leaders at the Jefferies AI and Capital Market Summit. There, they will discuss the most significant trends shaping AI development and their implications for various industries and markets. Ahead of the event, we spoke with Brent to get his perspective on a few key questions likely on attendees’ minds.
Q: Some recent media coverage has suggested that a lot of major corporate AI investments have yet to turn a significant profit. What’s your take on this narrative?
| Contrary to what you hear in some quarters, AI is making money today on multiple fronts. The backlog of contract signings is $700 billion across the tech vendors we watch in the hyperscale market. That’s over 200% greater than the Capex increase. So, revenue exceeds the investment. Over the last two years, Microsoft has shown that it has expanded operating margins while investing in AI. So, when you look at both revenue-to-costs and margin, we continue to believe there’s pricing power and that the economic output of AI will exceed its costs. |
| The long-term ROI case study for AI is yet to be seen. We don’t believe you’re seeing significant jobs reduced or eliminated at this point. AI is effectively augmenting rather than replacing today. Long-term, there will no doubt be job loss. The question is, what’s the magnitude of that loss? What are the most notable trends you have seen unfold in the last six months to a year? The number one trend is investment in infrastructure for AI. The other things, like applications, are not selling as robustly as energy, land, data centers, power, water cooling and all the other things that go into enabling AI to run. That’s what’s booming right now. The applications and the number of people inside using AI are yet to come, which helps explain why many application stocks have lagged materially. Where is the lion’s share of the investments going? For AI to work, you must have three things: users, data, and capital. You can count on two hands the number of companies that have those things: Microsoft, Amazon, Oracle, Google, Nvidia are the big ones of course. That is why we believe, as an investment thesis, that you want to stay long with the biggest companies that are leading investment in this space. The number one constraint is energy. So, energy is a big winner. Some energy stocks are up 50% to 500%. No one would have seen that coming five years ago. There will also be specialized vendors that benefit, such as CoreWeave and Nebius, which handle data center infrastructure, along with chip companies like Avago and Broadcom. However, for now, this remains a very narrow trade. What are the risks and concerns we should have about the AI story today? The risk is the ROI. We’re putting such a big investment into infrastructure. It’s the biggest investment we will see in our lives. The big question is, can AI do what everyone says it can do? And when are we going to see the adoption path inside the enterprise? Today, that process is early and the path to revenue is uncertain. Nonetheless, we expect substantial returns from AI investments, and we will see them in the near future. There are so many AI tools available to us, but many users aren’t trained. Enterprise still must go through the crawling and walking stages before this thing really takes off. Then, data governance and security are critical issues for AI. One of the biggest challenges we hear about is that the adoption of these technologies can’t succeed if you’re using poor data, because that leads to bad results. However, these are solvable problems. They also present great opportunities for companies that can provide effective data security solutions. What are the major things you see happening with AI in the next two years? Brent: What we will see in the next couple of years is an AI shakeout. We will go from the tryout phase to the implementation phase. I think many of the systems we are just seeing now will be used in a big way in 2026 and 2027.And that’s when we start to see bigger ROI. That’s when we see more capital allocated. Another thing you should keep in mind is that during past major technological shifts – such as the arrival of the internet or the shift from PCs to mobile devices – entire boards of directors did not rush to demand the adoption of these new technologies. Today, however, boards are frenetically pushing for the adoption of AI because of its potential future impact. There has really been nothing like this in the last 30 years. |
