November 21, 2025

Key takeaways
- There is a rising risk that AI company enthusiasm may be outpacing the return on investment for these technologies
- AI revenue pales in comparison to the investment being made and valuations are high
- AI bellwether, NVIDIA’s earnings release and forecast were strong but didn’t alleviate all concern, we expect more volatility
In Shakespeare’s Hamlet, the main character wrestles with whether to continue enduring life’s uncertainties (“to be”), or to end his suffering (“not to be”). Recently, the big uncertainty for market participants has
been whether we are in an Artificial Intelligence (AI) bubble or not. Should they stay invested and ride the euphoria, or step aside, fearing the consequences of overvaluation and potential collapse?
While we are concerned about rising levels of concentration in the AI theme, we are not calling this a “bubble” just yet. We have been overweight AI for the past few years and have benefited from this position. At present, we are modestly
trimming our positions in light of rising risks.
AI bellwether reports earnings
On Wednesday, NVIDIA — the largest AI company and the largest stock in the world — released its earnings report. Investors were laser focused on this release as it serves as a key indicator of the health of the AI industry. The company delivered
a surprisingly strong revenue forecast, pushing back on the idea that the AI sector is in a bubble. However, while this initially eased concerns, it fell short of fully alleviating them.
What could go wrong?
While AI companies like NVIDIA have very strong earnings, the current price that investors pay for those earnings is high, and valuations risk falling amid rising concerns about the future growth of these companies. The other big fear is that the direct
revenue being generated by AI pales in comparison to the hundreds of billions of dollars being spent on AI infrastructure by hyperscalers like Microsoft, Amazon, Alphabet and Meta. Technological advancement is always led by spending, however, the
more dollars that go into the technology, the more revenue investors want to see to believe they will receive a good return on the investment. In the chart below we compare revenues of top United States (US) AI companies with their capital expenditures
(capex), or investment spending.

Getting through with a long-term view
Our global equity team continues to monitor these dynamics and actively manage our AI positions, which includes overweight positions in each of the above-mentioned companies. Our long-term view is that we are in the early innings of the greatest technological
evolution of our lifetime. In the near term, we are monitoring two key themes:
- Supply and demand for chips: We anticipate sustained demand for AI graphics processing units (GPUs) continuing to exceed supply through 2026.
- Data center buildout to support AI growth: The ability of AI hyperscalers to monetize their capabilities remains unproven. Further, spending on data centers is increasingly funded by debt, raising concerns about a potential liquidity
crunch if market conditions change.
Longer-term, we are monitoring enterprise adoption as this is an indicator of potential productivity gains and future AI revenue. We are also monitoring hyperscalers’ progress towards Artificial General Intelligence (AGI), which would potentially
enable AI to replace humans in certain tasks. This could significantly raise AI-based revenues while lowering company costs.
We remain positive about the next 12 months, with AI compute demand continuing to exceed supply. If we are in a bubble, the big selloff is typically preceded by other large selloffs, which we haven’t seen yet.
“To be or not to be” was Hamlet’s struggle with uncertainty. Much like the AI predicament facing investors today, the answer — for now — lies between conviction and caution. Bubble or no bubble, this story is still being
written. And the ending will depend not on hype, but on earnings, innovation, and time.