Janus Henderson Applies Dot-Com Lessons to AI Boom

Janus Henderson is applying lessons from the dot-com era to the current AI boom, according to co-portfolio manager Joshua Cummings. He argues that the real winners may lie outside the data center, a perspective shaped by his 30 years in equity research and experience during the dot-com bubble.

Cummings, who co-manages the firm’s $45 billion research funds, recalls the dot-com era when financial media focused on about ten stocks. Investors who bought those popular names would have performed poorly since, despite the internet’s real potential. He said an investor who had bought the ten most popular names of that period would have done very poorly since, even though the attention was pointed at something real.

AI’s potential and market concentration

“We knew there was this thing that was going to be really, really big, the internet. We didn’t really know what we were going to do with it, but it ended up being exponentially more important and bigger than anybody could have conceived of, even the people who were right in the middle of it, like me at the time. I’m using that as the mental model for AI.” The Nasdaq Composite peaked at 5,048.62 on 10 March 2000 and bottomed at 1,114.11 in October 2002, a fall of 78 per cent. The S&P 500 fell 49 per cent over the same period.

Today’s market is more concentrated, with the top ten companies making up 39.1% of the S&P 500 in September 2026, compared to 27% at the dot-com peak. Alphabet, Amazon, Microsoft, and Meta have increased their combined capital expenditure plans to $725 billion for 2026.

Cummings believes the investment debate will shift from data center infrastructure companies to other sectors. “The real alpha long term is going to be in the other nine sectors of the market, and who is essentially answering the question of who’s going to use AI to their advantage versus be disrupted by it,” he said.

Incumbents must disrupt themselves to avoid being disrupted, Cummings argues. “In order for an incumbent to not get run over by something like AI, they have to disrupt themselves as a company. You can’t just be the same company and say, we’ll just plug in AI,” he said. The firm’s analysts assess companies’ preparedness by speaking to chief executives.

Cummings uses movie scripts as an analogy for pattern recognition. The current AI story starts with a 30-minute action scene, representing the scale of revenue at companies like Anthropic.

Despite potential setbacks, Cummings remains optimistic about AI’s long-term prospects. “Episodes like this are part and parcel of really important structural secular changes, and we’re prepared for volatility,” he said.

Cummings highlights biotech as an exciting area where AI can run simulations and analyze trial data. He also praises Meta’s AI product as a glimpse into the technology’s consumer potential.

Nvidia and key person risk

Janus Henderson’s research funds, which run sector-neutral, avoid hugging the largest stocks. However, Nvidia remains their largest position, making up 6.65% of the Global Research Equity Fund in August 2026.

Cummings addresses the investigation involving Nvidia CEO Jensen Huang, emphasizing leadership character and succession planning. He said technology is around 35 per cent of the universe for the product and that the tech team could shift most of that allocation from semiconductors to software if its view changed.

The firm takes corporate governance seriously and won’t ignore the issue, regardless of financial materiality. He was less concerned about the allegations of intellectual property being passed to China, saying the companies involved appear confident in containing the issue.

Cummings described the firm’s investment philosophy as centered on long-term business model analysis.