Hunting for "forever businesses" and 5 stocks for your watchlist

Please note this interview was recorded Tuesday 18 November 2025

It was only in September that I last spoke with Janus Henderson Investors’ Joshua Cummings, when he participated in a panel I hosted at Livewire Live. The world has changed considerably since then. 

The everything rally has developed some cracks, the epicentre of which lies in AI valuations, with tremors spreading into the global economy, inflation dynamics and the debate around US debt. Leadership remains narrow, and the pressure on the mega-cap cohort has only intensified.

Against that backdrop, Cummings offered a line that neatly captures the moment investors now face;

“I think the money from here is going to be made in the 493, not the seven. I mean, it has to be or the market’s got issues.”

This isn’t a contrarian soundbite, it’s a structural observation. While the hyperscalers still dominate earnings revisions, Cummings argues the next leg of returns depends on where earnings breadth emerges, how far AI’s influence pushes into non-tech sectors, and whether the rising capital intensity of big tech creates new opportunities elsewhere.

He also sees subtle shifts in the US labour market becoming increasingly important. Growth remains intact, but the pace of hiring has eased just enough to raise new questions about corporate behaviour, demand, and the early productivity effects of AI. Add in an AI CapEx cycle that continues to accelerate, and investors must now consider a market supported by strong fundamentals in a narrow band, but with meaningful upside forming outside it.

In our latest discussion, we explored these themes, including where Cummings believes genuine earnings power is emerging and how he’s positioning for the next phase of market dispersion. Watch the video above for the full experience, or read a summary below. 

INTERVIEW SUMMARY

Earnings momentum still concentrated in mega-cap tech

Cummings opens with a straightforward assessment: the centre of gravity for earnings momentum remains in mega-cap technology. While market breadth has improved at the margin, the strongest upgrades are still coming from the hyperscalers. 

“The fundamental momentum, meaning upward sales and earnings revisions within the market, is still right where it’s been - in mega cap tech,” he says.

He points to the extraordinary lift in CapEx expectations heading into 2026. Before third-quarter results, hyperscaler CapEx growth estimates were around 20%; after earnings, they surged to roughly 36%. 

Despite debate about whether AI spending is running too hot, he argues valuations remain grounded and physical constraints - power, water, permitting, cooling - limit the pace at which excess capital can actually distort fundamentals. “They’re not priced on the moon,” he says.

Labour market signals growing harder to read

Beyond AI and tech, Cummings sees the labour market as one of the more important swing variables. Data is still “fine,” but the rate of hiring has slowed just enough to raise eyebrows. Importantly, he doesn’t believe this is simply a continuation of post-COVID rebalancing. After a normalisation phase in 2023 and early 2024, companies are again reducing headcount — this time without any collapse in demand.

That disconnect raises new questions. Are companies acting pre-emptively? Are they budgeting for AI-driven productivity? Are they simply being overly cautious? 

“It begs the question, do they see something? Are they simply being conservative? Is it just more optimization post COVID? I don’t know if it’s all of it,” he says. 

The upcoming holiday period will be an important signal for how the consumer and corporate confidence are actually evolving.

AI’s next chapter moves beyond the Magnificent Seven

Cummings believes investors need to prepare for AI’s impact far beyond the narrow group of current market leaders. While he fully acknowledges the strength of the hyperscalers today, he thinks investors are underestimating how quickly the economic benefits of AI will spread horizontally across sectors. 

“I think the money from here is going to be made in the 493, not the seven,” he says.

He highlights the inversion of past technology cycles: instead of enterprise-first adoption trickling down to consumers, AI has moved in reverse, with consumer applications exploding first and enterprise use cases building later. This inversion is already reshaping competitive moats, especially in e-commerce. The rise of “agentic commerce,” where large language models mediate product discovery, means companies must rethink product descriptions, digital discoverability and logistical sophistication.

Inventory quality and logistics, the least digital pieces of the chain, become critical differentiators. That naturally favours platforms like Amazon and Walmart, but Cummings stresses the field is far from settled. “A lot remains to be seen,” he says. 

Process, mistakes and the Nvidia case study

One of the most revealing segments of the conversation centres on process. Specifically, how the team handles mistakes. Cummings distinguishes sharply between a thesis break and external shock. That distinction underpinned their approach to Nvidia during the 2022 downturn.

Their semiconductor analyst identified GPUs early in 2021, and the team built a position. But the macro unwind of 2022 dragged the stock sharply lower. This is where Cummings makes the critical distinction:

“We make lots of mistakes, but there’s a key difference when your thesis is breaking. That’s one thing. When the stock performance is bad because of external stimuli or external shock, that’s quite another. 

And those are the types of situations where we like to double down on businesses we love.”

They sharpened their work, increased the position materially, and today Janus Henderson is one of Nvidia’s largest active shareholders. “It seems easy in hindsight — it was not easy,” he adds.

Stocks he likes now: Formula One, European banks, GLP-1s and biotech

Cummings points to Formula One (NASDAQ: FWONK) as the quintessential “forever business” in the portfolio. The pandemic created a moment of genuine uncertainty about the sport’s viability, one that many investors interpreted as a structural impairment. That’s precisely when his team took the opposite stance. They dramatically increased their position during the downturn and are now the largest active shareholder globally. 

“That’s an example where cyclicality and macro gave us an opportunity to really jump in,” he says.

He then turns to Europe, where he continues to see earnings leverage in banks. Years of regulatory headwinds are easing, capital requirements have stabilised, and returns on equity are improving. The team is overweight names such as Erste Bank (VIE: EBS), NatWest (NYSE: NWG) and UniCredit (BIT: UCG), and Cummings is keen to point out the core reason for these investments, which is not valuation; "This is about earnings growth and relative earnings", says Cummings. 

GLP-1 therapeutics, after a year of volatility, are moving into a new, more constructive phase. Pricing clarity is improving, tablet versions are approaching launch, and supply-chain bottlenecks are easing. Cummings continues to favour Eli Lilly (NYSE: LLY).

Finally, he highlights biotech as a sector that has quietly begun to turn. Years of political and pricing pressure depressed valuations, but M&A activity is picking up and sentiment is healing.

“If you want heat and you want what’s the beta in the market… biotech is quite interesting here,” he says.

Managed Fund

Janus Henderson Global Research Equity Fund

Global Shares

View