Investors are shifting how they evaluate AI stocks after the latest earnings season, moving from questions about whether Big Tech's massive spending will pay off to which companies will sustain profits over the long run, according to asset managers and Reuters analysis of market data.
Results from Microsoft and Amazon reassured markets that demand for AI infrastructure remains strong. Cloud growth is accelerating and capacity constraints persist. But for some of the world's largest asset managers, the real question is which companies can keep growing profits once those constraints ease.
Many funds still hold significant semiconductor positions, even after a sector rout in July when doubts emerged about AI spending and competition from China. At the same time, they are adding exposure to hyperscalers - the largest cloud service providers that can rapidly expand AI infrastructure to meet customer demand.
Hyperscalers gain favor
"The hyperscalers are being recognised in this moment as companies that are likely to be very large beneficiaries of this AI paradigm shift," said Brian Barbetta, co-head of the technology platform at Wellington Management, which manages about $1.3 trillion in assets. "They remain core holdings in our portfolios, and we've in fact increased our positioning in many of these companies recently."
Barbetta's view is part of a broader embrace of companies like Amazon, Microsoft, and Google - the operators of the cloud infrastructure that has become a foundation of AI workloads. These firms have the scale and customer relationships needed to monetize AI capex over time.
The biggest spenders lag chips
Shares in the four biggest AI capex spenders all lagged a 75% surge in the Philadelphia Semiconductor Index. They also trailed Nvidia-backed CoreWeave, up about 50%, and Nebius, up over 200%. Known as neocloud providers, those companies rent computing power to customers and have profited from elevated spot pricing for scarce AI capacity.
Richard Clode, portfolio manager at Janus Henderson's Bankers Investment Trust, said hyperscalers should eventually benefit from their investments. "By later next year into 2028, we think you're going to start seeing these companies growing profits and cash flow faster than the incremental capex growth," he said. Clode said Amazon was one of his fund's largest overweight positions. "Today's capex is tomorrow's sales," he said.
Reuters analysis estimates hyperscalers will generate about $340 billion more in annual operating cash flow in 2027 than in 2025, while capex is expected to rise by roughly $534 billion.
Not just chips versus cloud
John Lamb, equity investment director at Capital Group, which manages about $3.6 trillion in assets, said investors should treat AI as an expanding ecosystem. "It's not about whether chips are better investments than hyperscalers. It's about having both in your portfolio," he said, noting that data centres typically take 12 to 18 months from construction to revenue production. "We're just starting to see from the latest quarterly earnings this inflection point."
Pick your winners
Clode says companies that control both computing capacity and the layers that help customers deploy AI efficiently across different models - optimising cost and performance - will gain a competitive edge. He says Amazon, Microsoft, and Google have more lasting advantages than neocloud providers because of their scale and customer relationships.
Hyperscalers' valuations have compressed this year and remain below post-pandemic peaks. Microsoft trades on the highest multiple of forward earnings at 24.6 times; Meta trades at just 17.6 times.
Noah Weisenberger, chief US equity strategist with BCA Research, said neocloud providers could be vulnerable if new computing capacity comes online and pricing normalises, given heavier reliance on debt and uncertain growth. He recommends a long-hyperscalers, short-neoclouds trade.
Fewer future winners than players today
Even among the winners, there are challenges. Alberto Conca, CIO of Swiss wealth manager LGF+ZEST, estimates AI monetisation needs a fivefold to thirteenfold increase to justify current spending plans.
Barbetta expects competition to narrow the field of AI winners as the market matures. Companies with the broadest technology portfolios, deepest customer relationships, and most control over their own infrastructure are likely to pull ahead of specialised rivals. "There are absolutely going to be fewer winners in the future than there likely are players today," he said.
Why this matters for finance professionals
For portfolio managers and equity analysts, the distinction between hyperscalers and neocloud providers is not just about tech strategy - it's about cashflow visibility and risk. Hyperscalers may not deliver the same stock-price pop as pure chipmakers, but they also carry a broader revenue trajectory and a more predictable operating model. Neocloud firms offer potential upside, but they could face margin compression as capacity catches up to demand.
The actionable takeaway: pinpoint which level of exposure matches your client's risk tolerance, and start monitoring capex-to-cashflow conversion now, not in 2027 - because the market is already pricing in the gap.
Your membership also unlocks:
