UBS Warns AI Stocks Near Dot‑Com Peak - Tech Trading ~35x P/E as Big Tech Plows $350B into Capex

UBS (NYSE:UBS) says the prices on artificial-intelligence names are getting uncomfortably close to the bubble territory we saw in the late 1990s.

The bank's HOLT Economic model shows the U.S. tech complex trading at an aggregate price-to-earnings north of 35x. That's roughly the multiple seen near the dotcom peak - a sign that a big chunk of market value is pricing in future cash flows rather than cash on the books today. Short version: there isn't much room for disappointment.

Part of the story is cash being shoveled into infrastructure and R&D. Meta Platforms (NASDAQ:META), Alphabet (NASDAQ:GOOGL), Amazon (NASDAQ:AMZN) and Microsoft (NASDAQ:MSFT) are on track to commit roughly $350 billion this year - more capex than the listed energy and utilities sectors in the U.S. and Europe combined. Add Apple (NASDAQ:AAPL), Nvidia (NASDAQ:NVDA) and Broadcom (NASDAQ:AVGO) and the technology-heavy R&D tab dwarfs what many entire regions are spending.

UBS points out that much of the AI story is built on anticipated revenue streams rather than current sales. That's been visible in a swathe of pilots and proofs-of-concept that haven't immediately translated into top-line growth. When the math depends on future cash, timing and execution become everything.

There are operational headaches too. Massive capex raises the bar on returns, and data-center growth bumps up against energy availability and cost. Add intensifying competition from China and you get a picture where profit margins and cash-flow resilience could come under pressure.

The bank's HOLT team sees consensus forecasts implying drops in Cash Flow Return on Investment for a set of big tech names over the next couple of years - specifically for AMZN, META, MSFT and GOOGL. Put bluntly: high valuation plus falling CFROI is a recipe for headline risk if revenue ramps slip.

UBS also flags a concentration issue. The handful of hyperscalers and chip leaders are projected to produce an outsized share of U.S. economic profit next year - roughly 37% by the bank's estimate. That's a level of market influence that increases beta for the overall market when these names wobble.

On the flip side, UBS highlights areas that historically don't move in lockstep with U.S. mega-cap tech: overseas quality growth names, sector exposure outside the AI frenzy, and subsectors like real estate, utilities, energy, communication services and staples. The bank frames those as diversification routes rather than answers - wording worth noting if you're mapping correlation risk.

For traders, the takeaways are straightforward facts: sky-high multiples, huge capex commitments, and concentrated profit pools. Volatility is the likely companion. Is this dotcom redux or a different beast? That's the debate heating up on desks right now.