Billionaire Bill Ackman Says There’s a Huge Disconnect in Wall Street’s Expectations for Amazon, Microsoft, Meta, and Alphabet

The stock market hasn't been as keen on the hyperscalers in 2026 as it was in years past. Amazon (NASDAQ:AMZN) and Microsoft (NASDAQ:MSFT) have delivered better returns than the S&P 500 (SNPINDEX:^GSPC) year to date thanks to strong relative performances over the past month or so. Alphabet (NASDAQ:GOOG) (NASDAQ:GOOGL) and Meta Platforms (NASDAQ:META) haven't fared nearly as well — they’re down for the year.

But that could be an opportunity for long-term retail investors. Billionaire hedge fund manager Bill Ackman believes the market has soured on these four stocks due to a "notable disconnect" between market expectations and the businesses' fundamentals. That's reflected in growing skepticism among analysts that the massive amounts these companies are spending on new data centers will yield meaningful returns on investment.

But the actual financial results and management's commentary should give investors confidence that the future may be brighter for these hyperscalers than the average analyst expects.

Pershing Square CEO Bill Ackman. Image source: Getty Images.

The huge disconnect on Wall Street

In his third-quarter letter to Pershing Square shareholders, Ackman expressed his belief that most analysts aren't properly accounting for the potential revenue growth stemming from the stepped-up capital spending among Amazon, Microsoft, Meta, and Alphabet. He wrote:

Consensus analyst estimates for the hyperscalers reflect a notable disconnect: they expect the current level of capital intensity to persist for years into the future, yet they assume minimal acceleration in revenue growth beyond this year and next. Consensus estimates treat the capex buildout as a structural increase in maintenance capital intensity rather than as an investment capable of producing outsized future revenue growth and attractive returns on capital.

That is to say, Wall Street is acting like the increase in capital expenditures won't result in a commensurate increase in revenue over time. That's despite the fact that their prior spending has already been doing so for all four hyperscalers.

Meta, notably, saw accelerating growth in its advertising business as it accelerated its capital expenditures. And with the potential to grow its Muse agentic AI service, expand business messaging, and collect licensing revenue from its large language models (LLMs), it could see a sharp rise in direct artificial intelligence monetization in the near future.

Alphabet has seen Google Cloud revenue accelerate significantly, more than doubling year over year in its most recent quarter. Amazon and Microsoft are also experiencing accelerating cloud revenue growth, with expectations for sustained growth near or above current levels.

Nonetheless, Ackman points to consensus expectations on Wall Street for stable or declining revenue growth over the next few years for the companies, while capital intensity remains unchanged relative to 2026. That's missing a fundamental factor in the potential of hyperscalers' capital spending.

Is the return on invested capital there?

Ackman estimates that 1 gigawatt of AI data center capacity currently costs the hyperscalers between $40 billion and $50 billion to build. Importantly, they have to spend a significant portion of that money well before the compute capacity in question can actually start generating revenue. Amazon CEO Andy Jassy said it takes about two years from breaking ground on a data center until it's ready to be filled with servers and networking equipment.

Once a new data center starts generating revenue, however, Ackman estimates hyperscalers can bring in $15 billion to $20 billion in annual revenue per gigawatt. At a 50% incremental operating margin, that's $7.5 billion to $10 billion in operating income per year from a $40 billion or $50 billion investment. That's a return on invested capital of 15% to 25%.

Based on the consensus estimate for capital expenditure growth, that return on invested capital should result in revenue growth and earnings growth well in excess of Wall Street's consensus over the next three years.

What's more, the hyperscalers are set to produce improving returns on invested capital over time. As more and more capital spending goes toward the actual revenue-generating assets of servers and networking equipment instead of the data center structures, the revenue generated per dollar of capex should increase. On top of that, the cost of new servers will likely decline over time as a larger percentage of servers get built around custom silicon rather than GPUs, memory prices come back down, and overall supply constraints ease.

As such, investors should expect the hyperscalers to produce very strong revenue and earnings growth over the next few years, exceeding analysts' expectations. At their current valuations of between 21 and 28 times forward earnings, all of them look like good buying opportunities, but the low-priced shares of Alphabet and Meta look especially appealing right now.

Adam Levy has positions in Alphabet, Amazon, Meta Platforms, and Microsoft. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, and Microsoft. The Motley Fool has a disclosure policy.