- Big tech is spending hundreds of billions on AI infrastructure while near-term profits remain uncertain
- Chipmakers and memory companies are capturing most of the current financial upside from the boom
- Hyperscalers face rising costs and pressure as AI revenue lags behind massive capital investments
- The next phase of the cycle depends on whether AI adoption translates into sustainable profitability
The artificial intelligence boom has triggered one of the largest capital spending cycles in tech history. The world’s biggest companies are pouring unprecedented sums into AI infrastructure—data centres, chips, power systems, and software platforms—in a race to dominate the next computing era.
But as spending accelerates, a key question is emerging for investors: When will the returns actually show up?
A US$700 billion bet on AI
In 2026 alone, the four largest hyperscalers—Microsoft (NASDAQ:MSFT), Amazon (NASDAQ: AMZN), Alphabet (NASDAQ:GOOG), and Meta (NASDAQ:META)—are expected to spend roughly US$725 billion on capital expenditures, up sharply from about US$410 billion the year prior.
That figure rivals the GDP of entire countries.
The bulk of this spending is targeted at AI:
- Massive GPU clusters
- Next-generation data centres
- Networking and power infrastructure
In fact, analysts estimate that roughly 75 per cent of hyperscaler capex is now AI-related.
The reason is simple: demand for AI compute is outpacing supply, leaving cloud providers capacity-constrained and scrambling to build.
This article is a journalistic opinion piece that has been written based on independent research. It is intended to inform investors and should not be taken as a recommendation or financial advice.
The winners—for now
While hyperscalers are doing the spending, chipmakers are capturing the profits.
Nvidia: The face of the boom
Nvidia (NASDAQ:NVDA) sits at the centre of the AI buildout, supplying the GPUs that power everything from ChatGPT to enterprise copilots. Its data centre revenue has surged as nearly every AI dollar ultimately flows through its chips.
Micron: The hidden beneficiary in plain sight
Less obvious—but increasingly important—is memory supplier Micron (NASDAQ:MU).
AI systems require massive quantities of high-performance memory, and shortages have created a surge in pricing power:
- Micron’s revenue nearly tripled year-over-year in recent results
- Demand tied to AI infrastructure has driven record margins and earnings growth
Memory has effectively become a bottleneck in the AI supply chain, turning a historically cyclical industry into one of the biggest near-term winners.
The big spenders feel the pressure
For hyperscalers, however, the story is more complicated.
Companies like Microsoft, Amazon, Alphabet, and Meta are investing aggressively—but returns are less immediate:
- Capital intensity has surged to historically high levels
- Some firms are directing over 45–50 per cent of revenue into capex
- AI services still generate only a fraction of the infrastructure spend required to support them
This mismatch is starting to show up in investor sentiment.
Despite strong revenue growth, markets have at times reacted negatively to rising capex guidance, reflecting concern about near-term profitability.
The bottlenecks beneath the surface
Beyond chips, the AI boom is exposing deeper constraints.
Power and infrastructure
AI data centres require enormous electricity capacity—often comparable to small cities. As a result, power availability is becoming a limiting factor in where new facilities can be built.
Memory constraints
AI models demand exponentially more memory with each generation. The shift toward high-bandwidth memory (HBM) is consuming manufacturing capacity and tightening supply across the broader market.
Time lag
Most importantly, infrastructure takes years to build. Even as capex rises today, the revenue associated with that spending may not fully materialize for several years.
The monetization question
At the heart of the debate is a simple issue: AI usage is growing—but monetization is still evolving.
- Enterprises are experimenting with AI tools, but ROI remains uneven
- Pricing models for AI services are still being refined
- Many applications are still in early adoption phases
The long-term opportunity is enormous—potentially trillions in value—but the near-term financial payoff remains uncertain. There is a short squeeze happening right now and we are paying the price for it. Lenovo (OTC Pink:LNVGF) has said as much and warns that higher RAM prices are the “new normal” and we might never see them go back down to pre-2025 levels.
A split market story
This dynamic has created a divide across the AI ecosystem:
Clear winners (short-term)
- Nvidia, AMD (NASDAQ:AMD), Broadcom (NASDAQ:AVGO)
- Micron and memory suppliers
These companies benefit immediately from infrastructure demand.
Waiting for returns (long-term)
- Microsoft, Amazon, Alphabet, Meta
- Software platforms integrating AI
These firms are making the largest bets—but their returns depend on future adoption and monetization.
What investors should watch
The next phase of the AI cycle will hinge on a few key indicators:
- Revenue per AI workload
– Are companies able to charge meaningfully for AI services? - Margin trends at hyperscalers
– Does profitability stabilize as infrastructure scales? - Supply chain expansion
– Can chip, memory, and power constraints ease? - Enterprise adoption
– Do AI tools move from experimentation to necessity?
The bottom line
The AI boom is real—and so is the spending behind it.
But markets are beginning to shift from asking“Who is building AI?” to: “Who is actually making money from it?”
For now, the biggest profits are flowing to the companies selling the tools—chips, memory, and infrastructure.
The companies spending hundreds of billions?
They’re still waiting for their return.
Closing insight
The AI revolution may ultimately be measured not by how much companies spend—but by how efficiently they turn that spending into sustainable profits.
Join the discussion: Find out what the Bullboards are saying about software infrastructure and check out Stockhouse’s stock forums and message boards.