Marketdash

The AI Arms Race Is Making Big Tech Heavy Again

MarketDash
The era of asset-light tech giants is over. AI spending is gobbling up cash flow at a rate not seen since the dot-com boom, forcing a dramatic shift in how the biggest companies allocate capital.

Get Amazon.com Alerts

Weekly insights + SMS alerts

For over a decade, the biggest tech companies—think Alphabet (GOOGL), Meta Platforms (META), Microsoft (MSFT), and Amazon.com Inc. (AMZN)—were masters of the asset-light game. They printed cash, watched their margins grow, and sent piles of that excess capital back to shareholders through buybacks and dividends. They were financial efficiency machines.

That game is over.

According to Goldman Sachs analysts Ben Snider and Ryan Hammond, Wall Street now expects these "hyperscalers" to spend a staggering $667 billion on capital expenditures in 2026. That's up $127 billion from estimates just a few months ago and implies a 62% year-over-year jump. But the headline number isn't even the wildest part.

The real story is what that spending represents. Capex is now on track to eat up roughly 92% of these companies' cash flow from operations. That's a higher share than during the dot-com boom. "This dynamic has dramatically reduced hyperscaler free cash flows," Goldman Sachs said. In plain English, nearly every dollar these companies generate internally is being plowed right back into building AI infrastructure—data centers, chips, networking gear, and power capacity.

The Great Capital Allocation Shift

This isn't a tweak to the business model; it's a full-scale pivot. At the start of 2023, these tech giants were allocating about 43% of their cash flow to buying back their own stock. Today, that share has collapsed to just 16%. Collective buybacks fell 15% year-over-year in 2025. Management teams aren't optimizing for shareholder returns right now; they're optimizing for scale in AI.

This is a full-blown reinvestment cycle, and it's so big that some firms, including Oracle (ORCL) and Alphabet, have started borrowing money to help pay for it. Goldman expects we might see even more upward revisions to 2026 spending, potentially pushing the total toward $700–$725 billion.

Capex Growth Is Peaking — But Not Yet Slowing

The good news, if you can call it that, is that the breakneck growth rate of this spending is expected to start decelerating in the second half of 2026. Companies will still be building out their AI fortresses, just at a slightly less frantic pace.

Interestingly, Goldman argues that this slowdown could actually be a positive catalyst for the stocks. Why? Because it would give investors a clearer view of when free cash flow might bottom out. Once the spending frenzy eases, the market could start valuing these companies on their earnings again, rather than on the narrative of endless reinvestment.

"Decelerating capex growth should also signal a deceleration in the revenue and earnings growth of the infrastructure complex that has benefited from that investment spending," the analysts added. In other words, the companies that sell the picks and shovels for this AI gold rush might see their own growth cool off too.

Get Amazon.com Alerts

Weekly insights + SMS (optional)

What This Means For Investors

So here's the trade-off, laid bare. Every dollar that Microsoft, Amazon, and the rest are funneling into AI chips and data centers is a dollar that isn't coming back to you as a shareholder via buybacks or dividends anytime soon. The capital-light champions have become capital-heavy builders.

The bet they're making is purely strategic: spend aggressively today to protect what they hope will be a dominant, monopoly-like position in the AI-powered world of tomorrow. But it forces a pretty fundamental question. Will the future revenue and profits from all this AI wizardry be big enough to justify the historic scale of investment happening right now?

In that sense, the AI boom isn't just a technology story. It's a capital allocation story. And it's forcing investors to make a choice: are the tech giants temporarily sacrificing returns to win the next era, or are they permanently rewriting the rules of how Big Tech manages its money?

The AI Arms Race Is Making Big Tech Heavy Again

MarketDash
The era of asset-light tech giants is over. AI spending is gobbling up cash flow at a rate not seen since the dot-com boom, forcing a dramatic shift in how the biggest companies allocate capital.

Get Amazon.com Alerts

Weekly insights + SMS alerts

For over a decade, the biggest tech companies—think Alphabet (GOOGL), Meta Platforms (META), Microsoft (MSFT), and Amazon.com Inc. (AMZN)—were masters of the asset-light game. They printed cash, watched their margins grow, and sent piles of that excess capital back to shareholders through buybacks and dividends. They were financial efficiency machines.

That game is over.

According to Goldman Sachs analysts Ben Snider and Ryan Hammond, Wall Street now expects these "hyperscalers" to spend a staggering $667 billion on capital expenditures in 2026. That's up $127 billion from estimates just a few months ago and implies a 62% year-over-year jump. But the headline number isn't even the wildest part.

The real story is what that spending represents. Capex is now on track to eat up roughly 92% of these companies' cash flow from operations. That's a higher share than during the dot-com boom. "This dynamic has dramatically reduced hyperscaler free cash flows," Goldman Sachs said. In plain English, nearly every dollar these companies generate internally is being plowed right back into building AI infrastructure—data centers, chips, networking gear, and power capacity.

The Great Capital Allocation Shift

This isn't a tweak to the business model; it's a full-scale pivot. At the start of 2023, these tech giants were allocating about 43% of their cash flow to buying back their own stock. Today, that share has collapsed to just 16%. Collective buybacks fell 15% year-over-year in 2025. Management teams aren't optimizing for shareholder returns right now; they're optimizing for scale in AI.

This is a full-blown reinvestment cycle, and it's so big that some firms, including Oracle (ORCL) and Alphabet, have started borrowing money to help pay for it. Goldman expects we might see even more upward revisions to 2026 spending, potentially pushing the total toward $700–$725 billion.

Capex Growth Is Peaking — But Not Yet Slowing

The good news, if you can call it that, is that the breakneck growth rate of this spending is expected to start decelerating in the second half of 2026. Companies will still be building out their AI fortresses, just at a slightly less frantic pace.

Interestingly, Goldman argues that this slowdown could actually be a positive catalyst for the stocks. Why? Because it would give investors a clearer view of when free cash flow might bottom out. Once the spending frenzy eases, the market could start valuing these companies on their earnings again, rather than on the narrative of endless reinvestment.

"Decelerating capex growth should also signal a deceleration in the revenue and earnings growth of the infrastructure complex that has benefited from that investment spending," the analysts added. In other words, the companies that sell the picks and shovels for this AI gold rush might see their own growth cool off too.

Get Amazon.com Alerts

Weekly insights + SMS (optional)

What This Means For Investors

So here's the trade-off, laid bare. Every dollar that Microsoft, Amazon, and the rest are funneling into AI chips and data centers is a dollar that isn't coming back to you as a shareholder via buybacks or dividends anytime soon. The capital-light champions have become capital-heavy builders.

The bet they're making is purely strategic: spend aggressively today to protect what they hope will be a dominant, monopoly-like position in the AI-powered world of tomorrow. But it forces a pretty fundamental question. Will the future revenue and profits from all this AI wizardry be big enough to justify the historic scale of investment happening right now?

In that sense, the AI boom isn't just a technology story. It's a capital allocation story. And it's forcing investors to make a choice: are the tech giants temporarily sacrificing returns to win the next era, or are they permanently rewriting the rules of how Big Tech manages its money?