Big Tech Plans $720 Billion in AI Spending as Investors Question When the Profits Will Come
Amazon, Google, Microsoft and Meta are continuing their massive AI infrastructure expansion, but rising computing costs, weaker search monetization and growing concerns about data center expansion are putting pressure on the investment story.
The biggest technology companies are preparing to spend more than $720 billion on capital expenditures in 2026, with artificial intelligence infrastructure accounting for much of the investment. Yet, as Amazon, Microsoft, Alphabet and Meta continue expanding their data centers, investors are becoming increasingly concerned about a difficult question: When will all this spending translate into meaningful profits?
The spending plans have also exposed a growing divide between Big Tech’s confidence in AI and Wall Street’s expectations for financial returns. Despite strong revenue growth across several businesses, the companies have yet to demonstrate that their enormous investments can generate profits at the scale investors were initially promised.
And there are few signs that the spending will slow down anytime soon.
Amazon, Google, Microsoft and Meta are committing hundreds of billions to AI
The four companies have outlined combined capital expenditure plans of at least $720 billion for 2026. Most of that spending is directed toward expanding computing capacity, building data centers and acquiring the hardware needed to support increasingly demanding AI systems.
Amazon has the largest spending plan among the group, with approximately $220 billion in capital expenditures expected this year. Alphabet has projected spending of between $195 billion and $205 billion, while Meta expects to spend between $135 billion and $145 billion.
Microsoft’s spending plans are also substantial, although its latest quarterly expenditure came in below Wall Street expectations.
These figures represent companywide capital expenditures rather than spending exclusively on AI. However, AI infrastructure is a major driver of the expansion.
The commitments extend beyond 2026, with the companies continuing to arrange substantial future investments in computing infrastructure.
For Amazon CEO Andy Jassy, the opportunity is significant enough to justify the expense. During the company’s earnings call, he described the expansion as a “once in a lifetime opportunity” to grow Amazon into a $1 trillion company.
Jassy also argued that the opportunity cannot be approached with a slower investment schedule.
“Would it be more convenient if this type of acceleration and opportunity happened over 25 years and you could spread the CapEx over 25 years? Yeah, it might be more convenient,” he said. “But that’s not the world we live in, and that’s not the opportunity.”
Google’s AI expansion is creating an unexpected problem for Search
Alphabet provides one of the clearest examples of the financial challenge facing AI companies.
Google’s AI investments are helping expand its cloud business, but the company’s newer AI-powered search experiences are also creating questions about how its traditional advertising model will evolve.
Google Cloud’s revenue more than doubled over the year, reflecting strong demand for computing and AI services.
However, the growth of AI Mode queries has introduced a monetization challenge.
Philipp Schindler, Alphabet’s senior vice president and chief business officer, acknowledged during the company’s earnings call that the increase in AI Mode queries had affected revenue because those interactions were generating lower monetization.
“It’s worth noting that the strong growth in AI Mode queries had an impact because of the lower monetization we were seeing in AI Mode,” Schindler said.
That creates an unusual situation for Google.
Its traditional Search advertising business has historically helped finance its investments in new technologies. But as more users interact with AI-powered search experiences, Google must find ways to generate comparable advertising revenue from those interactions.
Search advertising revenue grew 12% in the third quarter, compared with 17% growth during the same quarter last year.
Schindler said Google’s advertising experiments had produced promising results, but he did not provide specific examples demonstrating how AI Mode could eventually match the monetization of traditional Search.
Alphabet’s earnings report also triggered a significant market reaction. Its shares fell approximately 13% in a single trading session, marking the company’s worst one-day decline.
The larger concern is not whether Google can attract users to AI-powered search. It is whether the company can make that experience financially sustainable without weakening the business that currently supports its investments.
Amazon and Alphabet are already preparing to spend more in 2027
The latest earnings calls did little to reassure investors looking for evidence that Big Tech’s spending cycle might be approaching its peak.
Alphabet CFO Anat Ashkenazi indicated that the company expects capital expenditures to increase significantly in 2027.
Amazon also expects to spend meaningfully more next year than the $220 billion planned for 2026, according to Jassy.
Meta has not provided a specific 2027 spending figure, but financial institutions expect its investment requirements to increase as the company expands its AI capabilities.
The company’s Muse agentic AI program is another reason for its infrastructure requirements.
CEO Mark Zuckerberg explained that making the program profitable will depend partly on reducing the cost of the computing power required to operate it.
He expects compute efficiency to improve tenfold next year, but achieving that improvement will require significant investment.
That highlights another complication in the AI business: building more powerful systems is expensive, but making those systems cheaper to operate is equally important.
Anthropic’s computing costs are catching up with its revenue
The profitability challenge is not limited to the largest publicly traded technology companies.
Anthropic’s updated S-1 prospectus revealed how expensive it has become to operate its rapidly expanding AI business.
The company’s quarterly revenue reached $17 billion, representing growth of more than 1,000% compared with the previous year.
However, its computing costs reached approximately $17.5 billion during the same quarter.
In other words, the company’s reported computing expenses alone exceeded its revenue.
Anthropic has also outlined nearly $900 billion in future computing and infrastructure commitments and plans to begin building its own data centers.
Some of its infrastructure spending will involve Google, which supplies its TPU chips.
That arrangement has raised questions about Google’s own capacity constraints. If the company needs additional computing resources for its internal AI operations, investors want to understand why it is allocating capacity to external AI developers.
Google executives have not provided a straightforward explanation of whether revenue from these hardware agreements will meaningfully reduce the company’s own infrastructure spending.
There is also an important distinction between future contractual commitments and immediate cash generation. Google is funding the hardware it sells, while payments under those arrangements are tied to future delivery and use.
The deals may create future revenue opportunities, but they do not automatically solve the immediate financing demands of the AI buildout.
Microsoft raises $25 billion as AI infrastructure spending continues
Microsoft appeared to fare better than its competitors during the latest round of earnings-related market reactions.
The company reported $45.2 billion in capital expenditures for the quarter, below the approximately $50 billion Wall Street had expected.
However, the lower figure was not the result of a decision to reduce its AI investment ambitions.
Microsoft attributed the difference to delays in third-party capacity.
The company had also previously stood out among the major hyperscalers because it had not tapped debt markets to finance its AI expansion.
That changed shortly afterward.
Microsoft CFO Amy Hood had discussed the company’s access to debt markets during its earnings call. By the following morning, news emerged that Microsoft had launched a $25 billion, 10-part bond sale.
The move illustrates how financing has become an increasingly important part of the AI infrastructure race.
Companies are not simply deciding how much computing capacity they want. They must also determine how to fund the construction, hardware purchases and long-term commitments required to secure that capacity.
Data center restrictions are becoming another concern for Big Tech
The financial questions are developing alongside growing public and political scrutiny of data center expansion.
The latest U.S. election results have introduced additional uncertainty for technology companies, particularly in areas where candidates supporting data center moratoriums and other restrictions gained political support.
A moratorium can temporarily halt or restrict new development while authorities consider issues such as electricity demand, water consumption, infrastructure requirements and environmental effects.
That debate is becoming increasingly relevant as AI companies compete for access to larger computing facilities.
Data centers require substantial electricity, cooling systems and physical infrastructure. Their expansion can therefore create pressure on local resources, particularly in communities already facing infrastructure constraints.
The possibility of tighter regulations could complicate future construction plans and increase uncertainty around the speed of AI infrastructure expansion.
It could also give investors something they have been looking for: a potential reason for the industry’s spending growth to moderate.
However, there is no indication from the latest earnings calls that the major technology companies are independently preparing to abandon their expansion plans.
Wall Street is beginning to question the AI investment timeline
For years, investors largely accepted the argument that enormous AI infrastructure spending would eventually produce substantial financial returns.
There are genuine signs of business growth. Amazon reported its fastest growth in five years for the quarter, Meta’s advertising revenue continued to surge, and Google Cloud delivered significant revenue growth.
But strong revenue does not automatically mean that AI investments are profitable.
Building data centers, purchasing advanced chips, securing electricity and operating large computing systems all involve substantial costs. Companies must eventually generate enough revenue from AI products and services to cover those expenses while delivering returns that justify the original investment.
That is the uncertainty now facing Wall Street.
The possibility of an AI bubble is becoming a more prominent concern among investors, particularly as technology companies continue increasing their capital expenditure plans despite questions about monetization.
A slowdown in construction or tighter data center regulations could give companies additional time to improve their financial models. But it could also limit the computing capacity needed to meet future AI demand.
For now, the biggest technology companies appear committed to spending their way into the next stage of the AI race.
The unresolved issue is whether the revenue generated by that infrastructure will eventually justify the extraordinary cost of building it.