After Earnings Season, Which Tech Giant Can Turn AI Compute Into Cash Fastest?
The pace at which each company converts investment into returns varies, and investors are beginning to scrutinize the return outlook for each business more closely. Market divisions emerged quickly after the latest earnings reports were released. Amazon’s cloud business posted its fastest growth since early 2022, sending its shares up nearly 10% after hours. Microsoft’s Azure revenue hit a record for a single quarter, and its shares jumped more than 15% the day after earnings were released. Meta was weighed down by a weaker revenue-growth outlook and high capital spending, sending its shares down more than 7% cumulatively after the report. Google posted the first quarter of negative free cash flow since its listing, and its shares fell more than 6% the following day. **Capital spending growth with no clear end** Industry data show that Google, Amazon, Microsoft and Meta together spent $1.1 trillion on capital expenditures from the beginning of 2023 through the end of June 2026. The latest earnings reports showed that Amazon’s capital spending surged 69% year over year to about $54 billion, while Meta’s rose 83% to $31.1 billion. Microsoft’s capital spending also increased 69%, and Google raised its full-year capital-spending target to as much as $205 billion. Rishi Jaluria, an analyst at RBC Capital Markets, said capital-spending growth currently shows almost no sign of ending. Investors, however, still want companies to balance AI investment with their existing high-return businesses. Importantly, capital spending reported for the current period does not fully reflect the financial pressure. The four tech giants have also signed numerous multiyear contracts, creating substantial off-balance-sheet financial commitments that will result in fixed cash outlays over the next several years. These commitments do not entirely equal future capital spending. Equipment-purchase agreements, for example, will later be recorded as capital expenditures, while data-center leases and power-purchase agreements are operating expenses. Google disclosed that, as of the end of the second quarter, its future financial commitments related to AI investment had increased by about $500 billion from three months earlier, mainly because of long-term technology-infrastructure and power-purchase agreements. Meta added $233 billion in commitments during the quarter, including $96 billion in data-center and network leases, $112 billion in purchase commitments and $25 billion in new debt. In July, it added another $68 billion in data-center leases. Microsoft signed more than $130 billion in new data-center lease agreements in the second quarter. Amazon has not disclosed comparable figures. Industry observers said Google, Meta and Microsoft added nearly $900 billion in AI-related commitments in just those three months. The continued expansion of infrastructure is putting significant pressure on free cash flow. Industry data show that the four companies’ combined free cash flow fell to a 10-year low of $7 billion, with only Microsoft and Meta generating net cash inflows. Google posted the first negative-free-cash-flow quarter in more than 20 years as a public company, with a free-cash-flow shortfall of $6 billion. Amazon CEO Andy Jassy said the company would continue to face headwinds because many data centers are under construction simultaneously. Completing and commissioning a data center does not mean it can immediately provide computing capacity to customers; it typically takes another two years before the facility reaches the stage at which it can generate revenue. The latest earnings reports show that some investment is translating into faster revenue growth, particularly in cloud computing. Cloud growth accelerated at Google, Amazon and Microsoft. Google Cloud revenue rose 82% year over year. Amazon Web Services revenue increased 37% to $42.2 billion, its fastest growth since the fourth quarter of 2021. Microsoft said revenue from its Azure cloud-computing business rose 43% year over year. Industry observers said the acceleration in cloud revenue growth is benefiting, on one hand, from AI companies such as OpenAI and Anthropic renting computing capacity to run large models. On the other hand, many traditional companies are gradually deploying AI tools, creating substantial demand for cloud-computing capacity. At the same time, the contract backlog generated by demand for AI computing has become a key indicator supporting market growth expectations. Amazon, Google and Microsoft together have nearly $1.7 trillion in contracted backlog, more than double the amount from a year earlier. Meta CEO Mark Zuckerberg told investors that many institutions were interested in leasing the company’s computing capacity, at rental rates significantly above cost. He also said, however, that the profit margin from selling intelligent services directly is far higher than the margin from simply renting out computing capacity. Dec Mullarkey, managing director at alternative asset manager SLC Management, said Meta’s failure to provide a clear plan for leasing computing capacity was one reason its shares fell 8% on Thursday. Cloud-business growth and a contract backlog measured in the hundreds of billions of dollars have demonstrated that demand for AI computing is real. But massive current-period capital spending, extensive off-balance-sheet commitments and the multiyear gap between data-center commissioning and revenue generation continue to weigh on free cash flow and amplify pressure for near-term returns. The central question of this AI cycle is shifting from whether to spend money building computing capacity to which will move faster: capital spending or revenue realization. Going forward, investors may no longer be satisfied with an “AI narrative.” They are likely to assess each company’s investment efficiency more rigorously. The companies that can convert computing capacity into high-margin intelligent services at a lower cost will be the ones that emerge as the true commercial winners.
Amazon’s cloud business posted its fastest growth since early 2022, sending its shares up nearly 10% after hours. Microsoft’s Azure revenue hit a record for a single quarter, and its shares jumped more than 15% the day after earnings were released.
Meta was weighed down by a weaker revenue-growth outlook and high capital spending, sending its shares down more than 7% cumulatively after the report. Google posted the first quarter of negative free cash flow since its listing, and its shares fell more than 6% the following day.
**Capital spending growth with no clear end**
Industry data show that Google, Amazon, Microsoft and Meta together spent $1.1 trillion on capital expenditures from the beginning of 2023 through the end of June 2026.
The latest earnings reports showed that Amazon’s capital spending surged 69% year over year to about $54 billion, while Meta’s rose 83% to $31.1 billion. Microsoft’s capital spending also increased 69%, and Google raised its full-year capital-spending target to as much as $205 billion.
Rishi Jaluria, an analyst at RBC Capital Markets, said capital-spending growth currently shows almost no sign of ending. Investors, however, still want companies to balance AI investment with their existing high-return businesses.
Importantly, capital spending reported for the current period does not fully reflect the financial pressure. The four tech giants have also signed numerous multiyear contracts, creating substantial off-balance-sheet financial commitments that will result in fixed cash outlays over the next several years.
These commitments do not entirely equal future capital spending. Equipment-purchase agreements, for example, will later be recorded as capital expenditures, while data-center leases and power-purchase agreements are operating expenses.
Google disclosed that, as of the end of the second quarter, its future financial commitments related to AI investment had increased by about $500 billion from three months earlier, mainly because of long-term technology-infrastructure and power-purchase agreements.
Meta added $233 billion in commitments during the quarter, including $96 billion in data-center and network leases, $112 billion in purchase commitments and $25 billion in new debt. In July, it added another $68 billion in data-center leases.
Microsoft signed more than $130 billion in new data-center lease agreements in the second quarter. Amazon has not disclosed comparable figures.
Industry observers said Google, Meta and Microsoft added nearly $900 billion in AI-related commitments in just those three months.
The continued expansion of infrastructure is putting significant pressure on free cash flow. Industry data show that the four companies’ combined free cash flow fell to a 10-year low of $7 billion, with only Microsoft and Meta generating net cash inflows.
Google posted the first negative-free-cash-flow quarter in more than 20 years as a public company, with a free-cash-flow shortfall of $6 billion.
Amazon CEO Andy Jassy said the company would continue to face headwinds because many data centers are under construction simultaneously. Completing and commissioning a data center does not mean it can immediately provide computing capacity to customers; it typically takes another two years before the facility reaches the stage at which it can generate revenue.
The latest earnings reports show that some investment is translating into faster revenue growth, particularly in cloud computing.
Cloud growth accelerated at Google, Amazon and Microsoft. Google Cloud revenue rose 82% year over year. Amazon Web Services revenue increased 37% to $42.2 billion, its fastest growth since the fourth quarter of 2021. Microsoft said revenue from its Azure cloud-computing business rose 43% year over year.
Industry observers said the acceleration in cloud revenue growth is benefiting, on one hand, from AI companies such as OpenAI and Anthropic renting computing capacity to run large models. On the other hand, many traditional companies are gradually deploying AI tools, creating substantial demand for cloud-computing capacity.
At the same time, the contract backlog generated by demand for AI computing has become a key indicator supporting market growth expectations. Amazon, Google and Microsoft together have nearly $1.7 trillion in contracted backlog, more than double the amount from a year earlier.
Meta CEO Mark Zuckerberg told investors that many institutions were interested in leasing the company’s computing capacity, at rental rates significantly above cost. He also said, however, that the profit margin from selling intelligent services directly is far higher than the margin from simply renting out computing capacity.
Dec Mullarkey, managing director at alternative asset manager SLC Management, said Meta’s failure to provide a clear plan for leasing computing capacity was one reason its shares fell 8% on Thursday.
Cloud-business growth and a contract backlog measured in the hundreds of billions of dollars have demonstrated that demand for AI computing is real. But massive current-period capital spending, extensive off-balance-sheet commitments and the multiyear gap between data-center commissioning and revenue generation continue to weigh on free cash flow and amplify pressure for near-term returns.
The central question of this AI cycle is shifting from whether to spend money building computing capacity to which will move faster: capital spending or revenue realization.
Going forward, investors may no longer be satisfied with an “AI narrative.” They are likely to assess each company’s investment efficiency more rigorously. The companies that can convert computing capacity into high-margin intelligent services at a lower cost will be the ones that emerge as the true commercial winners.