Earnings from Microsoft, Alphabet and Amazon this week will test whether the largest cloud platforms can show that their escalating AI infrastructure spending is producing cash returns quickly enough to support current valuations. The three companies have collectively spent about $134.9 billion on property and equipment in their latest reported quarter, while planned 2026 outlays point to an even larger buildout of data centers, servers and networking equipment.
Roughly 160 S&P 500 companies are scheduled to report third-quarter results, with Microsoft and Alphabet due after the Oct. 28 close and Amazon and Intel following after the Oct. 29 close. Eli Lilly and Caterpillar report before the market opens on Oct. 29, while Exxon Mobil and Chevron release results later in the week. LSEG’s aggregate estimates indicate the S&P 500 is heading toward roughly 9% blended earnings growth for the quarter.
The market’s immediate focus has shifted from AI revenue growth to the speed at which booked cloud demand turns into billed revenue, operating cash flow and free cash flow. That shift has become more visible in recent trading: the MAGS ETF, which tracks the so-called Magnificent Seven technology companies, fell nearly 7% from its 2026 high in the prior week, while Meta shares were down about 18% from their peak after the company reported negative free cash flow.
Capital spending is rising faster than free cash flow
Microsoft, Alphabet and Amazon are funding a physical expansion that is unusual even by the standards of previous cloud-computing cycles. Their combined capital spending is estimated at about $670 billion for 2026, based on the companies’ disclosed plans and cited forecasts.
Alphabet has provided one of the clearest illustrations of the acceleration. The company reported capital spending of $32.2 billion in 2023 and $91.4 billion in 2025, while the midpoint of its 2026 guidance stands near $200 billion. Reaching that figure would require quarterly expenditure of roughly $57.2 billion to $62.2 billion in the second half, above the $44.9 billion reported for the second quarter.
Amazon reported $54.2 billion of capital spending in the second quarter, up 68% from a year earlier. Its free cash flow was negative $18.2 billion in the first quarter and negative $8.8 billion in the second, leaving the company with two consecutive quarters in which cash investment exceeded internally generated cash after operating and capital expenditures.
Microsoft’s reported capital spending has also climbed sharply, reaching about $115.9 billion in fiscal 2026 from $28.1 billion in fiscal 2023. Free cash flow increased over the same period, from $59.5 billion to about $67 billion, but the pace has been far slower than the increase in spending on data-center capacity.
That gap does not necessarily indicate weak demand. It reflects the timing of the AI buildout: cloud providers must buy land, power equipment, chips, servers and networking capacity before customers begin using enough computing resources to generate recurring revenue. The earnings calls will show whether executives expect that gap to narrow or widen through 2027.
Backlogs offer demand evidence but leave a delivery challenge
Large order backlogs provide support for the spending plans, though they also establish a high bar for execution. Alphabet has disclosed a cloud backlog of roughly $514 billion. With its second-quarter cloud revenue at $24.8 billion, an annualized run rate of about $99.2 billion places the backlog at around 5.2 times current annual cloud revenue.
Amazon has said some compute orders extend into 2028, indicating that customers are reserving capacity well ahead of delivery. Yet the company’s cited 2026 capital spending of about $220 billion exceeds an estimated $168.8 billion in AWS revenue over the past four quarters, calculated by annualizing the reported $42.2 billion quarterly figure. The comparison shows how much infrastructure is being financed today for workloads that may arrive over several years.
The concern is less about whether AI demand exists than whether cloud operators can convert that demand at margins sufficient to absorb the cost of building capacity. Higher capital spending eventually feeds into depreciation expense once new facilities and equipment enter service. Based on combined 2026 spending of about $670 billion, straight-line depreciation over five to six years could add roughly $110 billion to $130 billion in annual depreciation when those assets are fully deployed.
Microsoft, Alphabet and Amazon will therefore face detailed questions on the useful life of AI equipment, data-center utilization rates, power availability, customer contracts and the extent to which depreciation will pressure operating margins.
Treasury yields sharpen the valuation comparison
The cash-flow issue is particularly sensitive while the 10-year U.S. Treasury yield remains near 5%. Based on trailing free cash flow and Oct. 23 market values, Alphabet’s free-cash-flow yield is estimated at about 1.4%, while Microsoft’s is about 1.7%.
A Goldman Sachs forecast cited in market discussions projects Microsoft free cash flow could fall from about $67 billion to $40.1 billion. If that projection materializes, the company’s forward free-cash-flow yield would fall to around 1% under the same valuation approach.
Energy companies reporting later in the week present a contrasting cash profile. Using the same calculation method, Exxon Mobil’s free-cash-flow yield is estimated near 4.1% and Chevron’s near 5.5%. LSEG’s consensus figures point to Exxon Mobil posting about $16.2 billion in third-quarter profit, up roughly 101% year over year, and Chevron earning about $9.3 billion, up about 158%.
Exxon Mobil reported $23.555 billion in operating cash flow and $17.028 billion in free cash flow in the second quarter, with capital spending around $6.5 billion. Chevron reported second-quarter free cash flow of $16.807 billion and has indicated annual capital spending is trending toward the lower end of its $18 billion to $19 billion target range.
Caterpillar and memory makers extend the AI test beyond cloud platforms
Caterpillar’s Oct. 29 results will offer another read on data-center construction and power demand. Analysts tracked by the market expect quarterly revenue of about $20.7 billion, up 17% from a year earlier, and adjusted earnings per share of roughly $7.
Bernstein estimates Caterpillar’s backlog at about $77 billion, equivalent to approximately 93% of an annualized $82.8 billion revenue estimate based on third-quarter expectations. The firm has described the backlog as covering all projected second-half 2026 revenue, compared with a historical average near 60%. Despite that visibility, TIKR data show Caterpillar shares were down about 33.89% from their June 30 closing high in the week before earnings.
Samsung Electronics, SK Hynix and Kioxia will add an Asian hardware perspective through their own reports, particularly on demand for high-bandwidth memory and other components used in AI servers. Their results can indicate whether cloud operators’ capital budgets are translating into orders throughout the semiconductor supply chain.
Eli Lilly’s report offers a different version of the cash-generation question. Market estimates place its third-quarter revenue near $22.97 billion and earnings per share around $8.38, representing annual growth of about 31% and 35%, respectively. The company generated about $10.8 billion in free cash flow in the first half, above the cited $9 billion total for full-year 2025.
For the cloud giants, revenue growth alone is unlikely to settle the market’s concerns. Their guidance on 2027 capital spending, depreciation, backlog conversion and cash-flow timing will determine whether traders view AI infrastructure as an expanding source of recurring returns or a longer and more expensive investment cycle than current valuations assume.
For deeper context on AI-driven markets and trading strategies, explore our guide on AI copy trading today.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
