--- title: "$667 billion frenzy! Giant AI spending approaches its peak, free cash flow may hit bottom and rebound" description: "Goldman Sachs data shows that capital expenditures for hyperscale cloud computing companies are expected to reach $667 billion by 2026, accounting for 92% of operating cash flow. This spending frenzy " type: "news" locale: "en" url: "https://longbridge.com/en/news/276898503.md" published_at: "2026-02-25T14:00:21.000Z" --- # $667 billion frenzy! Giant AI spending approaches its peak, free cash flow may hit bottom and rebound > Goldman Sachs data shows that capital expenditures for hyperscale cloud computing companies are expected to reach $667 billion by 2026, accounting for 92% of operating cash flow. This spending frenzy has caused free cash flow yields to drop to historic lows, significantly shrinking stock buybacks. Analysts believe that the growth rate of capital expenditures will slow in the second half of 2026, potentially providing an opportunity for a rebound in free cash flow. Currently, the valuations of hyperscale cloud computing companies are at their lowest levels in the past decade, and the market is confident in the sustainability of AI investments Artificial intelligence-driven capital expenditures are reshaping the balance sheets of large technology companies. According to Goldman Sachs, capital expenditures for hyperscale cloud companies are expected to approach $667 billion this year, accounting for about 92% of their operating cash flow, leading to a significant reduction in stock buybacks and a drop in free cash flow yield to historic lows. Goldman Sachs analysts believe that **the growth rate of this capital expenditure frenzy is expected to slow down in the second half of 2026, providing visibility for a rebound in free cash flow.** Once free cash flow begins to recover, investors will be able to value these companies based on profitability, and the currently suppressed valuations may face a repricing opportunity. In terms of valuation levels, hyperscale cloud companies are currently trading at a forward price-to-earnings ratio of 24 times, which is at the 14th percentile of the past decade; the valuation premium of the "Tech Seven Giants" relative to the S&P 493 constituents has also dropped to a ten-year low. ## Record Capital Expenditures Consume Most Cash Flow According to FactSet data, **analysts expect total capital expenditures for hyperscale cloud companies to reach $667 billion in 2026**, with a year-on-year growth rate of about 62%, **slightly lower than the 73% increase in 2025, but the absolute scale remains historically rare.** It is noteworthy that **market expectations for capital expenditures are being continuously revised upward.** Since the fourth quarter earnings season, analysts have raised their consensus expectations for 2026 capital expenditures by $127 billion, and expectations for 2027 have been raised by $162 billion, indicating a strong confidence in the sustainability of AI investments. However, **this scale of capital expenditure is putting severe pressure on cash flow.** According to Goldman Sachs data, capital expenditures for hyperscale cloud companies are expected to account for 92% of their operating cash flow, with a significant decline in free cash flow conversion rates, and the drop in free cash flow yield far exceeds that of the S&P 500 overall. ## Sharp Decline in Stock Buybacks Pressures Shareholder Returns The sharp expansion of capital expenditures is directly impacting shareholder returns. According to Goldman Sachs data, total stock buybacks for hyperscale cloud companies are expected to decline by 15% year-on-year in 2025, with the proportion of buybacks in cash flow plummeting from 43% at the beginning of 2023 to about 16% currently, a decline of over 60%. This shift means that **during the peak phase of the AI investment cycle, tech giants are clearly prioritizing capital allocation from shareholder returns to infrastructure construction.** For investors who rely on buybacks to support earnings per share growth and stock performance, this structural change poses a significant pressure. ## Slowing Growth May Mark a Turning Point, Valuation Repair Path Becomes Clearer Goldman Sachs expects that **the growth rate of capital expenditures for hyperscale cloud companies will begin to slow down in the second half of 2026.** Scenario analysis shows that by then, the year-on-year growth rate of capital expenditures will be around 70% to 75%, with the annual growth rate roughly comparable to that of 2025, but the quarter-on-quarter growth rate will show a significant slowdown in the second half. Goldman Sachs points out that **the slowing growth rate of capital expenditures will give investors visibility into the potential bottoming of free cash flow, and the recovery of free cash flow will restore the market's ability to value these companies based on profitability.** From a valuation perspective, ultra-large-scale cloud computing companies are currently trading at a 24 times expected price-to-earnings ratio, which is at the 14th percentile over the past decade; the valuation premium of the "Magnificent Seven" relative to the S&P 493 constituents has also fallen to a ten-year low. Technically, the stock prices of the Mag 7 have recently rebounded after touching the 200-day moving average at the lower end of the range, and the market may be waiting for confirmation signals of a fundamental turning point. Risk Warning and Disclaimer The market carries risks, and investment should be approached with caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at one's own risk ### Related Stocks - [MSFT.US - Microsoft](https://longbridge.com/en/quote/MSFT.US.md) - [GOOG.US - Alphabet - C](https://longbridge.com/en/quote/GOOG.US.md) - [GOOGL.US - Alphabet](https://longbridge.com/en/quote/GOOGL.US.md) - [DTCR.US - Global X Data Center & Dgtl Infrs ETF](https://longbridge.com/en/quote/DTCR.US.md) - [MSFL.US - GraniteShares 2x Long MSFT Daily ETF](https://longbridge.com/en/quote/MSFL.US.md) - [CLOU.US - Global X Cloud Computing ETF](https://longbridge.com/en/quote/CLOU.US.md) - [IDGT.US - iShares US Digital Infrastructure and Real Estate ETF](https://longbridge.com/en/quote/IDGT.US.md) - [MSFX.US - T-Rex 2X Long Microsoft Daily Target ETF](https://longbridge.com/en/quote/MSFX.US.md) - [MSFU.US - Direxion Daily MSFT Bull 2X Shares](https://longbridge.com/en/quote/MSFU.US.md) - [IGV.US - iShares Expanded Tech Software Sector ETF](https://longbridge.com/en/quote/IGV.US.md) ## Related News & Research | Title | Description | URL | |-------|-------------|-----| | Brynne Tillman and Social Sales Link Partner with Alex Kutsishin's Fuel to Launch Brynne.ai: The 24/7 "Trust-Based" AI Sales Coach | Brynne Tillman, CEO of Social Sales Link, has partnered with Alex Kutsishin's Fuel to launch Brynne.ai, a 24/7 AI sales | [Link](https://longbridge.com/en/news/276924524.md) | | Google ‘Deeply Sorry’ for BAFTA News Alert That Included N-Word, Says the Message Was Not AI-Generated | Google has apologized for sending an offensive notification regarding the BAFTA Film Awards that included the N-word. Th | [Link](https://longbridge.com/en/news/276772906.md) | | 08:37 ETCapxel Launches LLM-LD, the First Open Standard for Making Websites Readable by AI Agents | Capxel has launched LLM-LD, the first open standard designed to make website content readable by AI systems. This new sp | [Link](https://longbridge.com/en/news/276895952.md) | | TCS and GitLab partner to bring AI-powered orchestration | TCS and GitLab partner to bring AI-powered orchestration | [Link](https://longbridge.com/en/news/276877058.md) | | CloudAlpha Capital Management Limited Hong Kong Has $107.04 Million Stock Holdings in Alphabet Inc. $GOOGL | CloudAlpha Capital Management Limited Hong Kong has reduced its holdings in Alphabet Inc. (NASDAQ: GOOGL) by 38.1% in Q3 | [Link](https://longbridge.com/en/news/276706039.md) | --- > **Disclaimer**: This article is for reference only and does not constitute any investment advice.