# AI Capex to Hit $1 Trillion – And Estimates Are Still Too Low > Source: > Published: 2026-08-05 00:00:00+00:00 # AI Capex to Hit $1 Trillion – And Estimates Are Still Too Low August 05, 2026 ### Beth Kindig #### Lead Tech Analyst - Analysts have persistently underestimated Big Tech AI spending, with guided 2026 spending now over 150% higher than initial estimates. - JP Morgan and Goldman Sachs are now forecasting AI capex of $1 trillion or more in 2027, and trillions more in the coming years. - Several Big Tech names continue to point to undercapacity compared to demand. Big Tech capex is the driving force behind the AI infrastructure trade, yet Wall Street has repeatedly underestimated the sheer scale of the buildout. Looking back to 2024, actual spending exceeded early forecasts by roughly 25%, and then exceeded forecasts by a whopping 62% in 2025. Currently, in 2026, *the guidance for $732.5 billion is now 158% higher than forecasts issued two years ago. * Last week, the market received confirmation that Microsoft, Meta, Amazon and Google are expected to deploy approximately $432 billion in the second half of 2026 alone. This means the flashy forecast that capex will reach $1 trillion in 2027 is now firmly within reach. Because of the outsized increase in 2026, capex will need to grow by only 36.5% next year - meaning, Big Tech could cut its capex growth rate by more than half and still reach the $1 trillion threshold. For investors tracking this trend, capex-related selloffs have created opportunities to generate alpha in key suppliers. Below, we review the single most important line item to the AI trade and what it’s communicating regarding the strength of the AI market in the coming months. ## AI Capex Estimates Have Consistently Underestimated Big Tech Spending Historical trends offer strong evidence that actual capex spending far exceeded analyst estimates. For example, we noted that analysts initially expected [Big Tech capex to come in at around $200 billion in 2024](https://io-fund.com/artificial-intelligence/market-trends/big-tech-ai-stocks-to-showcase-ai-gains-capex-in-q4-reports), or growth of 30% YoY. This ended up being well below actual capex outlays, which came in at just over $250 billion, or growth of 62% YoY, more than double initial growth expectations. In September 2024, initial estimates from Goldman Sachs placed [2025 Big Tech capex](https://www.goldmansachs.com/insights/articles/why-ai-companies-may-invest-more-than-500-billion-in-2026) at $253 billion, before moving to $280 billion in early 2025. [Estimates began to move up to $308 to $325 billion](https://www.spglobal.com/market-intelligence/en/news-insights/research/big-tech-earnings-preview-microsoft-meta-amazon-n-apple) by mid-2025, yet Big Tech would end 2025 at $410 billion, coming in 62% above the initial estimate as growth accelerated to 64% YoY. Initial expectations for 2026 of $284 billion -- just 12% YoY growth versus the initial 2025 figure of $253 billion -- would prove laughably low. By November 2025, 2026 capex forecasts had surged to $527 billion, an 86% increase over early estimates. *Goldman Sachs data shows analysts repeatedly raising AI hyperscaler capex forecasts. Estimates for 2025 increased from approximately $253 billion to nearly $394 billion, while 2026 forecasts surged from about $284 billion to more than $525 billion. The trend highlights growing expectations for AI infrastructure and data center investment. Source: FactSet, Goldman Sachs Research* ## Where Big Tech Capex Stands After Q2 2026 Capex spending in the first half of 2026 already hit $301 billion across Microsoft, Meta, Amazon and Google, with updated guidance from the four pointing to annual capex of $732.5 billion this year. For comparison, this latest figure is 39% higher than expectations from Nov. 2025, and 158% higher than expectations from Sept. 2024. At $732.5 billion, 2026 capex growth would be 79% YoY, a 15 point acceleration from 2025, and notably 6.5X higher than early forecasts off a much larger base than was previously expected. It also implies a strong acceleration into 2H, with ~$432 billion in spending on deck over the next two quarters. Here is where Big Tech’s 2026 spending stands so far in 2026, and their guidance for the full year: As seen in the chart above, capex intensity on a dollar basis is increasing substantially in into 2H. All four are currently expected to see 2H spending between $27 to $39 billion higher than 1H, adding more than $10 billion to quarterly capex bills. *Sign up for our free newsletter where our next article will discuss what this capex intensity means for Big Tech’s cash flows heading into 2027.* ## AI Capex Could Reach $7.6 Trillion Between 2026 and 2031 Based on current guidance of $732.5 billion, Big Tech capex would need to rise by 36.5% YoY to hit $1 trillion next year, representing a rather stark deceleration from 2026’s guided 79% growth. In dollar terms, Big Tech would need to add just $267.5 billion in 2027 to reach $1 trillion, less than the $332.5 billion increase guided for 2026. At present, current consensus estimates point to Big Tech spending around $934.5 billion in capex in 2027, or only 7% below the $1 trillion threshold with sixteen months to go. Multiple banks are already penciling in capex to easily surpass $1 trillion next year, and commentary from management teams and signals across the supplier ecosystem suggest this is easily doable. *The table displays consensus 2027 capital expenditure forecasts for the four largest hyperscalers. Google is expected to spend $284.8 billion, followed by Amazon at $256.5 billion, Microsoft at $207.6 billion, and Meta at $185.6 billion. Combined, the companies are projected to invest $934.5 billion in 2027, putting Big Tech AI and infrastructure spending within reach of the $1 trillion milestone. Source: MarketScreener* For example, JP Morgan currently expects [total hyperscaler capex](https://www.goldmansachs.com/insights/articles/tracking-trillions-the-assumptions-shaping-scale-of-the-ai-build-out) to rise from its estimate of $800 billion in 2026, which may include other players outside of the big four tech companies, to $1 trillion next year. Goldman Sachs is [forecasting capex of $1.01 trillion in 2027,](https://www.goldmansachs.com/insights/articles/tracking-trillions-the-assumptions-shaping-scale-of-the-ai-build-out) rising 32% from its 2026 estimate of $765 billion; overall, the firm is forecasting an astronomical $7.6 trillion in cumulative AI capex from 2026 to 2031, with capex hitting nearly $1.64 trillion by 2031 – exceeding the GDP of many smaller countries. *This infographic summarizes Goldman Sachs' baseline estimate for AI-related capital expenditures from 2026 through 2031. Total annual AI capex is projected to grow from $765 billion in 2026 to $1.64 trillion in 2031, with cumulative spending reaching $7.6 trillion. Spending is divided across three categories: compute, data centers, and power infrastructure, with compute representing the largest share throughout the forecast period. The projections illustrate the scale of investment expected to support continued AI infrastructure expansion. Source: Goldman Sachs Global Institute, Goldman Sachs Global Investment Research* ## Hyperscalers Continue to Face AI Capacity Constraints Overall, the key themes delivered in Q2’s earnings reports across Big Tech underscore future capex increases as a rather necessity, with demand continuing to outpace capacity delivered as supply constraints layer into the picture. Alphabet explicitly noted that it expects capex to increase “significantly” in 2027, explaining that despite its substantial capacity investments of the past three years, demand still outpaces these investments and that it continues to be supply constrained. mid When asked about the risk of overcapacity on its earnings call, Microsoft said the current “situation is obviously that demand exceeds available supply in a sort of relatively extreme moment.” This indicates that Microsoft could be planning substantial capex increases into 2027 and beyond as it expects to double its total data center capacity from FY2025 to FY2027. Even after adding $20 billion to its capex outlook to account for rising memory costs, Amazon CEO Andy Jassy said AWS “will still not have enough capacity to meet all the demand we have in 2026, and I believe this dynamic will also be true in 2027 too.” Similar to Microsoft, Amazon also expects to double its data center power capacity by the end of 2027 versus 2025 as capacity reservations for 2028 enter the mix, implying capex intensity is likely to remain elevated as its data center footprint gets built out Meta is floating the idea of selling “excess capacity” to third parties. However, the company still says “our current plans are geared towards maximizing 2026 and 2027 capacity," signaling that the firm does not expect to slow its spending in the near term. ## Higher Than Expected Capex Translates to Supplier Sales Revisions As analysts have underestimated the scale of Big Tech capex, we can see that subsequent capex guidance has led to huge jumps in sales expectations for suppliers. Nvidia provides the cleanest example of this. *The chart tracks analyst estimates for Nvidia's revenue in the next fiscal year from mid-2025 through August 2026. Forecasts increase consistently over the period, rising from approximately $245 billion to $561.5 billion. Several sharp upward revisions occur in early and mid-2026, reflecting growing expectations for Nvidia's AI-related revenue growth. The trend highlights strengthening demand for AI infrastructure and continued optimism surrounding Nvidia's position in the AI semiconductor market. Source: * [YCharts](https://go.ycharts.com/io-fund) Nvidia’s revenue estimates for FY2027 now sit near $561.5 billion, a 123% jump compared to estimates at the beginning of Q2 2025. The largest jump pictured comes in early February 2026, which corresponded with Google, Amazon and Meta releasing their capex guidance for 2026. The combined midpoint of [capex guidance](https://www.cnbc.com/2026/02/06/cnbc-daily-open-amazons-projected-capex-dwarfs-that-of-its-peers-which-have-already-spooked-markets.html) among these names was $505 billion, or around 36% higher than the $371 billion analyst had modeled. Nvidia’s 2027 revenue estimates would then rise by approximately $90 billion in one week, or an over 27% increase. Following higher capex, sales estimates of Broadcom and AMD, also led to strong growth in the following quarters. Over the course of 2026, the next fiscal year estimates for Broadcom and AMD’s sales have increased by 32% and 29%, respectively. Additionally, AMD has raised its server CPU market CAGR forecast to over 50% from now through 2030, up over 2.7X from its 18% CAGR forecast in November 2025. We can also see downstream markets, such as memory and networking, benefiting immensely as capex surpasses expectations. ## Higher Capex Foreshadowed Memory Boom; Networking Boom The memory boom is directly related to the higher capex we’re seeing in 2026 as memory content is rising 2X in AI systems and 4X next year, accounting for more than half of AI chip component costs. Networking also sits downstream as another key beneficiary of rising capex, with both component markets showing a disproportionate flow-through with growth far outpacing capex growth. Looking back to 2023, the memory market bottomed at $89.9 billion, declining (36%) YoY. In 2026, spurred by increasing content in GPU systems and supply shortages driving prices significantly higher, the global memory market is forecast to hit $889.3 billion – or nearly 10X growth in just three years. Current projections for 2027 estimate [the market hitting $1.28 trillion](https://www.trendforce.com/presscenter/news/20260529-13068.html) next year, or a 94% CAGR from 2023. Not only is this above current capex estimates for 2027 at $1.01 trillion, but it also outpaces the four-year capex CAGR of 61% by 33 points. Networking is more fragmented than memory, yet if we look at 2025-2027, we see that optical transceiver growth far outpaced capex growth. From 2025 through 2027, shipments of >800G transceivers are forecast to rise at a 104% CAGR, versus a 57% CAGR for capex, while certain subsegments of the optical networking landscape are projected to see more than 4X growth next year to over $30 billion next year. This illustrates the importance of identifying the beneficiaries that are downstream from capex spend. ## Conclusion Top Wall Street forecasters are now eyeing more than $1 trillion in AI capex for 2027. Given the expected growth rates from giants like Nvidia, Broadcom, along with AMD’s forecasts for the CPU market forecast, this level of spending appears easily obtainable. The real question is how far Wall Street will underestimate the AI buildout this time. However, the most important question for investors is ** who are the beneficiaries?** If we look at previous years, we see that increased capex was correlated to a chip boom (Nvidia, Broadcom), a memory boom (Micron, SK Hynix) and significant gains across lesser-known networking names in the year following an acceleration in spending. Astonishingly, many high-profile hedge funds and Tech ETFs took the easier path with their AI allocations by concentrating in the relative laggards of the AI buildout, which is Big Tech. This creates immense opportunity cost when a tech bull market is led by memory and networking winners, and meanwhile, tech portfolios with significant AuM see their tech positions sit out the bull run. In sharp contrast, the I/O Fund has identified some of the strongest beneficiaries of the AI trade over the past several years – leading to a performance that would rank #1 if we were a hedge fund and #3 if we were a tech ETF. Our latest **90-page Top 20 AI Stocks for Q3 2026 report **offers investors a comprehensive deep dive into the AI stack, mapping out the companies best positioned to capture capex spend. Previous winners identified in the report include **Bloom Energy up 1150% **since our first entry, **Micron up 210%** since our entry a few months ago, **a lesser-known networking stock up 370%** since November. All of this from the team that first identified Nvidia as [an AI stock in 2018](https://www.youtube.com/watch?v=WSaEsYBSjqw&t=1s), up 6700% since our first entry. **Don’t miss out on the AI trade. Subscribe Now.** *Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance, and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA, MU at the time of writing and may own stocks pictured in the charts.* *Leo Miller, AI and Semiconductor Investment Writer at I/O Fund, contributed to this analysis.* **👉🏻 Share with a Fellow Investor** *Help someone else benefit from this insight.* **Recommended Reading:** ### More To Explore ### Newsletter ### AI Capex to Hit $1 Trillion – And Estimates Are Still Too Low Big Tech capex is the driving force behind the AI infrastructure trade, yet Wall Street has repeatedly underestimated the sheer scale of the buildout. Currently, in 2026, the guidance for $732.5 billi ### Token Growth is Surging - Here Are the Beneficiaries The reality of AI demand growth has shattered early estimates for token processing, yet expectations continue moving up and to the right. In the second installment of our token processing series, we e ### AI Token Demand is Shattering Forecasts Total annual token processing is no longer measured in billions or trillions of tokens, but in the quadrillions and beyond. As annual token processing is now tracked in units with 15 trailing zeros, i ### Nvidia and Google Are Crowding TSMC’s N3 Node - Can Intel Fill the Gap? Nvidia is moving its next-generation Rubin GPUs from 4nm to 3nm, yet Google’s latest TPUs are already on N3 and are expected to remain there. Meanwhile, a growing number of AI CPUs from Nvidia, Amazon ### Intel vs TSMC: How CoWoS Packaging Constraints Could Create an Opportunity for Intel Foundry Taiwan Semiconductor (TSMC) is the single, most important company to the AI industry. However, to compete with the incumbent, Intel does not need to beat TSMC at leading-edge manufacturing. It only ne ### Big Tech’s Free Cash Flow is Turning Negative – Who's Next? Big Tech’s AI revenue is accelerating, but free cash flow is moving sharply in the opposite direction. Across Google, Microsoft, Meta and Amazon, capex is rising much faster than operating cash flow a ### Big Tech Earnings Preview: Is AI Monetization Finally Catching Up to Capex? The most pronounced difference between 2026’s tech rally compared to rallies in the past is which companies have been left out of it. The names most associated with the AI trade have hardly participat ### Nvidia, CXL, and the Battle to Improve AI Inference Economics This is Part 2 of our two-part series on AI inference economics. In Part 1 — Why Nvidia's Next AI Battle Is About Tokens per Watt, we laid out why tokens per watt has become the defining metric for in ### Why Nvidia’s Next AI Battle Is About Tokens per Watt As hyperscalers move from building AI infrastructure to monetizing it, tokens per watt helps to reflect if revenue is scaling and if profitability is improving. Offload engines can increase tokens per ### Micron Is Up 900%. Here’s Why the AI Memory Trade May Still Have Room to Run Over the past 10 months, memory chip stocks have gone from being solid beneficiaries of the AI boom to capturing a massively outsized piece of the return pie. The inflection in Micron’s performance de