--- title: The Bigger The Spend, The Harder The Fall? description: "The Point #55: The four largest AI spenders lifted 2027 capex to $1.05 trillion and investors balked. Why services look sturdier than hardware - and the risks building." --- [Skip to content](https://thepoint.ciogroup.com/episode/55#main-content) [![The Point logo](https://thepoint.ciogroup.com/hs-fs/hubfs/ThePoint_logo_WHITE-nav.png?width=150&height=50&name=ThePoint_logo_WHITE-nav.png)](https://ciogroup.com/) Open main navigation Close main navigation [All posts](https://thepoint.ciogroup.com/episode/all) October 29, 2026 # The Bigger The Spend, The Harder The Fall? By  Steven Wieting & David Bailin  ·  5 minute read [Download PDF](https://thepoint.ciogroup.com/hubfs/EP%2055%20-%20Online.pdf) [All Episodes](https://thepoint.ciogroup.com/home) With their earnings reports this past week, the four largest public firms building AI capacity revealed their capital spending plans for 2027 collectively increased to $1.05 trillion (see figure 1). Investors balked. It seems too risky to bet so much on AI. ![Figure 1: Capital Spending of the Four Largest Tech Spenders vs the S&P 500 Semiconductors Share Price Index. Source: CIO Group, Haver Analytics, Bloomberg](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%201.png) But the leaders of the AI revolution won’t stand down. Assuming others spend similarly (some facts below) our tracking estimates for AI capex for the coming year indicate a 48% growth rate after a massive 89% surge in 2026. Alphabet, Amazon, Meta and Microsoft are the “strong hands” in the data center buildout. While Meta chose to dilute current shareholders with an equity offering to fund its record spending, these larger firms use the capacity themselves to support existing profitable operations. This is unlike many of the “neoclouds” who are trying to fill a niche supplying computing power to address excess demand. Yet investors in both the spenders and the beneficiaries of the spending (from hyperscaler to semiconductors to software) saw their share prices plunge. ## Another Round of Tech Investor Soul Searching or “Gut Check” While there are political and geopolitical fears driving near-term risk aversion, tech investors appear to be gripped in another “soul searching” moment, wondering if the return on AI investments will pay off or generate consequential busts that will crush the economy. Investor concerns do not signal a fundamental turn in the economy today. Alphabet and Meta share prices since 2023 can help illustrate. Meta’s shares have fallen as much as 34% in a period when investors doubted its spending would generate future profits. Alphabet has traded off sharply on fears of being displaced by private firms such as Anthropic and OpenAI. Both traded lower on shocks including the 2025 tariffs and Iran war. But both recovered while spending even more. ![Figure 2: Alphabet and Meta Share Prices, Indexed to 100 on Jan 3, 2023. Source: CIO Group, Financial Modeling Prep](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%202.png) The projected path of spending for 2027, like 2026 before it, has only risen, though there are limits to how far it can go, and how far lenders will be willing to finance it before borrowing costs become prohibitive. Shares of semiconductor designers, including Nvidia, are not rising as rapidly as profits. This is because markets rightly believe future growth must eventually slow (see figure 3). However, the restraint from investors is a bullish sign for share price performance heading into early 2027 assuming the AI economy powers through today’s building doubts. ![Figure 3: S&P Semiconductors Y/Y% vs EPS and Estimates Y/Y%. Source: CIO Group, Haver Analytics, Bloomberg](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%203.png) More broadly, advances in AI services have the potential to drive growth for innovators for many years to come. Yet, history suggests shortages of supplies like memory chips to support computing networks won’t follow the same path. Innovation finds ways to avoid shortages of commodity inputs. Supply is flexible and responsive to demand over time. Most importantly, AI service providers are “building factories” that they plan to run profitably. The initial investment is very large, far in excess of annual profits or cash flow, but the operating capacity will last for years and the scale of the spending can be pulled back without interrupting services. Firms that innovate and find new demand are not hit by booms and busts the way hardware makers are. An example of this can already be detected in Anthropic’s early disclosures related to its IPO filing. As figure 4 shows, Anthropic’s revenues are parting ways with its spending on compute capacity (hardware, leased capacity, etc.). Revenue is accelerating faster than its spending on compute. ![Figure 4: Anthropic: Quarterly Revenue vs Compute and Infrastructure Expenses ($ Billions). Source: CIO Group, Anthropic Form S-1](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%204.png) ## Stronger Growth Is Not the Same as Less Risky Almost every objective data source we analyze has suggested a stronger rise in AI hardware demand in 2027. The inflation of input costs and rapid growth rate more clearly define the spending as unsustainable. (As we wrote in June, the 65% pace of spending of the past three years extrapolated to the middle of the next decade would leave AI spending larger than the whole US economy, an impossibility.) While the demand side continues to boom, constraints on the spending are building too. Demand for power and water is a constraint that plays out locally, suggesting the industry must adapt to new political realities. In 2026, credit finally flagged a warning too (see figure 5). The break in lending to CoreWeave, the largest neocloud borrower, is a tiny “blip” within the high yield market. However, the lending constraint will rise as the Fed tightens in the face of rising credit demand from speculative borrowers. ![Figure 5: Yields on CoreWeave’s 2030 Bonds, US High Yield Bonds and the 5-Year US Treasury (Daily). Source: CIO Group, Haver Analytics, Bloomberg](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%205.png) ## Innovation is Constant, Profitability Is Uncertain It is natural that during periods of aggressive innovation, markets will be uncertain in sorting out winners and losers. Early in 2026, markets appeared to fear the “death of software” on AI displacement. Some displacement is certain (when was the last time you purchased a map?), but the idea that software borrowers would be displaced en masse hampered the loan market and private credit lenders in 1Q 2026. In June, software equities underperformed hardware by the widest margin in data going back to 1997, wider even than in the internet boom (see figure 6). Yet since then, the scope of the cybersecurity threat has become clearer, and cybersecurity shares reached a record high this month. ![Figure 6: Software vs Hardware and Semiconductors: Relative Share Price Performance, 12-Month %. Source: CIO Group, Haver Analytics](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%206.png) Anthropic, once public and included in key indices, will almost certainly be classified as software (application software) under the Global Industry Classification Standard. Of course, the most innovative firms are rarely classified well. To this day, Amazon is classified as a retailer. AI services and software remain better placed to survive the market’s boom/bust dynamic in our view. Importantly, our broader asset allocation is filled with risk-mitigating positions to manage our tech exposure (see figure 7). This includes dividend growth equities far removed from the AI trade, energy producers and long-term Treasury securities with depressed values. ![Figure 7: S&P 500 Health Care vs Information Technology: Total Returns, Indexed to 100 on Jan 1, 2026. Source: CIO Group, Bloomberg](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%207.png) ## Fear is Busting Out All Over Signs that the hardware boom will continue in 2027 actually raise the likelihood of a sharper drop at some point. This is despite the fact that a high level of maintenance capital spending for AI infrastructure is almost certain to be a new feature of the economy looking forward. AI is a new kind of productive capital. We need not have a “weak” view of its impact to assess cyclical risks to markets and the economy. As we look ahead to 2027, analysts project even greater spending in 2028 (see figure 8). This would mean a record high share of S&P 500 profits for the semiconductor industry, well above what current share prices indicate. This can keep us tactically bullish in the face of investor fears. ![Figure 8: Semiconductor Share of S&P 500 Operating Profit and Market Cap, with Projections Through 2028. Source: CIO Group, Haver Analytics, Bloomberg](https://thepoint.ciogroup.com/hubfs/243932535/EP%2055%20-%20FIG%208.png) Yet “no tree has ever grown to the sky” and booms don’t end with a whimper. Malcolm Spittler contributed to this report. ## Important Information This material has been prepared by CIO Capital Group, LLC (“CIO Group”) for informational purposes only and does not constitute investment advice or an offer to buy or sell any security or the solicitation of an offer to buy any security or investment advisory service. All opinions are subject to change without notice. Past performance is not indicative of future results. Investment returns may vary significantly over time. Indexes are unmanaged, are not available for direct investment, and do not reflect the deduction of fees or expenses. This material has been prepared without regard to the specific investment objectives, financial situation, or particular needs of any individual investor. CIO Group believes that the information obtained from third party sources contained herein is reliable, but no representation or warranty is made regarding its accuracy or completeness. CIO Group accepts no responsibility or liability (including for indirect, consequential, or incidental damages) for any error, omission, or inaccuracy in such information and for results obtained from its use. Certain information contained herein constitutes ‘forward-looking statements’ which are based on current expectations, estimates, projections, and beliefs. Forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially. Any forecasts, projections, estimates, or market expectations are based on current assumption and should not be considered guarantees of future events or investment results. References to portfolio positioning, asset allocation, sectors, or investment strategies reflect CIO Group's views as of the date of publication and should not be interpreted as recommendations to purchase or sell any particular security or investment strategy. CIO Group is an investment adviser registered with the SEC. Copyright ©2026 CIO Group. All rights reserved. Share [facebook-f icon](http://www.facebook.com/share.php?u=https://thepoint.ciogroup.com/episode/55) [linkedin-in icon](http://www.linkedin.com/shareArticle?mini=true&url=https://thepoint.ciogroup.com/episode/55) [twitter icon](https://twitter.com/intent/tweet?url=https://thepoint.ciogroup.com/episode/55) [pinterest-p icon](http://pinterest.com/pin/create/link/?url=https://thepoint.ciogroup.com/episode/55) [envelope icon](mailto:?body=https://thepoint.ciogroup.com/episode/55) [![CIO Group Logo](https://thepoint.ciogroup.com/hubfs/CIO_Logotype_Neg-120x96.png "CIO Group Logo")](https://ciogroup.com/) ciogroup.com ```json { "@context" : "https://schema.org", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "Steven Wieting & David Bailin", "url" : "https://thepoint.ciogroup.com/episode/author/steven-wieting-david-bailin" }, "dateModified" : "2026-11-02T02:20:52.214Z", "datePublished" : "2026-10-29T17:39:51.000Z", "headline" : "The Bigger The Spend, The Harder The Fall?", "mainEntityOfPage" : { "@id" : "https://thepoint.ciogroup.com/episode/55", "@type" : "WebPage" }, "publisher" : { "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://thepoint.ciogroup.com/hubfs/CIO_Colophon_Neg.png" }, "name" : "The CIO Group" } } ``` ```json { "@context" : "http://schema.org", "@type" : "Article", "author" : { "@type" : "Person", "name" : [ "Steven Wieting & David Bailin" ] }, "datePublished" : "2026-10-29T17:39:51+0000", "description" : "The Point #55: The four largest AI spenders lifted 2027 capex to $1.05 trillion and investors balked. Why services look sturdier than hardware - and the risks building.", "headline" : "The Bigger The Spend, The Harder The Fall?", "image" : "", "publisher" : { "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://243932535.fs1.hubspotusercontent-na2.net/hubfs/243932535/CIO_Colophon_Neg.png" }, "name" : "The CIO Group" }, "url" : "https://thepoint.ciogroup.com/episode/55" } ```