US hyperscaler companies are projected to spend $800 billion on capital expenditures in 2026, a 94% increase from 2025, driving roughly half of S&P 500 earnings growth through chipmakers and tech suppliers. However, Goldman Sachs forecasts slower capex growth in subsequent years, which combined with rising depreciation charges will dampen AI investment's contribution to earnings. Market analysts warn of structural headwinds including valuation compression from higher interest rates and massive new equity supply from IPOs and insider lockup expirations.
US President Trump sold tens of millions in shares of major tech companies in July, including up to $31 million in Microsoft, $26 million in Amazon, and smaller amounts in Meta, Alphabet, and Nvidia, according to ethics filings. The sales may signal his views on the tech sector amid intensifying US-China competition.
Goldman Sachs reports that AI investment drove nearly half of S&P 500 earnings growth in 2026, with hyperscalers spending $800 billion on capital expenditures, up 94% from 2025. However, analysts warn that slowing capex growth rates, rising depreciation charges, and increased equity supply from IPOs pose structural headwinds to the AI trade's continued outperformance.
Major AI companies borrowed $121 billion last year versus $28 billion annually from 2020-2024, with projections reaching $400 billion next year. Debt now covers 32% of their capital expenditures, up from 9% two years ago, as spending outpaces revenue growth. Bond markets are tightening terms and demanding higher premiums, potentially increasing costs for consumers of cloud services, streaming, and online storage.
Meta launched Muse, a personal AI agent that can send emails, plan travel, and make purchases without specialized training. The agent achieved record adoption with 2.8 million downloads in its first 12 days, outpacing ChatGPT's initial launch and reaching the No. 1 most-downloaded free app on iOS and Android.
A pangram is a word, phrase, or sentence that uses every letter of an alphabet at least once. The most famous English example is 'The quick brown fox jumps over the lazy dog,' historically used to test Telex equipment and now used to display computer typefaces. Perfect pangrams containing each letter exactly once are rare and often require obscure words or abbreviations.
Google has agreed to fund upgrades at Georgia Power's Vogtle and Hatch nuclear plants to increase electricity capacity for its AI data centers, as local opposition to AI facilities mounts over concerns about strain on power grids and rising utility costs. The deal reflects a broader trend of AI companies turning to nuclear power for infrastructure projects, partly due to the Trump Administration's opposition to solar and wind energy development.
Social media discussion on AI infrastructure spending and data center economics. Major tech companies (Microsoft, Alphabet, Amazon, Meta, Oracle) are projected to spend heavily on AI infrastructure in 2026-2027, with Nvidia reporting $89B in data center revenue (up 117% YoY). Investors debate whether the rally is supported by real demand or hype, with focus on stock valuations and power capacity constraints.
Mauboussin and Callahan's analysis shows hyperscaler AI capital returns averaging 23%, above the 8% cost of capital, but with divergent company trajectories and recovery dependent on consensus forecasts proving accurate. Discussions highlight risks in AI infrastructure stocks: capacity monetization timelines, customer concentration among hyperscalers, and competition in custom chip development could materially alter return expectations.
Major tech companies face a $1.1 trillion gap between AI capital expenditures projected to exceed $1 trillion in 2027 and sharply declining free cash flow, forcing greater reliance on debt and equity financing. India's power infrastructure is shifting capex from generation to grid upgrades, while markets show growing confidence in AI investments despite negative near-term cash flows.
Investment in data center and AI infrastructure has surpassed residential housing investment in the U.S., with spending on information processing equipment reaching $752 billion compared to $748 billion for housing in Q2. Hyperscalers are expected to spend over $1 trillion annually on AI infrastructure, driven by companies' belief in high future returns despite rising borrowing costs, while the housing market remains constrained by high mortgage rates and limited supply.
Danny Bate explores how the letter W emerged in the fifth and sixth centuries when Germanic-speaking peoples migrated into the collapsing Roman Empire and needed their names written in Latin script. Latin scribes, lacking a letter for the Germanic /w/ sound, innovatively doubled the letter U to create W, which became standardized in texts across post-classical Europe.
US stock futures showed mixed premarket movement on September 18, 2026, with Mag 7 tech stocks mostly higher led by Alphabet, while cryptocurrency-linked stocks rallied following SEC approval of digital securities trading. Individual movers included Macom gaining 2.1% after a BMO upgrade, Netflix dropping 2.9% on Wells Fargo downgrade, and Xenon plummeting 27% after pausing clinical trials.
Zhang Yiming, founder of TikTok parent company ByteDance, has become Asia's richest person with a $105.5 billion net worth, surpassing Gautam Adani. His wealth surge of $12 billion this month was driven by new company valuations and ByteDance's AI investments, which leverage data from its social media platforms to train advanced models. The AI boom is reshaping billionaire rankings globally, with entrepreneurs like Michael Dell and Larry Page also dramatically increasing their wealth through AI-related ventures.