The AI Capital Shift: From Attention to Intelligence

Value is migrating from consumer attention to AI infrastructure, driving a massive hyperscaler capital expenditure boom in 2026.

/ Article
The AI Capital Shift: From Attention to Intelligence
Photo by imgix on Unsplash

The Monthly Thesis: A New Center of Gravity

The tech world is undergoing a profound reordering, a seismic shift in where true value resides. We are witnessing the twilight of one era and the dawn of another, marked by the evolution of tech leadership acronyms. The market’s center of gravity is moving, decisively, from capturing human attention to building artificial intelligence.

First, there was FAANG: Facebook (now Meta), Apple, Amazon, Netflix, and Google. This cohort defined the “attention economy,” dominating consumer hardware, social media, and streaming content. Their success hinged on engaging billions of users. Then came the Magnificent 7, or “Mag-7,” a term coined by Bank of America in 2023. This group, comprising Alphabet, Amazon, Apple, Meta Platforms, Microsoft, Nvidia, and Tesla, reflects the mega-cap index-weighting story of the current public market. It drops Netflix, adding Microsoft, Nvidia, and Tesla, signaling a broader industrial and enterprise reach beyond pure consumer plays.

Now, we are entering the MANGOS era: Meta, Anthropic, Nvidia, Google, OpenAI, and SpaceX. This emerging cohort represents the AI-native next wave. It emphasizes foundational reasoning models, silicon dominance, and orbital data backbones. Crucially, MANGOS includes private AI-natives like Anthropic, OpenAI, and SpaceX, which the public Mag-7 index cannot yet hold. The editorial thesis is clear: value is migrating along this chain, from FAANG’s legacy of attention capture, through the Mag-7’s diversified mega-cap dominance, to MANGOS’s focus on artificial intelligence. The overlap is the story. Nvidia, Google/Alphabet, and Meta sit in both the Mag-7 and MANGOS, embodying this critical transition.

The CapEx Scorecard: Defensive Survival vs. Offensive Dominance

The backdrop to this narrative is a massive AI Capital Expenditure (CapEx) boom. Wall Street, in a complex dance, fears short-term free-cash-flow compression while simultaneously rewarding long-term hyperscaler growth. The numbers are staggering. The top five hyperscalers are projected to increase their combined CapEx from approximately $256 billion in 2024 to an estimated $443 billion in 2025, and then to a colossal $602 billion in 2026. This represents a 36% year-over-year increase for 2026. Approximately 75% of this aggregate hyperscaler CapEx in 2026, or about $450 billion, is earmarked for AI-related infrastructure.

Amazon, for instance, has raised its full-year 2026 capital expenditure forecast to approximately $220 billion, up from prior guidance of $200 billion, primarily due to elevated memory chip costs. Despite this immense spending, Amazon CEO Andy Jassy stated the company will still fall short of meeting customer demand in both 2026 and 2027. Its AWS contracted backlog jumped by $132 billion in a single quarter, reaching $496 billion, with commitments extending into 2028.

Alphabet, Google’s parent company, plans to spend between $175 billion and $185 billion on capital expenditures in 2026, more than doubling the $91 billion spent in 2025. Some reports indicate Google’s CapEx could even reach $205 billion for fiscal year 2026. This heavy investment targets Google DeepMind, its AI research lab, and Google Cloud expansion. The sheer size of this projected CapEx has, at times, disturbed investors, despite strong overall results.

Microsoft is also accelerating its AI infrastructure expansion. The company expects approximately $175 billion in calendar-year 2026 capital spending, after accounting for a change in data center lease classification. Microsoft added 88 data centers in fiscal 2026 and another 1 GW of capacity in the fourth quarter. Demand for Azure continues to exceed available capacity, prompting Microsoft to shorten GPU deployment times by nearly 50%.

Nvidia, a key enabler of this boom, saw its data center revenue grow 92% year over year to a record $75 billion in the fiscal first quarter of 2027. Its data center segment now accounts for over 90% of its total revenue. CEO Jensen Huang has expressed confidence in Nvidia’s ability to achieve at least $1 trillion in cumulative Blackwell and Rubin chip sales from 2025 through 2027.

The combined spending of Amazon, Alphabet, Meta, and Microsoft is projected to be around $725 billion in 2026, a 77% increase from 2025. This represents the largest single-year infrastructure investment cycle in technology history. While some investors have punished heavy CapEx plans, favoring free cash flow, others reward companies demonstrating clear monetization from these investments.

Server Rack
Photo by Kevin Ache on Unsplash

The Status Table: Tiering the Tech Evolution

| Tier | Market Sentiment
The following is an independently produced analysis and does not constitute financial advice.

| Tier | Market Sentiment to building AI. | Tier | Market Sentiment

Works Cited