On July 1, 2026 (Eastern Time), the US stock market appeared calm on the surface—the Dow Jones edged down 0.03%, the S&P 500 slipped 0.22%, and the Nasdaq fell 0.66%. Yet beneath the surface, a dramatic revaluation of assets was underway.
The Philadelphia Semiconductor Index plunged over 6%. Micron Technology (MU) and SanDisk (SNDK) both dropped more than 10%, Corning tumbled over 13%, Intel fell 9.03%, AMD slid 6.89%, and TSMC lost 6.98%. Nvidia (NVDA) closed at $197.58, down 1.25%, with a market cap of $4.781 trillion.
However, the catalyst behind this sharp sell-off—Meta (META)—soared 8.81%, closing at $612.91. Microsoft (MSFT) also bucked the trend, rising 3.02% to $384.28.
The immediate trigger for this market divergence was the revelation that Meta is planning to launch a cloud infrastructure business, aiming to sell its surplus AI computing power to external clients. The market quickly interpreted this as a signal that the AI industry narrative is shifting from a "capability race" to a "monetization contest."
$725 Billion in Capital Expenditure and a Fundamental Question
To understand the logic behind this sell-off, we first need to grasp the scale of AI infrastructure investment.
In 2026, the combined capital expenditures of the world’s four tech giants—Meta, Microsoft, Alphabet, and Amazon—are expected to reach about $725 billion, up 77% from approximately $410 billion in 2025. Meta alone has raised its capex guidance to between $125 billion and $145 billion.
But there’s a fundamental difference between Meta and the other three: Microsoft has Azure, Google has GCP, and Amazon has AWS—these massive capital outlays are directly offset by mature cloud service revenues. Meta does not have this advantage. Every dollar Meta previously spent on infrastructure was a pure cost item.
This explains a seemingly paradoxical phenomenon: Meta beat Wall Street’s earnings expectations for two consecutive quarters in 2026, yet its share price is still down about 7% year-to-date. The core market concern: If you spend $135 billion a year building data centers, where’s the payoff?
Meta’s answer is, in essence, that it has bought itself a "put option"—if internal AI monetization succeeds, all computing power is used in-house; if internal demand falls short, the surplus capacity can be sold for revenue. If the bet pays off, it’s a great innovation; if not, at least there’s rental income.
The AI Industry Has Crossed the First Threshold of "Self-Sustaining Cash Flow"
Meta’s plan to "rent out computing power" sparked such a strong market reaction because it touches on the core issue of the AI industry’s cycle shift: moving from "infrastructure-driven" to "revenue realization-driven."
A recent "AI Economy Status Report" from Exponential View, a research institute founded by renowned investor Azeem Azhar, provides key supporting data. As of June 2026, the global generative AI industry (excluding China) had an actual annualized revenue of about $175 billion, with real revenue of approximately $110 billion over the past 12 months.
More importantly, in Q1 2026, the AI industry’s quarterly revenue surpassed the depreciation cost of AI infrastructure for the first time. This means that current AI business cash flow is sufficient to cover the accounting depreciation of servers, GPUs, and data centers. The AI industry has crossed the first threshold of being able to "sustain itself."
However, this doesn’t mean AI investment has entered a "harvest phase." The report estimates that by the end of 2026, cumulative AI-related capital expenditures by global hyperscale cloud providers and emerging AI cloud platforms will reach about $2 trillion. Annual depreciation of AI infrastructure is expected to approach $111 billion in 2026. While quarterly revenues now cover depreciation, cumulative revenues still fall short of offsetting the historical depreciation burden from past investments.
Generative AI revenue continues to grow at around 200% year-over-year—about three times the pace of any previous IT platform upgrade, outpacing the early stages of the internet, cloud computing, and smartphones. According to the revenue growth curve, it took about 180 days to add $1 billion in cumulative AI industry revenue in 2023; now, that process takes less than two days.
From "CapEx Expansion" to "Token Production and Commercialization"
A research report from China Merchants Securities, released July 2, 2026, clearly states: The main theme of the AI industry in 2026 is shifting from "CapEx expansion" to "token production and commercialization." The market focus is moving from "who builds more GPUs and data centers" to "who can produce and monetize more tokens at lower cost and latency."
This shift is already playing out in secondary markets. On July 2, Goldman Sachs commented that it’s too early to call the current AI rally a bubble; the market is still more akin to a profit-driven bull run than a speculative frenzy based solely on valuation expansion. Goldman Sachs continues to favor companies that can generate revenue and earnings growth directly from AI capital expenditures.
Huaan Securities similarly believes that in the first half of 2026, the global AI industry has moved from a "technology explosion phase" to a "rational implementation phase." Computing power supply is becoming more diversified, model capabilities continue to iterate, and the application layer is starting to deliver revenue and profits. The token economy is shifting from a hidden cost to an explicit operational variable.
Revenue Realization Leaders: Microsoft’s $37 Billion ARR and $627 Billion Backlog
When it comes to AI commercialization, Microsoft provides a highly instructive example.
Microsoft’s AI division now boasts an annualized revenue run rate (ARR) of $37 billion. Even more impressive, its remaining performance obligations (i.e., contract backlog) have soared 99% to $627 billion. This massive backlog ensures a highly visible revenue stream for years to come as enterprises lock in cloud services.
Despite Microsoft’s stock dropping as much as 22% earlier this year, its core business engine remains robust—Azure cloud service revenue continues to grow at a 40% pace. At current valuations, Microsoft’s forward P/E ratio is about 22x, a significant discount to its 10-year average of 31x. The latest surveys show 35 analysts rate Microsoft a "buy," with an average target price of $562.10.
Microsoft’s case demonstrates: The market is punishing capex models that "only spend without earning," while rewarding business models that can turn AI investment into verifiable revenue.
Commercialization Accelerates at the Application Layer
The most direct manifestation of the AI industry’s cycle shift is the acceleration of commercialization at the application layer.
Barclays estimates that AI industry ARR was $44 billion at the end of 2025 and is expected to reach $200 billion by the end of 2026. Citigroup projects that global AI revenue for 2026–2030 has been sharply revised up from $2.8 trillion to $3.3 trillion.
In the enterprise AI market, leading companies are seeing especially strong revenue growth. Anthropic expects annualized revenue to reach $26 billion in 2026, while OpenAI’s annualized revenue has already surpassed $25 billion. HSBC predicts that from 2026 to 2030, B2B AI industry revenue forecasts have been raised by 74%, mainly driven by the rise of agent-based AI and the ongoing expansion of enterprise application scenarios.
According to the latest IDC survey, by 2026, 72% of global enterprises have deployed AI agents in production, and 51.6% have embedded agents into core business processes. AI agents are becoming the "new entry point" for enterprise software.
The application layer’s divergent paths are also becoming clearer: On one end are consumer-facing (C-end) products, tasked with educating the market and competing for user entry points; on the other are enterprise-facing (B-end) services, which are beginning to take on more direct commercialization goals and are gradually becoming the key profit centers of the industry. Enterprises purchase AI services to improve business outcomes—lowering costs, boosting efficiency, optimizing processes, and enhancing decision quality.
Rebuilding the Valuation System: From "Growth Expectations" to "Profitability"
The shift in the AI industry cycle ultimately points to a deeper change: the restructuring of how tech companies are valued.
In the "infrastructure-driven" phase, the market paid a "growth premium" for AI concept stocks—whoever built bigger GPU clusters and spent more on capex got higher valuations. Nvidia’s valuation surge in 2024–2025 was the ultimate expression of this logic.
But in the "revenue realization-driven" phase, valuation anchors are shifting from "capex scale" to "revenue quality and sustainable profitability." Meta’s plan to rent out computing power is viewed as a positive, not a negative, precisely because it demonstrates a commitment to financial discipline over massive capex.
This transition was vividly reflected in the July 2 market action: hardware stocks plunged, while application and platform stocks rallied. Microsoft rose, Palantir (PLTR) jumped 7.77%, and Meta soared, creating a stark contrast with the chip stock collapse. Capital is flowing from "the shovel sellers" to "those digging for gold with shovels."
The AI narrative is shifting to performance-driven results. Several public fund managers believe that as the industry enters the revenue realization phase, investment logic is moving from valuation-driven to performance-driven.
Conclusion
The market turbulence on July 2, 2026, was not a random sector rotation, but a concentrated signal of the AI industry’s cycle shift.
From Meta’s "computing power rental" to AI’s quarterly revenue surpassing depreciation costs, from China Merchants Securities’ "token production and commercialization" theme to Microsoft’s $627 billion backlog—all signs point to the same conclusion: The AI industry is moving from the first phase of "infrastructure-driven growth" to the second phase of "revenue realization-driven growth."
In this new era, the core market question is no longer "Who built the biggest data center?" but "Who can turn AI capabilities into sustainable cash flow?" The valuation system is already being rebuilt: hardware valuations are under pressure, while revenue validation at the application and platform layers is becoming the new pricing anchor.
For investors, this signals a fundamental shift in AI investment logic—from chasing the scale of capital expenditure to tracking the quality of revenue realization. AI’s "money-making ability" is becoming the most crucial pricing variable in this industry cycle.
FAQ
Q1: What does it mean that the AI industry’s "quarterly revenue surpassed depreciation costs for the first time"?
This marks the key transition for the AI industry from the "cash-burning phase" to the "self-sustaining phase." As of Q1 2026, cash flow from AI businesses can now cover the accounting depreciation costs of servers, GPUs, and data centers. However, there’s still a gap before recouping all historical investments—cumulative capex stands at about $2 trillion, with annual depreciation around $111 billion.
Q2: Why did Meta’s sale of computing power trigger a semiconductor stock sell-off?
The market interpreted Meta’s move as a sign that AI infrastructure capex may have peaked. If hyperscalers start selling surplus computing power instead of buying new hardware, the supply-demand balance for GPUs, memory chips, and other hardware could reverse. This directly triggered a revaluation of future earnings expectations for semiconductor stocks.
Q3: What does the shift in AI investment logic from "infrastructure" to "application layer" mean?
It means the market’s pricing anchor is shifting from "capex scale" to "quality of revenue realization." The valuation premium for hardware is narrowing, while companies at the application and platform layers that can turn AI capabilities into verifiable revenue are being repriced. China Merchants Securities calls this the main theme shift from "CapEx expansion" to "token production and commercialization."
Q4: Which types of companies are most likely to benefit in the second phase of AI commercialization?
Companies with clear AI revenue models are best positioned to benefit. These include: cloud platforms with large enterprise customer bases (like Microsoft Azure), software companies that can embed AI into core business processes, and leading AI firms with proven large-model commercialization paths (such as OpenAI and Anthropic). B2B scenarios are viewed as the main battleground for large-scale AI application, thanks to stronger payment capacity and verifiable results.
Q5: What are the implications of the second phase of AI commercialization for the crypto industry?
AI and crypto are becoming increasingly intertwined. The Gate platform has launched the "Gate for AI" infrastructure layer, integrating AI into core areas such as trading, risk management, and data analysis. AI agents are shifting from information retrieval to executing economic activities—calling paid APIs, conducting on-chain transactions, and purchasing computing resources. The logic of AI commercialization applies to crypto as well: whoever can turn AI capabilities into verifiable on-chain revenue will be revalued in the new cycle.
