Mag 7 earnings week arrives: Can $600 billion in capital expenditures support the AI narrative?

One of the most intensive earnings weeks in history has arrived as expected, and the AI capital expenditure narrative is facing a key stress test.

This Wednesday, Alphabet, Microsoft, Amazon, and Meta will simultaneously release their first-quarter earnings reports, with Apple following on Thursday. Current consensus expectations have fully priced in over $600 billion in combined capital expenditures for these four companies by 2026.

Goldman Sachs Delta One Head Rich Privorotsky points out that the core issue for the market isn’t the strength of demand—demand is clearly strong—but whether capital expenditures can continue to increase. If capital expenditures remain flat, amid rising investment costs, this essentially signals a slowdown, which will directly challenge the current AI market narrative.

In this market structure dominated by the logic of AI capital expenditures, semiconductors have become the most direct beneficiaries. The sector is up 42% this year, far outpacing Mag 7’s roughly 2% rise; excluding Nvidia, Mag 7’s increase is only about 0.2%. However, Goldman Sachs also warns that the upside surprises from AI spending are "almost the entire game." Semiconductor sector valuations have risen to a 60x P/E, risk-reward is deteriorating, and technicals are turning from tailwind to headwind.

Capital Expenditure Expectations: Consensus Fully Priced In

The current race for computing power in the tech industry is fueling what Goldman Sachs calls the "Token Maxing" effect: engineering teams at major companies are competing to consume as much computing resources as possible. The perception that "insufficient spending equals career risk" has created a distorted incentive, driving companies to aggressively spend even inefficiently.

Top tech stocks have pledged over $740 billion in combined AI capital expenditures by 2026. Currently, consensus models have incorporated more than $600 billion in capital expenditure for Amazon, Microsoft, Meta, and Alphabet combined for 2026. Meanwhile, constraints have spread across multiple segments: shortages are evident in electricity, CPUs, GPUs, copper, and even engineering talent.

Goldman Sachs says that in terms of the current fundamentals, it’s hard to argue: "Supply is limited in multiple verticals, EPS is accelerating, and the market narrative (right or wrong) is that we’re on the cusp of the most important technological breakthrough in decades." However, the market is indefinitely projecting upside—more computing power, greater intelligence, ultimately leading to AGI. This week’s earnings reports will provide crucial reference for such expectations.

Semiconductors: Rally Has Reached Critical Levels

The semiconductor sector’s rally has evolved into a self-reinforcing cycle: expectations and prices drive each other, and the trend has shifted from "fundamental support" to "self-fulfillment." The sector’s P/E has risen to 60x, and panic-driven chasing continues to accumulate. According to Goldman Sachs, this is the area where "convexity risk is beginning to emerge"—parabolic rallies always have an end.

Within the sector, leadership has quietly shifted. CPUs and analog chips have become the main drivers, while the usual leader, GPUs, are relatively lagging. Goldman Sachs suggests that the current breakout "looks like an upward breakout, but the trading characteristics are more akin to a short squeeze"—positions, flows, and forced buying are all driving the market to chase right-tail risk.

From an earnings perspective, Micron Technology alone contributed more than half of the recent upgrade to S&P 500 EPS, with its quarterly results and guidance far beating consensus, and the expectation for "unlimited extension" of AI memory chip demand nearly doubling consensus valuations. The breadth of S&P 500 earnings upgrades is narrowing, with median company consensus essentially unchanged; the extreme concentration of earnings momentum is an undisputed fact.

Hyperscale Cloud Vendors: Divergence in Shareholder Logic

Semiconductor stocks and hyperscale cloud vendor stocks are showing sharply different market responses to the same wave of capital expenditure, revealing mismatched interests among two types of investors. Rich Privorotsky bluntly states that semiconductor stocks are keenly enthusiastic about capital spending growth—as these expenditures go straight into their pockets—but shareholders of hyperscale cloud vendors historically have not seen increased returns from more capital spending.

"Last week we heard from suppliers (semiconductor companies); this week, it’s the spending side (hyperscale cloud vendors), whose narrative is more complex," says Rich Privorotsky.

This divergence is already reflected in this year’s market performance: Mag 7’s overall gain is about 2%; excluding Nvidia, it’s roughly flat, starkly contrasting the semiconductor sector’s 42% gain. Hyperscale cloud vendors are investing heavily in the AI era, but there remains considerable uncertainty as to when and how these investments will translate into shareholder returns.

Corporate Buybacks: Structural Demand Provides a Floor

As market sentiment faces tests, US corporate stock buybacks are providing a layer of structural support for equities. Buyback authorizations have hit historic records as blackout periods end; total buyback demand this year is expected to exceed $1 trillion.

This buying is insensitive to price and is purely structural demand—companies are aggressively buying back shares at historic highs, mechanically shrinking the float and boosting EPS. For equities trading at high levels, this is a passive force that cannot be ignored.

Market Structure: Non-AI Supply Chain Means Outsider

Rich Privorotsky’s assessment of overall market risk-reward has become cautious: "It’s hard not to respect AI capital’s power, but the speed of the rally is extreme. Almost all surprises relative to expectations come from AI spending—that’s the whole game."

Outside the AI narrative, crude oil and refined product prices are consuming market attention, European equities are lagging, and market polarization is extreme. Goldman Sachs sums up: "You’re either in the AI supply chain, or you’re not part of this rally. From here, risk-reward has worsened, and technicals are turning from tailwind to headwind."

This concentration trend also extends to emerging markets: semiconductor-driven earnings growth is reshaping the entire structure, EMs are becoming more concentrated rather than diversified. For global investors, regardless of where you are, being part of the AI supply chain or not is increasingly the most crucial boundary for returns.

Risk warning and disclaimerThe market involves risk, investment needs caution. This article does not constitute personal investment advice, nor does it take into account an individual user’s unique investment goals, financial situation, or needs. Users should consider whether any opinions, views, or conclusions here fit their specific circumstances. Invest accordingly and assume responsibility.