Goldman Sachs on AI Trading: Risks of "AI Infrastructure" in the Second Half of the Year, "Losers" in "AI Applications" Struggling to Recover in the Short Term

As AI capital expenditures surge and valuations become increasingly expensive, Goldman Sachs warns the market: the real risk often appears at the moment when “growth begins to slow down.”

On February 24, Goldman Sachs Global Investment Research stated in its strategy report “The Broadening and Narrowing of AI Trading” that recent AI trading volatility has risen significantly, driven by two opposing forces pulling the market: on one side, tech giants’ capital spending continues to “beat expectations,” while on the other side, investor concerns about “AI disrupting traditional industry profit pools” are rapidly intensifying.

Fueled by strong capital expenditure guidance, storage chip concept stocks have risen an average of 55% so far this year; meanwhile, software stocks have plummeted 24% due to panic over “AI disruption.” The same “AI theme” shows almost opposite trends at different stages.

Goldman Sachs divides this intense AI trading volatility into four stages, with current stock price trends already diverging sharply:

• Stage 1 (Leading computing power providers, like NVIDIA): Facing questions about “excessive profitability,” recent earnings expectations have been sharply raised, but stock prices have stagnated, showing a disconnect.
• Stage 2 (AI infrastructure, such as storage, devices, servers): Driven by strong capital expenditure guidance from tech giants, stocks have continued to soar, with storage stocks up 55% year-to-date.
• Stage 3 (AI application enablement, such as software services): Due to market fears of traditional business models being disrupted by AI, these stocks have experienced panic selling, with software stocks down 24% this year.
• Stage 4 (AI productivity improvements in non-tech industries): Since actual financial returns remain unclear, stock prices have been consolidating recently.

Faced with this extreme divergence, the report indicates that both the currently soaring “infrastructure winners” and the plunging “application losers” are lurking with their own risks.

Capital expenditure growth is nearing its peak, and the “valuation killing” risk in AI infrastructure is approaching

The market first needs to digest an expectation: a further “upgrade” in capital expenditure outlook.

According to Goldman Sachs’ consensus estimates, by 2026, tech giants’ AI capital spending will reach $667 billion, which is $127 billion higher than at the start of Q4 earnings season, with a year-over-year growth rate of 62%.

On the flip side, this sharp upward revision in capital expenditure forecasts squeezes free cash flow.

The report emphasizes: “Super cloud providers’ capital spending is heading toward exceeding 90% of operating cash flow this year, even higher than during the dot-com bubble.” More specifically, Goldman Sachs estimates that in 2026, capital expenditure will account for 92% of tech giants’ operating cash flow.

To fill this huge funding gap, giants are forced to significantly cut shareholder returns. In 2025, overall stock buybacks by these giants decreased by 15%; cash flow allocated for buybacks dropped sharply from 43% at the start of 2023 to just 16%. Meanwhile, Oracle and Google are increasingly tapping into bond markets.

Goldman Sachs expects that the absolute value of capital expenditure this year still has room for upward revision. Since Oracle and Microsoft’s fiscal years end in May/June, upcoming Q2 earnings reports could serve as catalysts for further upward adjustments.

However, Goldman Sachs warns that the core risk is not in the “absolute amount” but in the “growth rate.”

“We expect consensus estimates for super cloud providers’ capital expenditure to have some mild upside potential, but we still expect the growth rate to peak later this year.”

This slowdown in growth rate will become the Achilles’ heel for AI infrastructure stocks.

H2 Risks in “AI Infrastructure”: Slowing Spending and the “Over-Profitable” Trap

Goldman Sachs emphasizes: “Once the growth rate of capital expenditure slows, the revenue growth and valuations of some AI infrastructure stocks will become extremely fragile.”

The logic is straightforward: orders, revenue, and profits in the infrastructure chain are highly sensitive to the pace of capital expenditure; when the market shifts from “accelerating each quarter” to “still growing but no longer accelerating,” the most vulnerable part of valuations is often the “growth premium.”

Goldman Sachs openly states that many AI infrastructure-related sectors have experienced significant valuation multiple expansion over the past few years, but historical experience shows that investors tend to assign lower multiples to companies with slowing growth.

This is the core meaning behind the report’s theme of “killing valuations”: Even if profits are still growing, once the market starts worrying about unsustainable growth, multiple compression can offset the support from earnings upgrades.

Among the detailed sectors listed, manufacturing equipment, servers and networks, wafer foundries and IDMs, power and utilities generally trade above their five-year averages.

Goldman Sachs believes that the current “new bottleneck” within infrastructure is concentrated in the memory segment.

The report states that major memory stocks (like Micron, Western Digital, SK Hynix, Samsung) have risen an average of about 145% since early Q4 2025, with an average increase of 55% this year. Goldman attributes most of this rise to strong demand and price hikes improving profitability.

They also note that the forward P/E ratio for memory stocks is around 12x, which is below the broader market and also below their five-year average, seemingly not “expensive.”

But Goldman Sachs issues a warning using NVIDIA as an example: When the market begins to worry that companies are “over-earning,” stock prices may no longer follow earnings upgrades.

From late 2022 to mid-2023, NVIDIA’s stock price and earnings grew in tandem by 12 times, with valuation multiples remaining roughly stable. But recently, the logic has changed.

Goldman Sachs points out: “In the past five months, despite NVIDIA’s forward earnings expectations being raised by 37%, its stock price has remained essentially flat.”

They summarize this phenomenon as a market psychology of “over-earning”: when a company performs too strongly at the cycle’s peak, it can ironically trigger concerns about increased competition and demand sustainability, ultimately leading to “profits still strong, but valuation multiples contracting.”

For trading, this means: Even if infrastructure chain performance continues to materialize in the short term, investors will become more selective about “second derivatives of growth” and whether multiples can still expand.

Short-term Divergence of Tech Giants: Focus on “Returns,” Not Capital Expenditure

Goldman Sachs believes that in the near term, the divergence in returns among tech giants will persist.

Because in the first half of 2026, when quarterly growth in capital expenditure stabilizes, market attention will shift to “whether AI investments are actually paying off.”

The report provides a clear comparison: the free cash flow yield of tech giants is about 1%, at its lowest in history; while the rest of the S&P 500 companies yield around 4%.

As free cash flow weakens and conversion rates decline, capital will naturally seek alternatives. Goldman Sachs straightforwardly states, “Investors are increasingly reallocating funds elsewhere.”

AI Application Layer: A “Very Thin Line” Dividing Winners and Losers

If the infrastructure layer’s dilemma is “how fast can capital expenditure grow,” the application layer’s dilemma is “who will be disrupted and who can capture new revenue.”

Goldman Sachs judges that the expansion of AI trading into the application layer is a natural progression of technological development: after building the infrastructure, value creation shifts from “selling shovels” to “transforming business models,” and by reshaping profit pools, early investments are recouped.

But this also makes stock market outcomes more “micro-focused.” Goldman Sachs emphasizes that future decisions will rely more on company-level assessments, such as competitive positioning, entry barriers, and pricing power.

A sentence in the report captures the core uncertainty of the application layer:

“In an environment where the final competitive landscape remains uncertain, the line between a company being seen as an AI revenue ‘winner’ and being feared for ‘disruption’ is very thin.”

One direct consequence is that investors currently do not assign overly high valuations to many listed companies for “additional AI-driven revenue.”

Goldman Sachs states, “Contrary to our expectations, investors are almost not pricing in upside potential for AI-driven revenue increases in listed companies; instead, the most attention is on private companies’ AI applications.”

The report lists several private companies’ product developments: Anthropic’s Claude Cowork tool (with legal, HR, and business service plugins); Insurify’s price comparison app within ChatGPT; Altruist’s tools for personalized tax strategies for wealth management clients.

These cases reinforce a market concern: even if AI generates new demand, the additional revenue may not accrue to listed companies.

Why “Losers” Struggle to Rebound Short-term: Disruption Concerns Are Hard to Falsify with “Short-term Performance”

On the flip side, the narrative of disruption has a damaging effect on valuations.

Goldman Sachs notes that recent weeks’ market focus has been on “AI disruption risks.”

The report states that software stocks have fallen about 23% over the past six weeks, and “despite short-term earnings remaining resilient, investors are increasingly questioning the industry’s long-term growth prospects.”

Goldman Sachs offers a very clear judgment: “Concerns about AI disruption are unlikely to be disproved in the short term.”

They further point out that for companies already labeled as “potentially disrupted by AI,” stock prices can only stabilize if earnings first stabilize; but “this disruption uncertainty is unlikely to be resolved in the near term.”

Goldman Sachs details the conditions under which “application layer losers” will find it hard to rebound: “Investors will need either multiple quarters of evidence proving business resilience or a significant valuation discount relative to the broader market before re-engaging at scale.”

This explains the current awkwardness in software and related sectors: short-term financial reports may look fine, but the market is trading on “whether long-term profit pools will be reallocated.”

Goldman Sachs quantifies the disruption risk with two indicators: exposure to AI automation and asset strength

Regarding how to assess “who is more vulnerable to disruption,” Goldman Sachs provides two vectors (and emphasizes that other dimensions like regulatory barriers and market power also matter).

First, exposure of the workforce to AI automation.

Goldman Sachs notes rising concerns about white-collar job displacement.

They collaborated with economists to estimate the proportion of each company’s wage expenses exposed to AI automation, using the “labor cost/revenue” ratio as an indicator.

Goldman Sachs warns that this metric is a “double-edged sword”: AI can both improve efficiency and replace jobs.

But in trading, over the past six months, markets have rewarded industries with “low exposure” and penalized those with “high exposure.”

Second, tangible asset intensity.

Goldman Sachs uses the ratio of “(assets - cash - intangible assets) / revenue” to measure asset intensity, constructing industry-neutral, equal-weighted baskets.

They observe that companies with heavier assets have recently outperformed those with lighter assets, with the outperformance exceeding what macro factors typically explain.

Similarly, manufacturing companies have outperformed service-oriented firms.

For investors, these two indicators suggest that the market is using them as “alternative measures of moats/entry barriers” to hedge against uncertainties in the application layer, rather than simply favoring asset-heavy companies.

Three Catalysts: Goldman Sachs Bets on a Turning Point in H2 2026

Goldman Sachs believes that for tech giants to regain market leadership, three catalysts are needed.

Their baseline view is that these catalysts are “more likely to occur in the second half of 2026.”

First, AI revenue must accelerate. The market response to earnings seasons has already shown that when revenue growth exceeds expectations (e.g., Meta’s 10% surge), investor confidence in AI investments quickly recovers.

Second, the visibility of free cash flow (FCF) bottoming out due to slowing capital expenditure growth. Goldman Sachs believes that once cash flow signals a bottom, the market may start to price stocks based on profitability rather than cash flow, reducing valuation volatility.

They explain: “Slowing capital expenditure growth will give investors hope that free cash flow is bottoming out and rebounding. This will prompt a re-pricing based on earnings capacity.” Currently, the giants’ forward P/E ratio of 24x is only at the 14th percentile over the past decade, making valuations very attractive.

Finally, the fading of macroeconomic tailwinds. Goldman Sachs economists expect the US economy’s cyclical acceleration to peak mid-year and decline in the second half. When macroeconomic benefits diminish, funds will inevitably flow back into these long-term, highly certain tech giants.

This insightful content is from Chasing Wind Trading Platform.

For more detailed analysis, real-time insights, and frontline research, please join the 【Chasing Wind Trading Platform - Annual Membership】.

Risk Disclaimer and Terms of Use

Market risks are inherent; investments should be cautious. This does not constitute personal investment advice and does not consider individual user’s specific investment goals, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions herein are suitable for their circumstances. Invest at your own risk.