The AI trade is not dead. But the strategy of simply owning the market’s hottest AI names and relying on momentum may be reaching its limits.

That distinction is becoming increasingly important after another violent deleveraging episode. According to Goldman Sachs trader Natasha Tiwana, the firm’s high-beta momentum portfolio fell 12% in one week, while an AI long-short basket lost 10% over five days. Software has now replaced semiconductors as the largest component of Goldman's short-term momentum longs, while the semiconductor-and-AI complex has shifted toward the short side.

The message for investors is less “sell AI” than “AI beta is no longer enough.” The next phase may depend more on earnings revisions, valuation and stock selection—and the first major test arrives almost immediately with NVIDIA Corporation(NVDA.US).

The Number That Matters: Momentum Leadership Is Breaking

One of Goldman's strongest signals is not the absolute decline in AI stocks, but the breakdown in the relationship between previous winners and current winners.

The overlap between stocks leading over the past 12 months and those leading over the past three months has fallen toward multi-year lows. At the same time, many longer-term winners have moved into the short-term loser camp.

That is classic market rotation: capital is not necessarily leaving equities altogether; it is moving between factors and sectors.

Goldman's own public commentary supports this broader interpretation. The firm recently noted that US equity returns are expanding beyond AI, with earnings strength spreading across nine of the S&P 500 index(SPX.US)'s 11 sectors, even as volatility remains elevated in AI-related themes.

This is why the software-versus-semiconductor shift matters. iShares Expanded Tech-Software Sector ETF(IGV.US), which provides concentrated exposure to North American software, cloud and digital-media companies, has become a useful market proxy for whether this new momentum leadership can persist.

Conversely, VanEck Vectors Semiconductor ETF(SMH.US) provides a broad gauge of the semiconductor complex that powered much of the previous AI rally.

But a move into the short side of a momentum factor does not mean semiconductor fundamentals have structurally turned bearish. It means recent price leadership has weakened enough for systematic strategies to treat the group differently.

That distinction matters.

Nvidia’s August 26 Earnings Are the First Major Test

The clearest near-term catalyst is NVIDIA Corporation(NVDA.US), which will report fiscal Q2 2027 results on Wednesday, August 26. Nvidia says its previous quarter produced record revenue of $81.6 billion, including $75.2 billion from Data Center, so expectations remain unusually high.

For the broader AI trade, the question is no longer simply whether Nvidia grows.

It is whether results and guidance are strong enough to reaccelerate earnings expectations faster than the market is de-rating crowded AI exposure.

Goldman argues that the more attractive tactical opportunities may now sit where share prices and EPS trajectories have diverged most sharply—particularly memory/storage and data-center infrastructure.

That makes memory names such as Micron Technology, Inc.(MU.US) especially relevant. If earnings revisions continue rising while valuations remain below earlier AI-cycle highs, the group could become one way for investors to differentiate between fundamental AI beneficiaries and pure momentum exposure.

The same framework applies to data-center infrastructure: investors may want to watch whether demand, orders and margins continue supporting earnings even when the broader AI factor remains volatile.

Software Takes the Momentum Crown—but Confirmation Is Still Needed

Software's new position as Goldman's largest short-term momentum long is striking because the sector had previously been one of the areas most challenged by fears that generative AI could disrupt traditional software economics.

The rotation suggests the market may be reconsidering that thesis.

Falling computing costs increasingly shift the AI debate from “who supplies the infrastructure?” toward “who can monetize widespread AI adoption?” Goldman has separately argued that enterprise AI deployment is accelerating after a relatively slow start.

The next confirmation would be sustained upward earnings revisions across software—not simply stronger share prices.

If iShares Expanded Tech-Software Sector ETF(IGV.US) continues outperforming while earnings expectations improve, the rotation has fundamental support. If prices rise without corresponding revisions, investors may instead be watching another short-term factor trade.

Beyond AI: Banks Are Becoming Part of the Broadening Trade

Goldman is also looking outside US technology entirely.

European and Japanese banks stand out because the interest-rate backdrop can support net interest income while fee income, efficiency and shareholder distributions have also improved.

Japan offers a particularly clear transmission mechanism:

Higher rates → wider loan-deposit spreads → higher net interest income → stronger bank earnings.

Mizuho Financial Group, for example, recently reported a 45% increase in quarterly profit and raised its annual forecast, while its domestic loan-deposit margin widened to 1.26% from 1.10%.

European financials provide another diversification route. iShares MSCI Europe Financials ETF(EUFN.US) holds European banks, insurers and other financial companies and was up more than 16% year-to-date as of August 19, according to BlackRock.

Goldman's thesis is therefore not simply “banks are cheap.” It is that relative earnings durability plus still-moderate positioning could allow financials to absorb capital rotating away from crowded technology exposure.

Gold and Copper Miners: The Hard-Asset Catch-Up Trade

The second major destination is hard assets.

Goldman says its gold-mining basket has already risen sharply in August but remains below its historical peak and trades at a discount to its five-year average valuation. The argument is increasingly tied to the dollar, long-duration yields and demand for portfolio hedges.

VanEck Vectors Gold Miners ETF(GDX.US) offers broad exposure to global gold producers.

Copper presents a different setup.

Copper itself reached a record $14,912 per metric ton on the London Metal Exchange during August's severe physical-market squeeze. Reuters reported that exceptionally tight inventories and concentrated positions pushed the cash premium over three-month copper to as much as $545 per ton.

Yet Goldman argues that copper equities have not fully reflected that strength.

That creates a straightforward thesis:

Tight physical supply + strong copper pricing → stronger miner earnings → potential equity catch-up if risk sentiment stabilizes.

Global X Copper Miners ETF(COPX.US) provides diversified exposure to copper producers, including Freeport-McMoRan, Inc.(FCX.US) and other global miners.

The main confirmation signal is whether copper miners begin outperforming copper itself rather than merely following commodity prices higher.

France Is the Risk Side of the Rotation

Not every diversification trade is bullish.

Goldman also flags French political risk, where sovereign bond markets have already been demanding a larger risk premium while domestic equities have remained relatively resilient.

The key question is whether that disconnect persists.

France's late-August political calendar is becoming denser, including a major MEDEF gathering on August 27. Bank of America has also identified late-August French political events and the 2027 budget process as important catalysts for French sovereign spreads.

For US-listed investors, iShares MSCI France ETF(EWQ.US) can serve as a broad France equity proxy, although it is not equivalent to Goldman's more domestically focused French basket.

A simultaneous widening in French sovereign spreads and weakness in domestic-sensitive equities would strengthen the risk signal.

What Exactly Is the Next Test?

The market's key debate is no longer whether AI remains transformational. It is whether AI-related earnings can keep outrunning expectations after positioning, valuations and financing requirements have become much more demanding.

Goldman estimates that nearly $500 billion of AI-related debt had already been issued in 2026 by early August, highlighting how the buildout is increasingly becoming a capital-markets story as well as a technology story.

That makes the next sequence unusually important:

Near term: Nvidia's August 26 earnings test whether AI infrastructure expectations can reaccelerate.

September: the industry conference calendar should provide fresh signals on demand, memory, networking, data centers and enterprise AI adoption.

Medium term: watch whether earnings revisions broaden into software and other AI adopters while semiconductors stabilize after their momentum unwind.

Outside AI: European/Japanese banks and gold/copper miners need to keep outperforming on fundamentals—not merely because investors are temporarily reducing technology exposure.

The rotation becomes more durable if earnings leadership broadens with it. If Nvidia delivers another major upside surprise and semiconductor earnings revisions accelerate again, the current factor shift could instead prove to be a positioning reset rather than the end of AI leadership.

That is the central distinction investors should watch.

Stocks & ETFs to Watch

Broader read-throughs: Mitsubishi UFJ Financial Group, Inc. Sponsored ADR(MUFG.US) and Sumitomo Mitsui Financial Group Inc Sponsored ADR(SMFG.US) for Japanese banks; Freeport-McMoRan, Inc.(FCX.US) for direct copper-miner exposure.

Disclaimer: This article is for informational purposes only and does not constitute investment advice or a recommendation to buy or sell any securities.