Overview

In late June and early July 2026, Morgan Stanley’s Chief Investment Strategist Michael Wilson and his team issued a research note that has become one of the most debated calls in the technology investment landscape. The note argues that momentum is fading in semiconductor stocks and that investors should rotate capital into AI hyperscalers—Microsoft, Amazon, Alphabet, and Meta—as well as other laggard sectors. This call arrives at a moment of extreme divergence within the AI trade: the Philadelphia Semiconductor Index surged approximately 88% in the second quarter of 2026, its best quarterly performance on record, while the Magnificent Seven stocks collectively lost roughly $2.3 trillion in market value during June alone [17][18]. Wilson’s thesis is not a blanket dismissal of AI infrastructure but a tactical and potentially structural reassessment of where value resides in the AI supply chain. This report examines the specifics of Morgan Stanley’s argument, compares the valuation metrics of the key chipmakers and hyperscalers, and assesses whether the recommended rotation represents a short-term trading opportunity or a fundamental repricing of AI infrastructure investments.

Morgan Stanley’s Rotation Thesis

Authorship and Timing

The research note was authored by Michael Wilson, Morgan Stanley’s Chief Investment Strategist, and his team. It was published in late June 2026, with widespread media coverage appearing on July 6, 2026, including reports from Yahoo Finance, CNBC, and CoinDesk [3][1]. Wilson also gave a CNBC interview around June 30, 2026, in which he elaborated on the thesis [3]. The timing is significant: it follows the Philadelphia Semiconductor Index’s record-breaking 81–88% surge in Q2 2026 and a sharp pullback in early July, during which the index fell nearly 14% from its June record high [3][18].

Core Argument

Wilson’s central argument is that momentum in semiconductor stocks is exhausted and that investors should shift toward AI hyperscalers and other underperforming sectors. He specifically stated that he “favors hyperscalers over semiconductor-related stocks in the near term,” citing their strong core businesses as a key attraction [3]. The rotation, he noted, is unfolding “in a choppy/weaker equity market overall” [3]. The call is not a bearish view on AI itself but a relative-value and momentum-driven repositioning within the AI ecosystem.

Supporting Data and Analysis

Morgan Stanley’s note draws on several data points to support the rotation thesis:

• Semiconductor Index Performance and Reversal: The Philadelphia Semiconductor Index (SOX) fell nearly 14% from its record high set in June 2026, though it remained 123% higher than its level in September 2025 [3]. The index experienced a sharp 6.3% single-day decline on July 1, 2026, and a 7.9% weekly drop, with the Cboe Semiconductor ETF Volatility Index rising 83% year-to-date—the largest annual increase ever [18][11].
• Hyperscaler Resilience: A UBS basket of hyperscaler stocks was down only 2% over the same period, significantly outperforming semiconductors on a relative basis during the pullback [3].
• S&P 500 Target Maintained: Wilson maintained his S&P 500 year-end target of 8,000, implying roughly 7% upside from early July 2026 levels, suggesting the rotation is about relative performance rather than an outright market call [3][1].
• Magnificent Seven Underperformance: The Magnificent Seven lost a combined ~$2.3 trillion in market value in June 2026, with Microsoft falling 20%, Nvidia ~13%, and Apple and Amazon ~8% each [17]. This underperformance created a valuation gap that Wilson’s team views as an opportunity.
• Chip Sector Valuations: The semiconductor index traded at 26 times forward earnings, above its 10-year average of 19 times, though not at extreme levels historically [18]. However, individual names like Nvidia traded at 18 times forward earnings—its cheapest since 2018—and Micron at 8 times, indicating that the valuation argument is more about the sector’s extended momentum than uniformly stretched multiples [18].

Catalysts for the Rotation

Wilson’s thesis identifies several specific catalysts that could drive the rotation from chipmakers to hyperscalers:

• Momentum Exhaustion in Semiconductors: After the SOX’s record Q2 rally, Wilson sees momentum fading. The index’s wild daily swings and the surge in volatility suggest that the easy gains have been made and that the risk-reward profile has deteriorated [18][11].
• Hyperscaler Core Business Strength: Wilson argues that Microsoft, Amazon, and Meta possess strong underlying core businesses—Azure, AWS/retail, and advertising, respectively—that provide a valuation floor. This contrasts with pure-play chipmakers, whose fortunes are tied entirely to the AI capital expenditure cycle [3].
• Potential Softening of Capex Expectations: Wilson expects hyperscalers may soften their AI spending guidance given recent stock underperformance. This is a nuanced view: he sees the potential for spending pullbacks as a catalyst for hyperscaler stocks to rally, because it would ease return-on-investment concerns, even though it would be negative for chip stocks [3].
• Meta’s Cloud Pivot: On July 1, 2026, Meta announced plans to launch a cloud infrastructure business to sell excess AI computing power, directly competing with AWS, Azure, and Google Cloud. Shares surged over 8–10% on the news, as it addressed investor concerns about massive AI spending without clear monetization [20][10][11]. CEO Mark Zuckerberg had previously indicated at the company’s annual meeting in May that a cloud business was an option if they overbuilt on data centers [20]. Jefferies analyst Brent Thill rated Meta a buy, calling the cloud plan a “strategic value creation option” [20].
• Broadening Market Rally: Wilson sees the rotation extending beyond hyperscalers to consumer discretionary, transport, and biotech sectors, indicating a broader shift in market leadership [3].
• Cybersecurity as an AI Beneficiary: A Wall Street Journal report that Zhipu AI’s open-weight GLM-5.2 model now rivals Anthropic’s restricted Mythos model at finding software vulnerabilities drove investors to expect higher cybersecurity spending, benefiting names like Palo Alto Networks and CrowdStrike, which hit all-time highs [9].

Risks and Counterpoints

Wilson’s thesis is explicitly contrasted with JPMorgan’s opposing view, creating a clear debate on Wall Street:

• JPMorgan (Mislav Matejka): Recommends buying the chip pullback, arguing that the AI-driven chip cycle remains strong and that “AI is unlikely to be the only story in town” [3].
• JPMorgan (Nikolaos Panigirtzoglou): Acknowledges two scenarios: a bullish one where hyperscalers improve AI monetization and “catch up” to chip stocks, and a bearish one where hyperscalers pull back spending, creating a feedback loop that hurts chip stocks—similar to the dot-com divergence. The team leaned toward the more bullish scenario [2].
• Michael Burry: On July 5, 2026, the famed investor posted “the end is nigh” for the AI trade, calling it “mass addiction” and warning it “may die a death by a thousand cuts.” He accompanied his posts with Bloomberg charts showing AI semiconductor stocks dramatically outperforming hyperscale cloud providers and that the Philadelphia Semiconductor Index was trading near the top of its 15-year valuation range on both absolute and relative forward P/E bases [12].
• Roger McNamee (Elevation Partners): Warned that the AI investing structure is creating a “humongous bubble” [15].
• Dan Niles (Niles Investment Management): Said he would “put shorts back on” ahead of hyperscaler earnings [16].
• Cantor Fitzgerald’s CJ Muse: Noted that “the story of the past six months is the market going all-in on AI infrastructure, but now people are asking if this is sustainable and if we should be worried,” adding that “the biggest concern is about whether hyperscalers will sustain and grow their investments beyond 2026” [18].
• Quant Hedge Fund Stress: Quant hedge funds experienced their worst two-week run since 2023 due to the violent rotation, with a long-short momentum strategy dropping more than 3% for a second straight week, taking its two-week loss to the worst in more than three years [13].

Notably, on the same day the rotation thesis was covered, Morgan Stanley analysts also hiked price targets on Lam Research, Applied Materials, and KLA Corporation, making them some of the best S&P 500 premarket performers on July 6, 2026 [11]. This indicates the firm is not uniformly bearish on all chip stocks—it favors equipment makers even as it recommends rotating out of the broader semiconductor space into hyperscalers.

Valuation Comparison: Chipmakers vs. Hyperscalers

Chipmaker Valuations

Nvidia (NVDA)

Nvidia remains the dominant force in AI chips, but its stock performance has been relatively muted in 2026. As of July 4, 2026, Nvidia’s stock price was $197.58, up only 6.07% year-to-date [4][10]. The company reported Q1 FY2027 revenue of $81.61 billion, up 85.2% year-over-year, with Data Center revenue alone reaching $75.25 billion [4][10]. Non-GAAP gross margin held at 75.0%, and free cash flow reached $48.55 billion in the quarter [4][10]. The board authorized an $80 billion share buyback [4][10].

On valuation, Nvidia trades at approximately 21.7 times forward earnings, well below its two-year historical average of 34 times forward earnings, and roughly in line with the S&P 500 [2][13][17]. This forward multiple is far below slower-growing peer AMD, which trades at 73 times forward earnings [2][13][17]. Goldman Sachs analysts have advised patience, arguing that Nvidia’s current valuation already reflects fears of lost market share in the AI chip market, making the stock potentially undervalued [1][14]. Wall Street projects next-quarter revenue growth of 96% year-over-year, an acceleration from current levels [2][13][17].

AMD (Advanced Micro Devices)

AMD’s valuation stands in stark contrast to Nvidia’s. The stock trades at 73 times forward earnings, far above Nvidia’s 21.7 times, despite growing more slowly [2][13][17]. This premium multiple reflects market expectations that AMD will capture a larger share of the AI chip market, but it also makes the stock vulnerable if growth disappoints. On July 1, 2026, AMD fell 7% on the first day of Q3, after a strong Q2 in which chip stocks added $2 trillion in combined value [21]. On July 6, 2026, Japanese self-driving technology startup Turing added AMD Ventures as an investor and began using AMD AI accelerators, reducing its reliance on Nvidia, highlighting AMD’s expanding role in the AI ecosystem [17].

Micron Technology (MU)

Micron has been the standout performer among chipmakers, with its stock surging approximately 241% year-to-date as of July 2026, making it the second-best gainer in the Nasdaq-100 [9][13]. The stock briefly surpassed a $1 trillion market cap on May 26, 2026, and traded at $973.59 as of early July [37][9][13]. Despite the massive rally, Micron trades at a remarkably low forward P/E of approximately 6.7 times, based on fiscal 2027 consensus estimates of $149.64 EPS on $236.82 billion in revenue [13]. This is far below Nvidia’s ~22 times and Broadcom’s ~32 times [7].

The low multiple reflects the market’s historical tendency to apply a deep cyclical discount to memory stocks, even as Micron’s fundamentals have transformed. The company reported fiscal Q3 2026 revenue of $41.5 billion, up 345–346% year-over-year, with adjusted EPS of $25.11, up over 1,200% [5][6][11][13]. Gross margin reached a record 84.9%, up from 39% a year earlier [13][21]. Management guided Q4 revenue to $50 billion, representing 315–342% year-over-year growth [12][14][15][16].

A critical structural development is Micron’s signing of 16 multiyear Strategic Customer Agreements (SCAs) with binding volume commitments and floor prices—an unprecedented move in the memory-chip industry, which historically used short-term contracts [5][6][13][15]. These five-year contracts cover roughly 25% of total revenue over their terms, with a cumulative floor-price revenue potential of approximately $100 billion [28]. Cantor Fitzgerald analyst C.J. Muse stated that these agreements “will shift pricing dynamics, reduce quarter-end volatility in negotiations, and support more stable long-term margin and price discovery versus prior cycles” [14]. Despite this structural shift, the market continues to price Micron as a deeply cyclical stock.

Michael Burry has shorted Micron at $1,051.87, arguing that the AI-driven rally has gone too far, citing Micron’s history of 34 drawdowns of over 30% in 42 years and the stock trading further above its 200-day moving average than at any point since 1984, including the dot-com bubble [20][22]. He dismissed the high-bandwidth memory business as “just another in a very long series” of products and noted the company’s median return on invested capital of just 4% [20][22].

Hyperscaler Valuations

Microsoft (MSFT)

Microsoft has been one of the hardest-hit mega-cap stocks in 2026, with its share price falling to $372.97 as of June 30, 2026, down 22.54% year-to-date, and trading 25% below its 52-week high [12][14]. As of July 2, 2026, the stock was at $384.28 [2]. This decline has compressed Microsoft’s valuation to levels that many analysts consider attractive: a trailing P/E of 22.87 times, significantly below its five-year median of 34.01 times, and a forward P/E of approximately 19 times [2][12][13].

The company’s AI business is growing rapidly. Microsoft’s AI annual revenue run rate reached $37 billion, up 123% year-over-year, and Azure grew 40% [12][14]. Microsoft Cloud revenue hit $54.5 billion, up 29%, and Commercial Remaining Performance Obligations—a measure of contracted future revenue—reached $627 billion, nearly double year-over-year [12][14]. The company’s operating margin stood at 45.62%, and operating cash flow reached $46.68 billion in Q3 FY2026, up 26% [12].

The primary concern weighing on the stock is capital expenditure. Microsoft spent $30.88 billion on capex in a single quarter (Q3 FY2026), up 84% year-over-year, and plans approximately $190 billion in capex for calendar 2026 [12][13]. CFO Amy Hood noted that most capex is pre-contracted for the infrastructure’s useful life, but investors remain anxious about the return on this massive investment [12]. Microsoft also holds a ~27% stake in OpenAI worth approximately $135 billion, plus $250 billion of incremental Azure spend and IP rights through 2032, though OpenAI’s leaked financials revealed a 2025 operating loss near $21 billion on $13.07 billion in revenue [12][8].

Of 55 analysts covering Microsoft, 52 rate the stock a buy or strong buy, with zero sells [12]. Forbes/Trefis projects 49% upside to ~$548 over three years, assuming revenue compounding at 15.2% annually and a constant P/E multiple of 21.9 times [13].

Amazon (AMZN)

Amazon’s valuation data is less granular in the available research, but the company’s position as the largest hyperscaler spender is clear. Amazon plans approximately $200 billion in capex for 2026, the highest among the hyperscalers [10]. AWS remains the market-leading cloud platform, but it represents only about 18% of Amazon’s total revenue; the other 82%—retail, advertising, and Prime—is tied to consumer spending rather than enterprise capex cycles, providing a degree of resilience that pure-play cloud companies lack [7]. Amazon gained roughly 1% in the first week of July 2026 [9].

Alphabet / Google (GOOGL)

Alphabet has been a relative outperformer among hyperscalers, with its stock up 14.2% year-to-date as of July 1, 2026, trading at $357.89 [3][7]. The company joined the Dow Jones Industrial Average on June 29, 2026, replacing Verizon Communications [6][7][9]. Alphabet’s Q1 2026 EPS of $5.11 beat consensus by 94%, and its operating margin expanded to 36.1% [3][7].

Google Cloud revenue reached $20.03 billion in Q1 2026, up 63% year-over-year, with a cloud backlog that nearly doubled quarter-over-quarter to over $460 billion [3][7]. The company’s Gemini model processes 16 billion tokens per minute via API, up 60% quarter-over-quarter [7]. Alphabet guided 2026 capex to $180–190 billion and moved to raise $80 billion through stock sales in June 2026 to fund its AI buildout [3][7][10]. CFO Anat Ashkenazi stated: “We are seeing unprecedented internal and external demand for AI compute resources” [3][7].

The primary risk is the sheer scale of spending. Free cash flow fell 46.63% as capex more than doubled [3][7]. Google’s 2025 sustainability report revealed that annual electricity consumption rose by 37%—the largest increase in company history—driven by AI data center buildout, with total usage surging over 250% since 2019 [9].

Meta Platforms (META)

Meta has been the most volatile hyperscaler stock in 2026. The shares were down about 12% year-to-date before a dramatic rally on July 1, 2026, when the company announced plans to build a cloud infrastructure business to sell excess AI computing power. The stock surged over 8–10% on the news, reaching $617 per share [10][12]. At that price, Meta trades at approximately 19.5 times forward earnings [2].

Meta’s AI spending is enormous: the company plans approximately $135 billion in capex for 2026, nearly double its 2025 spend of $69.6 billion, and has committed $182.9 billion to AI infrastructure overall, with massive data center projects in Louisiana and Ohio [10]. The cloud pivot—reportedly called “Meta Compute” and led by Santosh Janardhan, Daniel Gross, and Dina Powell McCormick—directly addresses the biggest investor concern: that Meta was spending hundreds of billions on AI infrastructure with no clear monetization path [10]. CEO Mark Zuckerberg had previously stated in late May that a cloud business was “definitely on the table” as a way to get a return on the investment [10][11].

However, on July 2, 2026, Meta shares fell approximately 5% after Zuckerberg told employees during an internal town hall that AI agent development has not accelerated as expected over the past four months, and that recent reorganization and job cuts were not as clean as he wanted [11][13]. This highlights the execution risk inherent in Meta’s AI strategy.

The Motley Fool describes Meta at 19.5 times forward earnings as a “screaming deal,” arguing that either the AI investments pay off—creating a new business unit—or they fail, and Meta sells the infrastructure, maintaining its strong social media and advertising business. Either way, long-term investors should be fine [2].

Comparative Analysis

The valuation divergence between chipmakers and hyperscalers is stark and forms the quantitative backbone of Morgan Stanley’s rotation call. The table below summarizes key metrics as of early July 2026:

Metric	Nvidia	AMD	Micron	Microsoft	Meta	Alphabet
Forward P/E	~22x	73x	~7x	~19x	~19.5x	Not specified
Trailing P/E	~30x	N/A	~26x	~23x	N/A	N/A
YTD Stock Performance	+6%	N/A	+241%	-23%	-12% (pre-cloud)	+14%
Most Recent Revenue Growth (YoY)	+85%	N/A	+346%	+18%	+33%	N/A
2026 Capex Plan	N/A	N/A	$27B	~$190B	~$135B	$180-190B

The data reveals a paradox: the chipmakers that are supplying the AI buildout trade at lower forward earnings multiples than the hyperscalers that are spending on that buildout, with the exception of AMD. Nvidia at 22 times forward earnings is cheaper than Microsoft at 19 times on an absolute basis, but Nvidia’s growth rate is far higher. Micron at 7 times forward earnings is the cheapest of all, yet its stock has already surged 241% year-to-date, raising questions about how much of the AI memory super-cycle is already priced in.

The hyperscalers, by contrast, have seen their multiples compress due to stock price declines, even as their AI revenue streams begin to materialize. Microsoft’s AI business at a $37 billion run rate growing 123% year-over-year, and Alphabet’s Google Cloud growing 63%, suggest that monetization is accelerating. The market’s reluctance to reward these stocks reflects deep skepticism about the sustainability of AI capex and the ultimate return on investment.

Tactical Rotation or Structural Repricing?

Evidence for a Tactical Rotation

Several elements of Morgan Stanley’s call suggest it is primarily a tactical, momentum-driven recommendation rather than a declaration of a permanent shift in AI infrastructure valuation:

• Momentum Language: Wilson’s team explicitly frames the call in terms of “momentum fading” in semiconductor stocks and a shift “toward laggards” [3]. This is classic tactical rotation language, focused on relative performance over a near-term horizon.
• Maintained S&P 500 Target: Wilson kept his year-end S&P 500 target at 8,000, implying only modest upside from current levels. The call is about where to be positioned within the market, not a directional bet on the market itself [3][1].
• Simultaneous Chip Equipment Upgrades: On the same day the rotation thesis was covered, Morgan Stanley analysts hiked price targets on Lam Research, Applied Materials, and KLA Corporation [11]. This indicates the firm is not bearish on all semiconductor exposure—it is differentiating within the chip space, favoring equipment makers over chip designers and memory producers.
• Near-Term Catalysts: The call is tied to specific near-term events: the exhaustion of the Q2 semiconductor rally, the upcoming hyperscaler earnings season in late July 2026, and the potential for capex guidance adjustments [5][6]. These are tactical catalysts with defined time horizons.
• Choppy Market Context: Wilson described the rotation as occurring “in a choppy/weaker equity market overall,” suggesting it is partly a defensive repositioning within a volatile environment [3].

Evidence for a Structural Shift

Despite the tactical framing, there are structural forces at work that could turn this rotation into a more enduring repricing of the AI supply chain:

• The Capex Sustainability Question: The combined hyperscaler capex for 2026 is on track to reach $725 billion, nearly double the prior year’s spend [2][7]. Goldman Sachs projects that hyperscaler AI capex could exceed the GDP of major economies like Japan by the end of the decade [2]. This level of spending is unprecedented, and the market is beginning to question whether it can be sustained without clear monetization. If hyperscalers signal any moderation in spending plans during upcoming earnings calls, it would structurally benefit their own stocks—by easing ROI concerns—while structurally damaging chipmaker revenue growth expectations.
• The Monetization Inflection: Microsoft’s AI business at a $37 billion run rate, Alphabet’s Google Cloud backlog of over $460 billion, and Meta’s cloud pivot all suggest that hyperscalers are moving from the infrastructure buildout phase to the monetization phase. This is a structural shift in the AI investment cycle, analogous to the transition from building railroads to operating them. If the market begins to reward AI revenue rather than AI infrastructure spending, the valuation gap between hyperscalers and chipmakers could close from the hyperscaler side.
• Memory Industry Structural Change: Micron’s Strategic Customer Agreements with floor prices and binding volume commitments represent a fundamental change in how memory chips are priced and sold. If these contracts succeed in reducing the historical cyclicality of memory earnings, the market’s deep cyclical discount on Micron—a forward P/E of 7 times—could prove to be a structural mispricing. However, this cuts both ways: if the memory cycle does turn, the contracts provide downside protection that previous cycles lacked, potentially making memory stocks more resilient than in past downturns.
• The JPMorgan Divergence Warning: JPMorgan’s Nikolaos Panigirtzoglou noted that the divergence between chip stocks and hyperscalers is reminiscent of the dot-com era, and that a bearish scenario—where hyperscalers pull back spending, creating a feedback loop that hurts chip stocks—could structurally reprice the entire AI supply chain lower [2]. This is not a tactical risk but a structural one.
• The Shiller CAPE Ratio: The S&P 500’s Shiller CAPE ratio has stayed above 40 since May 2026, a level only seen during the dot-com bubble [10][32]. While this is a broad market warning rather than specific to AI, it suggests that the overall equity market is priced for perfection, and any disappointment in the AI trade could trigger a broader repricing.

Implications for the AI Supply Chain

If Morgan Stanley’s rotation call proves prescient, the implications for the AI supply chain would be significant:

• Chipmakers would face a period of relative underperformance, particularly those with the highest beta to AI capex, such as Nvidia and AMD. However, the structural demand for AI chips is unlikely to disappear. Nvidia’s CEO Jensen Huang stated in June 2026 that “the buildout of AI factories, the largest infrastructure expansion in human history, is accelerating at extraordinary speed,” and that supply shortages across wafers, packaging, and silicon photonics will persist for several years [4][10][5][6]. A rotation would compress multiples but not necessarily destroy earnings.
• Memory makers like Micron face a unique tension. Their stocks have already priced in an enormous amount of good news, with Micron up 241% year-to-date. The low forward P/E of 7 times suggests the market is already discounting a cyclical downturn, even as the company’s SCAs provide structural earnings protection. If the memory cycle remains strong, Micron could continue to outperform even in a rotation environment; if it turns, the stock could fall sharply despite the low multiple.
• Hyperscalers would benefit from a re-rating if the market begins to value their AI revenue streams more highly. Microsoft at 19 times forward earnings with a $37 billion AI run rate growing 123% year-over-year, and Meta at 19.5 times with a new cloud business, appear undervalued relative to their growth potential—if AI monetization proves durable. The key risk is that AI revenue growth disappoints, or that the massive capex spend fails to generate adequate returns, in which case the stocks could de-rate further.
• The broader AI ecosystem—including neoclouds, cybersecurity, and power infrastructure—would see divergent impacts. Cybersecurity names like Palo Alto Networks and CrowdStrike have already benefited from the perception that AI models are becoming more capable at finding vulnerabilities [9]. Power grid technology has attracted $4.8 billion in VC deal value in 2025 and is tracking ahead through H1 2026, as AI data center energy demands soar [9]. These sectors could continue to attract capital even if chipmakers underperform.
• The rotation could accelerate the shift from hardware to software and services in the AI value chain. PitchBook’s 2026 Advanced Software Launch Report argues that the “SaaS-pocalypse” downturn is ending, giving way to a historic AI margin super-cycle driven by the shift from seat-based licenses to outcome-based digital labor and agentic AI [11]. If this thesis is correct, the next phase of AI investing will favor companies that monetize AI through software and services—precisely the hyperscalers and their ecosystem partners—over the hardware providers that enabled the initial buildout.

Conclusion

Morgan Stanley’s call for a rotation from AI chipmakers to hyperscalers, authored by Michael Wilson and published in late June 2026, is a nuanced and timely intervention in a market characterized by extreme divergence. The thesis rests on momentum exhaustion in semiconductor stocks, the relative resilience of hyperscaler core businesses, and the potential for capex guidance moderation to act as a positive catalyst for hyperscaler stocks. The valuation data supports the relative-value argument: Nvidia at 22 times forward earnings is not expensive, but the semiconductor sector as a whole has run too far too fast, while Microsoft at 19 times and Meta at 19.5 times forward earnings appear cheap relative to their AI revenue growth trajectories.

Whether this rotation proves tactical or structural depends on the answers to two questions that will be tested in the upcoming earnings season: Can hyperscalers demonstrate that their massive AI capex is generating sustainable revenue growth, and will they signal any moderation in spending plans? If the answers are yes and yes, the rotation could mark the beginning of a structural repricing that favors the monetizers of AI over the builders of AI infrastructure. If the answers are no and no, the rotation may prove short-lived, and chipmakers could resume their leadership as the AI buildout continues.

What is clear is that the AI trade is entering a new, more discriminating phase. The days of indiscriminate buying of anything AI-related are over. Investors are now asking harder questions about return on investment, competitive moats, and the sustainability of spending. Morgan Stanley’s call is both a reflection of that shift and a bet on its continuation.