After analyzing tens of thousands of financial reports, Morgan Stanley found that the "services + cyclical" sectors, which have been sold off, actually have the highest AI adoption rate and the strongest bargaining power.
```
In the recent period, a sense of “AI anxiety” has permeated Wall Street.
The market is worried that with the rise of generative AI (GenAI) and agentic AI, many traditional “service + cyclical” enterprises—especially software, information services, and financial intermediaries—will be thoroughly disrupted. This panic has led to indiscriminate sell-offs in the related sectors.
But this could be a huge pricing mistake.
On February 25, Morgan Stanley’s US equity strategy and thematic team released a report stating that the recent US stock market has overreacted to the “AI disruption theory.”
First, after the steep decline, this group perceived as “disruption targets” currently accounts for only 13% of the S&P 500’s total market cap. This proportion explains why the recent major indices have only experienced limited pullbacks, while there’s been a bloodbath within certain sectors.
Second, this group’s valuation and concentration are already at extremely low levels. According to Morgan Stanley’s data, the relative valuation of the “service + cyclical” sectors is at the 9th percentile since 2010, almost the cheapest in history. Institutional net exposure has dropped to the 20th percentile, indicating very underweighted positions.
Morgan Stanley puts it bluntly: “The bearish view on GenAI seems to seriously underestimate the legacy software vendors’ ability to participate in this innovation cycle.”
In fact, these groups that have been sold off were not only not disrupted, but are actually, according to Morgan Stanley’s thematic team’s AI mapping analysis, the groups with the highest AI adoption rates and strongest pricing power (top third).
In the market’s eyes, the “victims” are actually the biggest “beneficiaries.” These hard-hit groups coincidentally have extremely high AI adopter concentration.
Quantifiable gains have emerged—real money, not just promises
Investors broadly doubt: Can AI actually save or make money for companies at this stage? The data gives a positive answer.
Morgan Stanley’s team used AI models to analyze over 10,000 financial reports and meeting transcripts. The results show that companies are indeed gaining substantial AI dividends, and the momentum is growing.
In the just-concluded fourth quarter of 2025, among companies identified by analysts as “AI adopters,” 30% mentioned at least one “quantifiable financial impact” from AI in their conference calls.
That proportion was 24% in 3Q25, and only 16% in 4Q24. For the broader S&P 500 components, the proportion also rose to 21%.
Morgan Stanley says directly, “The most mentioned quantifiable gains are mainly ‘financial impacts’ (including revenue growth and cost savings), and the number of mentions doubled from the previous quarter.”
Reflected in fundamentals, AI adopters with strong pricing power are not seeing their forward net profit margins disrupted; on the contrary, they are accelerating expansion.
Morgan Stanley expects that AI adoption will contribute 40 basis points of margin growth to the overall S&P 500 in 2026.
Data is confirming the profit expansion of “AI adopters”
The report shows that between 2024 and 2025, AI adopters’ EBIT margin expanded by 310 basis points, a rate twice that of the MSCI Global Index over the same period. Morgan Stanley analysts expect about 80% of AI dividends will be reflected in cost efficiency improvements.
For example, Citigroup said: “Since the beginning of this year, AI-driven automated code reviews have exceeded one million, greatly improving developer productivity. This single innovation saves about 100,000 hours per week.”
European companies are the most aggressive. Surveys show that a net 35% of European firms plan to use AI to reduce workforce size, far above the slightly more than 10% in other regions. This points to even stronger profitability in the future.
Historical Mirror: Lessons from the Smartphone Era in 2007
To elucidate the current market logic, Morgan Stanley rewinds the clock to 2007.
At that time, the iPhone had just been released, and the market was similarly caught in “disruption panic.” Industries like gaming, PCs, printers, GPS, and desktop software were seen as doomed groups.
The data shows that in the years following the iPhone launch, performance of these “disrupted concept stocks” diverged drastically.
Faced with the same shock, Google, by capitalizing on mobile advertising, rose 28%, while Nokia plummeted 73%.
After testing multiple fundamental variables, Morgan Stanley found that the core metric determining stock performance when facing epoch-making technological shocks is the change in forward earnings.
In other words, those who can use AI to drive profit growth will have the last laugh in the capital markets. Since the end of 2023, the magnitude of earnings upgrades among AI adopters has been roughly twice that of those disrupted by AI. As returns on investment accumulate, the gap is widening.
For example, after the iPhone launch, the Spearman rank correlation coefficient between forward earnings and stock price performance reached a very high 0.9 (extremely strong correlation).
Morgan Stanley summarizes: “What we are experiencing now is a classic feature of a major investment cycle. Capital will flow not just to structural leaders, but also to cyclical leaders. Bottom-up stock selection is especially important at this time.”
Moats Deeper Than Imagined: Compliance, Trust, and Proprietary Data
Regarding AI impacts faced by specific industries, Morgan Stanley’s analysts provided detailed logical breakdowns, revealing what is true disruption and what is false alarm:
Software: Panic at its peak—AI is not a “new category” but a “new capability”
The software sector has recently experienced a sharp valuation hit. The current average EV/Sales multiple (about 4.4x) has returned to the lows of 2014-2016, when there was extreme public panic over cloud computing.
The market has “three major worries”: AI startups seizing share, the collapse of seat-based business models, and GPUs driving up costs and compressing margins.
Morgan Stanley directly says these worries are misplaced: “Generative AI fundamentally expands enterprise software’s capabilities. The issue is not whether software can monetize in this innovation cycle, but who will build these add-on capabilities.”
Morgan Stanley believes AI is, in essence, an expansion of enterprise software capabilities, solving the pain of traditional software’s inability to handle unstructured data. The biggest beneficiaries are current giants that possess distribution channels, proprietary data, and workflow control.
Consumer finance & payments: AI cannot replace trust and compliance
Recently, markets worry that “agentic AI” can shop autonomously, bypassing traditional credit card networks.
Morgan Stanley refutes this view: “We are skeptical that agentic AI can significantly disrupt credit card exchange networks. This overlooks the importance of trust, fraud protection, credit extension, and customer rewards.”
In these highly data-intensive and rule-driven industries, regulatory licenses and balance sheets are natural barriers. AI will only accelerate optimization of the back-end, such as underwriting, anti-fraud, and customer service.
Morgan Stanley expects banks and consumer finance firms to significantly improve operational leverage with AI. In 2026 and 2027, large banks’ profitability could be further boosted.
Internet & e-commerce: The next “agentic commerce” will enlarge the pie
Morgan Stanley forecasts that “agentic commerce” capable of autonomously helping users compare prices and place orders will be a major generative AI unlock point.
This makes the consumer funnel more conversational, personalized, and interactive. By 2030, agentic commerce is expected to add an extra $50–115 billion in US e-commerce spending.
Platforms with extensive logistics infrastructure, unique inventory, and strong fulfillment capability will not be replaced; instead, they’ll leverage AI to expand their online wallet share.
Transportation: Asset-heavy companies get the windfall, asset-light ones are truly at risk
Transportation is one of the industries most easily affected by AI. But Morgan Stanley points out huge differences within it.
Asset-heavy operators with fleets, rail, and warehouses will be pure beneficiaries of AI. Physical AI (self-driving trucks, humanoid robots) will structurally reduce labor costs (the largest component) and improve asset utilization.
In contrast, asset-light freight brokers (3PLs) that profit from information asymmetry face real disruption risks. Generative AI is commoditizing freight matching, continuing to squeeze broker profit margins.
Real estate & commercial insurance: Highly complex, non-standard business hard to replace
For commercial real estate services and large commercial insurance brokers, the market underestimates the complexity of their business.
Large commercial policies require complex contract parsing, risk tower building, and compliance reviews. Morgan Stanley points out: “AI cannot replace this kind of expertise that requires market access and regulatory oversight.”
In commercial real estate, AI is more of an “enhancement” than a “replacement.” These labor-intensive firms will use AI to cut back-office costs. Morgan Stanley estimates that AI automation in public REITs and CRE services could deliver up to $34 billion in financial impact, equivalent to 16% of operating cash flow.
Truth about the labor market: Will AI cause mass unemployment?
The ultimate worry throughout the “AI disruption theory” is that AI will cause massive white-collar unemployment, leading to economic recession and lower consumption.
Morgan Stanley, reviewing 150 years of technological change (electrification, tractors, computers, internet), points out that history shows every major technological innovation profoundly alters labor structure, but “has not replaced labor.”
On the contrary, technology creates entirely new jobs. Morgan Stanley expects that as AI is deployed more deeply, companies will need not just “Chief AI Officers,” but new jobs such as “product manager-engineer hybrids,” “AI supply chain forecasters,” “computational geneticists,” and so forth.
In short, it’s true that new technology brings the growing pains of old orders. But when the market blindly panics and undervalues many high-quality assets, the optimal way to navigate the tech cycle is to return to business fundamentals—focusing on proprietary data, physical asset moats, and long-term earnings power.
~~~~~~~~~~~~~~~~~~~~~~~~
The above content is from Chasing Wind Trading Desk.
For more details, including real-time interpretation and frontline research, please join [Chasing Wind Trading Desk • Annual Membership]
Risk Warning and DisclaimerThe market is risky; investments should be made with caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial situation, or needs of any particular user. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific situation. Investments based on this article are at their own risk. ```