To The Moon

1️⃣ After the AI Selloff, Which Segment Do You Favor?

Since October, AI-linked stocks with exposure to GPU compute, memory, optical interconnects and foundation models have suffered a sharp selloff. As some technology stocks rebound, investors considering a return to AI face a key choice: Will the next wave of AI investment returns remain concentrated in infrastructure companies supplying compute, or increasingly flow to application businesses turning AI into revenue and profits?

In its November 9 weekly commentary, BlackRock placed this question at the center of discussions at its 2027 Outlook Forum. The early AI trade largely priced in the conviction that infrastructure had to be built. The research focus is now shifting to whether the size and speed of the buildout match actual adoption, and who ultimately captures the gains. [1]

Being positive on AI demand and deciding which AI stocks to own are two separate questions. A company may supply compute, sell AI services or use AI in its own operations. Below, we group the institutional views by where the gains come from.

The comparison covers October 12–November 9, 2026 and uses closing prices. These stocks are observation samples, not indices representing their respective segments.

2️⃣ Compute Suppliers: Demand Has Support, but How Long Can Scarcity and Pricing Power Last?

In its November 6 CIO Daily, UBS argued that the selloff primarily reflected a reassessment of the cost, pace and execution of the data-center buildout. It did not see sufficient evidence of a fundamental change in the AI investment cycle. The bank pointed to support from cloud revenue growth and advance orders for computing capacity; its base case remained that the AI investment cycle was intact. [2]

Its preference for infrastructure businesses, however, came with clear conditions: durable pricing power, strong competitive positions and access to capital. UBS was more selective toward highly valued companies whose investment cases depended heavily on a single component, capacity bottleneck, customer or stage of the buildout.

BlackRock pushed the question further, to scarcity itself: Which constraints will persist? Where could supply become abundant as capacity expands? Can demand justify the scale of the buildout? These remained open questions in its Forum discussions, rather than a conclusion that the opportunity for suppliers had ended. [1]

The two perspectives fit along the same line of inquiry. UBS retains a preference for infrastructure businesses with competitive advantages; BlackRock asks investors to examine whether those advantages could weaken as supply and demand evolve.

For compute and networking suppliers such as NVDA, AMD and AVGO, the relevant questions therefore extend beyond whether AI will keep growing. They also include what customers will pay, whether orders will convert into delivery, and whether profits can grow alongside demand.

3️⃣ Platforms and Applications: Can More Usage Produce More Profits They Can Keep?

In the same November 6 report, UBS explicitly favored financially strong platform and application companies capable of monetizing growing AI adoption. That preference emphasizes two conditions: the business can fund the investment, and customers are willing to pay for its services. [2]

The bank also outlined what would make it reassess the AI cycle: slowing AI-related revenue growth, weakening cloud utilization, or falling AI costs that fail to generate enough additional usage to absorb expanding capacity.

BlackRock’s November 9 discussion highlighted a potential timing gap: capital spending could slow before productivity gains from AI adoption materialize. The firm also heard from corporate CFOs about deploying AI and measuring its benefits, but the public commentary did not disclose complete case studies or models sufficient to compare returns on investment across specific companies. [1]

These institutional views focus on whether adoption, revenue and investment can form a sustainable relationship. More people using AI validates only one part of that relationship. Platforms and applications still need evidence of paid adoption and profitability to establish themselves as the next winners.

4️⃣ AI Adopters: When Does Greater Efficiency Become Profit for Shareholders?

Adopters are businesses using AI in their own operations, such as banks, healthcare providers or retailers. They may not sell AI, but could benefit by reducing repetitive work, improving decisions and streamlining processes.

As earlier background research, Morgan Stanley Wealth Management argued in a September 23 public article that the next phase of opportunity could come from companies adopting AI deeply. Its criteria included large pools of structured work, proprietary data, balance-sheet flexibility and management teams willing to redesign workflows rather than simply add AI tools to existing processes. [3]

The firm also stressed that the payoff is not automatic. Data preparation, governance, cybersecurity, employee training and workflow redesign require upfront investment. Productivity and margin improvements may follow later, creating a productivity “J-curve”.

Research cited in the public article found that, as of July 2026, around 25% of S&P 500 companies had mentioned at least one quantifiable AI impact in earnings or conference transcripts, up from roughly 15% in the third quarter of 2025. This is an earlier disclosure statistic. Mentioning a quantifiable impact does not establish that a company has generated sustainable net profits from AI. [3]

Another question raised by BlackRock on November 9 completes the second half of the argument: Do the gains ultimately accrue to technology providers, businesses adopting AI, or their customers? Its commentary did not provide a definitive answer. [1]

The focus of this research is therefore whether companies can turn implementation spending into better operating performance, and retain the resulting gains. Greater efficiency can create opportunities for adopters, but competition and customer bargaining power may still affect how the benefits are distributed.

The framework groups businesses by their sources of gains; one company may operate across several parts of the chain. It is an analytical framework for this article, rather than evidence that opportunities have already shifted in sequence.

5️⃣ Our View: Broaden the Research Universe and Let Financial Evidence Guide Stock Selection

Our view is that investors can now broaden their research beyond compute buildout to AI monetization and operating profits, but there is not yet enough evidence to justify a wholesale shift in exposure from suppliers to adopters.

There are three reasons. UBS still sees support for suppliers’ underlying demand; the key test is how long their competitive advantages can last. Platforms and applications need to demonstrate that revenue growth can cover service and construction costs. Adopters need to navigate implementation spending, the timing of benefits and potential competitive pressure to pass those benefits on.

BlackRock raises the question of who captures the gains, while UBS offers conditional business preferences. Neither can be translated directly into a list of guaranteed winners in the next AI cycle.

We would update our assessment using three sets of evidence:

• Suppliers: Order conversion, pricing and cash collection. Are growing orders translating into actual delivery and cash receipts? As supply expands, do selling prices, gross margins and repeat orders remain supported? If growth requires sustained price cuts, the returns associated with scarcity need to be reassessed.

• Platforms and applications: Paid growth relative to incremental costs. As usage rises, are paid revenue, renewals and gross profit improving alongside it? Are compute expenses, capital spending and subsequent depreciation absorbing the incremental gains? If a company does not separately disclose AI revenue, its overall cloud or software growth cannot simply be attributed to AI.

• AI adopters: Net benefits and the ability to retain them. Does the company disclose specific workflows, unit costs and operating outcomes? Do savings remain after software, training and workflow-redesign costs? Can margin improvements persist, or are they offset by price cuts, competition or changes elsewhere in the business? Efficiency figures from a single pilot still need validation in subsequent operating results.

If suppliers maintain pricing, orders and cash collection, they can remain important beneficiaries of AI. If platforms or adopters begin consistently disclosing net benefits and upward earnings revisions—and the market has not fully priced in those improvements—there would be a more concrete basis for expanding exposure to them.

The ultimate comparison is which company creates gains beyond what its share price already reflects. AI-driven business growth does not guarantee rising stock prices. The next businesses worth watching may come from several parts of the chain at once.

6️⃣ Your Turn: Where Would You Put Your Next AI Investment?

• A. Compute suppliers. You prioritize demand, orders and technological advantages, and believe leading companies can continue to retain the profits.

• B. Platforms and applications. You want evidence that customers will pay and that revenue growth can cover AI investment.

• C. AI adopters. You prefer businesses embedding AI in their operations and beginning to deliver cost savings and profit improvements.

Which group do you favor? What earnings evidence would change your choice: orders and gross margins, paid AI revenue, or sustainable improvements in operating profit? Share the reasoning behind your view.

Research Sources

1. BlackRock Investment Institute: Midterms and AI in focus at 2027 Forum, November 9, 2026. Official public weekly commentary covering the alignment of buildout and adoption, the distribution of gains and competition for capital. Some questions remain unresolved.

2. UBS CIO: Chip rout tests AI investment case, November 6, 2026. Official public article covering the base case for the AI cycle, selection criteria for platforms and infrastructure, and signals that would prompt a reassessment.

3. Morgan Stanley Wealth Management: Why the Biggest AI Opportunity Is Still Ahead, September 23, 2026. Earlier official background research. The cited disclosure statistic on quantifiable AI impacts is as of July 2026, rather than new data from this week.

4. Investing.com: NVDA，AMD，MU，000660.KS，LITE，PLTR，CRM，GOOGL，META's historical prices. Sources 4–12 support the chart in section 1. All calculations cover October 12–November 9, 2026. Percentages were calculated by Capital_Insights from daily closing prices in each market, excluding dividends.