Ten Public Fund Leaders Diagnose the AI Market: No Broad Bubble But Clear Segmentation Emerging Across Subsectors
Table of Contents
You might want to know
Is the current AI-related equity rally a widespread valuation bubble, or a selective, structure-driven re-pricing across subsectors? Which AI supply-chain segments carry the most durable demand and justify higher valuations?
Main Topic
The recent volatility across global technology markets — from Nasdaq oscillations to sharp single-day falls in Korea — has intensified debate over whether AI-related equities have become a valuation bubble or whether the market is experiencing a structural, segmented reallocation of capital. Ten fund managers with concentrated AI exposure were interviewed to assess valuations, supply-chain vulnerabilities, and the likely trajectory of AI-related performance through 2026 and beyond. Their consensus: this is not an economy-wide or technology-wide bubble akin to the broad distortions of 2000, but it is a period of pronounced sectoral divergence and concentrated exuberance in parts of the space.
There are three analytical layers to unpack: valuation realism versus narrative excess; the real, observable supply-demand gaps in specific hardware segments; and the strategic implications for investors—both in terms of which subsectors may sustain elevated returns and how to position portfolios prudently.
On valuation, the interviewed managers almost uniformly argued against labeling the entire AI sector a bubble. Cross-market comparisons do not show the sort of universal index-level distortion typical of a systemic tech bubble. Instead, what is apparent is that some niches and individual stocks have priced in 1–2 years or more of future execution. Where high valuation is supported by rapid revenue growth and clear pathway to durable competitive position, the premium can be justified. Where valuations have outrun near-term cash-flow and order visibility, the risk of disappointment is materially higher.
Several managers pointed to observable segmentation within the supply chain. Segments such as optical modules, PCB and server suppliers, and memory have demonstrated real revenue expansion in recent reporting periods. For example, optical components faced a severe supply tightness amid strong demand for higher-speed interconnects; when revenues and margins expanded accordingly, valuations often reflected tangible earnings support. Yet even in these fast-growing pockets, concentrated investor interest has elevated multiples in some cases, which thins valuation cushions and raises sensitivity to execution misses or faster-than-expected supply relief.
Another principal theme is the role of domestic substitution (domestication of semiconductors) in China’s AI industrial story. Many managers view semiconductor domestic replacement as a high-conviction, high-certainly investment theme. The combination of strong local AI-driven demand and accelerating domestic production capability—across memory, advanced packaging, equipment, and certain specialty processes—creates the prospect of substantial order waves for domestic suppliers. If realized, that dynamic supports a structural re-rating of select domestic vendors. Equipment players, in particular, are highlighted because the product sets are fewer and value per unit tends to be large, which amplifies potential market-cap upside when domestic alternatives displace imports.
Alongside domestic substitution, a second durable investment thread is the hardware supply-demand mismatch for foundational AI infrastructure: optical modules, high-speed interconnects, memory and storage, and select photonics and packaging technologies. These segments have experienced demand accelerations driven by cluster-scale deployments and upgrades from 800G to 1.6T (and beyond). Because many upstream components have long lead times and require complex manufacturing capabilities, supply cannot quickly expand to meet sudden demand surges. This results in multi-year cycles of tightness, price increases, and outsized margin expansion for capable suppliers. Several interviewed managers forecast prolonged strong demand through 2027–2028, thereby implying multi-year earnings growth rather than transient hype.
Despite overall optimism on targeted segments, interviewees also cautioned about clear risk scenarios. Signposts that would indicate an inflection toward a more pervasive, price-driven unwind include: a sustained deterioration in capex-to-cash-flow dynamics; order momentum slowing or visible price concessions across multiple subsectors; concurrent sharp markdowns in previously leading names; and broad spillover of speculative capital into lower-quality, non-core technology stocks. If these developments materialize, the market could move from a validation phase to a broader de-risking environment.
From a portfolio construction perspective, managers emphasized layered allocation rather than blanket exposure. Tactical guidance included: avoiding concentration in names where valuation already discounts significant future growth; prioritizing firms with transparent order books and short- to medium-term margin visibility; and favoring equipment and component suppliers where supply constraints and long replacement cycles generate durable pricing power. Several managers also recommended considering a barbell approach—core holdings in high-quality suppliers and selective exposure to higher-risk, higher-reward innovators—combined with systematic strategies like dollar-cost averaging to mitigate timing risk.
Finally, many managers stressed the market’s transition from an expectation-driven phase—dominated by narrative and momentum—to a fundamentals-driven phase where mid‑term and near-term earnings will determine leadership. This implies that while AI is likely to remain a primary thematic axis for markets in 2026 and beyond, relative performance will be driven by proof of execution, order conversion, sustained margins, and the ability of firms to capture structurally increasing demand rather than by broad-based sentiment alone. As such, investors need to discriminate within AI, distinguishing durable structural winners from high-expectation, high-volatility names that may disappoint absent rapid revenue conversion.
Key Insights Table
| Aspect | Description |
|---|---|
| Valuation Diagnosis | Not a uniform bubble; evidence points to structural segmentation with localized overpricing in some themes. |
| Primary Investment Themes | Semiconductor domestic substitution and AI foundational hardware facing supply-demand mismatches (optical modules, memory, packaging). |
| Drivers of Outperformance | Genuine supply shortages, long lead times, and sustained order growth supporting multi-year margin expansion. |
| Key Risks | Rapid capex oversupply, demand deceleration, crowded trades, and under-delivery on revenue conversion. |
| Investor Guidance | Avoid chasing peaks; adopt staggered entries, focus on earnings visibility, and consider diversification strategies such as barbell allocations or dollar-cost averaging. |
Afterwards...
Looking forward, the AI investment landscape appears set to remain an important market theme, but its internal composition will matter more than ever. The next phase will reward firms and investors anchored in real demand signals, deliverable margins, and tangible production capabilities. Policymakers and industry participants advancing domestic capabilities may further reshape global supply chains, creating new winners and accelerating the secular shift toward onshore production in key technology nodes. For investors, patience and discriminating selection—favoring businesses with visible order books, durable pricing power, and realistic delivery timelines—will likely yield better outcomes than indiscriminate exposure to the broad AI narrative.
In short, AI is not a monolith: it is a constellation of opportunities and risks. The market is in a sorting phase, where execution and supply-side realities will determine which segments sustain elevated returns and which revert as sentiment normalizes.
