AI Rotation or Tech Bubble? Repricing the Crowded Growth Trade

AI-led technology and semiconductor names are no longer the only leadership in public markets. Large managers such as Invesco describe the current pattern as a healthy market rotation, not the start of a tech bust, arguing that the backdrop looks far more balanced than the late‑1990s setup and that breadth is finally improving beneath the surface. The framing matters: for institutional portfolios heavily overweight AI and growth, this is not just market color but an invitation to re-underwrite exposure and decide how much of the crowded trade to recycle into cyclicals and under‑owned sectors.

Why Invesco Says This Is Not 1999

Invesco’s recent note on the case for a healthy rotation versus a tech bubble makes a clear distinction: today’s market, in its view, “looks far healthier and far more balanced than the one that preceded the dotcom bust.” Leadership within technology has shifted from the hyperscalers that committed heavy capital to AI—companies such as Alphabet, Amazon, Meta and Microsoft—to firms that are now directly benefiting from that investment. To that author, “that’s a rotation, not a breakdown,” signalling factor rebalancing rather than a collapse in the core AI thesis.

Under the surface, the breadth data reinforces that view. Invesco highlights that the other 493 S&P 500 constituents have posted positive performance in June and are up more than 14% year to date, precisely the kind of broadening many investors argued was missing when 2024’s gains were concentrated in a small set of AI and mega‑cap names. Industrials, real estate and financials have all delivered strong performance over the past three months, and small‑cap indexes sit at all‑time highs. These are not the usual hallmarks of a late‑cycle, narrow bubble that depends on a handful of high‑multiple stocks to sustain the index.

Macro Backdrop: Inflation Moderates, Leadership Broadens

The rotation thesis rests in part on macro conditions. Oil prices have fallen sharply, market‑based inflation expectations have declined, and inflation breakevens point toward moderation rather than acceleration. While memory prices have risen—making laptops and tablets more expensive—Invesco notes that this category is a small slice of the consumer basket. In aggregate, the signals are of disinflation rather than renewed inflation pressure, reducing the risk that central banks are forced into aggressive tightening that would compress growth multiples.

Other commentators point to similar dynamics. Commentators on financial networks describe capital actively rotating out of tech into financials and healthcare and characterize that as healthy for the overall market. At the same time, there are visible cracks: revenue misses at key AI‑exposed names such as Broadcom remind investors that earnings expectations can overshoot, and some analysts argue that the “AI bubble is showing cracks” as high‑beta tech enters correction territory. The divergence in narratives—healthy consolidation versus early bubble deflation—puts the burden back on allocators to update probabilities rather than pick a side rhetorically.

Late-1990s Analogues and Structural Differences

Comparisons to the late‑1990s tech bubble are inevitable. In both cases, a new technology narrative—then the internet, now AI and related compute infrastructure—pulled capital into a narrow set of high‑growth names and drove index‑level gains. Some cycle analysts suggest that the current trend may be analogous to the early 1990s phase that preceded the final bubble years, raising the possibility that this is the beginning of a longer bull market rather than its end. Others emphasise that today’s leading platforms generate substantial cash flow and real earnings, unlike many of the pre‑2000 dot‑com listings.

Invesco’s “this isn’t 1999” framing rests on three structural differences. First, earnings visibility for AI infrastructure and hyperscaler leaders is higher, with capital spending tied to concrete workloads rather than purely speculative traffic growth. Second, the rotation is not confined to tech; performance in industrials, real estate, financials and small‑caps suggests a broadening support base rather than a single‑sector mania. Third, the macro backdrop shows moderating inflation and falling breakevens instead of the tightening cycle that ultimately undermined the dot‑com boom. None of this rules out sharp drawdowns, but it does support a Bayesian update away from a high probability of imminent bubble collapse.

Re-Underwriting the Crowded AI and Growth Trade

For institutions, the key question is not whether the move is called a rotation or a bubble, but how to size AI and quantum computing winners versus the beneficiaries of a sustained shift into cyclicals and real‑economy exposure. The AI and technology complex has delivered an “extraordinary run,” and some strategists argue that a rotation was overdue as capital reallocates toward areas with more reasonable valuations and clearer near‑term visibility. Whether this is a temporary reset or a durable regime shift will determine how much crowding risk remains in AI infrastructure, semiconductors and mega‑cap platforms.

From a decision‑theory perspective, the environment now blends several forces. Game‑theoretic dynamics are evident as large asset managers manage visible crowding in AI and semis while watching peers for signs of a more wholesale de‑risking. Bayesian updating is central: each inflation print, earnings revision and rotation datapoint adjusts the probability that the past year’s gains reflect a lasting re‑rating of AI cash flows versus a momentum‑driven overshoot. Prospect‑theory effects are visible in investor behaviour; fear of missing further upside in AI coexists with loss aversion after recent drawdowns, making rotations into industrial automation, infrastructure and energy transition plays emotionally and politically easier inside institutions than wholesale de‑risking.

Sizing Rotation Beneficiaries Versus Structural Winners

The practical allocation problem now is balancing asymmetric upside in AI and quantum “winners” against the cyclical catch‑up potential in sectors that have lagged but are now benefitting from broadening leadership. Industrial automation suppliers, infrastructure developers and energy transition companies stand to benefit if the rotation persists and if macro conditions remain supportive for capital expenditure and policy‑driven projects. Financials and healthcare, cited in recent media coverage as rotation destinations, offer different factor exposures—value, dividend yield, defensiveness—that can offset the growth and duration risk inherent in AI‑heavy portfolios.

Regret minimization becomes a relevant lens. Allocators must choose position sizes that minimise future regret across multiple macro paths: one in which AI earnings compound and current multiples prove sustainable; one in which rotation stabilizes, with broader sectors delivering mid‑single‑digit outperformance; and a less favourable scenario in which both AI and cyclicals reprice on weaker growth or renewed inflation. In each case, the convexity of AI and quantum exposure must be weighed against the stabilizing effect—and potential catch‑up—offered by industrials, infrastructure, financials, energy and healthcare. The current “healthy rotation” narrative provides cover to rebalance, but the real work lies in quantifying factor exposures, crowding and downside scenarios rather than assuming that breadth alone guarantees durability.