VanEck CEO Says AI Bubble Already Burst; Market Correction Cleared Excesses in 2025

Jan van Eck’s public statement represents a significant voice in the ongoing debate about AI sector valuations, coming from the leader of a major asset manager with over $100 billion in assets under management. The timing of this commentary—early 2026—provides critical context, as it follows the market turbulence of 2025 that van Eck characterizes as a necessary correction rather than the bursting of a speculative bubble [1].

The core thesis advanced by van Eck centers on the concept that the AI sector has already completed what strategist Josh Brown describes as “taking out its own trash”—a process of market purification where speculative excesses are expelled through price corrections while fundamentally sound companies remain standing. This differentiates the current AI cycle from the dot-com era, according to van Eck, because today’s AI companies are generating actual profits and cash flows rather than relying on growth narratives alone [1]. The valuation reset was triggered by disappointing share price performance from prominent AI plays including Oracle and CoreWeave, as well as bitcoin miners and leveraged compute companies that had ridden the AI wave [1].

From a market data perspective, recent indices performance suggests relative stability following the correction period. The S&P 500 has posted modest gains of approximately 1.52% over the past seven weeks, while the NASDAQ has shown similar resilience with a 1.42% increase, indicating that tech sector holdings have maintained their ground rather than experiencing catastrophic declines [0]. The Russell 2000’s outperformance at 7.53% suggests potential rotation toward small-cap opportunities, which may align with van Eck’s thesis about identifying beneficiaries that “may not appear on investors’ radar” [0].

Van Eck’s analysis introduces a nuanced perspective on AI sector differentiation that merits careful consideration. The assertion that “there is plenty of differentiation between winners and losers” suggests that the market correction served a valuable screening function, separating companies with genuine competitive advantages and sustainable business models from those primarily trading on AI enthusiasm [1]. This differentiation thesis has important implications for portfolio construction, as it implies that active selection rather than broad sector exposure may be the optimal approach going forward.

The “phase two” thesis presented by van Eck represents a forward-looking view of AI commercialization that extends beyond the infrastructure and model development phase that dominated 2023-2024. This next stage supposedly focuses on wider enterprise and consumer adoption of AI applications, driven by what van Eck describes as “insatiable demand for tokens and compute power” [1]. The infrastructure requirements to support this demand have prompted VanEck’s interest in nuclear energy as an indirect AI beneficiary, recognizing that power generation capacity may become a binding constraint on AI scaling [1].

The nuclear sector emphasis is particularly noteworthy as it illustrates van Eck’s approach to identifying AI-adjacent opportunities that may be overlooked by mainstream investors focused on direct AI software and hardware plays. This infrastructure angle addresses a real structural concern in the AI expansion narrative—the massive energy requirements of data centers and compute clusters—and suggests that energy sector allocations may become increasingly relevant to AI-focused investment strategies [1].

The analysis reveals several factors that warrant investor attention when evaluating van Eck’s thesis. The most significant consideration is the inherent bullish bias that comes with any sector commentary from a fund manager who has built products around that sector. VanEck’s suite of AI-focused funds creates structural incentives to maintain constructive views on AI adoption timelines and commercial viability [1]. This conflict of interest does not invalidate the analysis but contextualizes its optimism and suggests the need for independent verification of key claims.

The “phase two” thesis, while compelling as a narrative, remains speculative pending evidence of actual commercial AI adoption at scale. The velocity at which enterprises integrate AI into operations, and the willingness of consumers to pay for AI-enhanced products and services, will ultimately determine whether phase two materializes as envisioned or encounters adoption headwinds similar to those that triggered the 2025 correction [1]. Investors should monitor earnings reports from major AI players in Q4 2025 and Q1 2026 to validate the “profitable AI” claims that distinguish this cycle from historical technology bubbles.

On the opportunity side, van Eck’s emphasis on indirect beneficiaries presents a potentially underappreciated allocation theme. His suggestion that “the biggest beneficiaries may not appear on investors’ radar” implies that nuclear energy, grid infrastructure, and related sectors may offer asymmetric risk-reward profiles as AI compute demand continues expanding [1]. The introduction of nuclear-focused ETFs by VanEck represents a concrete product response to this thesis and deserves monitoring for flows and performance attribution.

VanEck CEO Jan van Eck’s characterization of the AI sector’s 2025 correction as a completed “bubble burst” rather than an impending event provides important context for current market positioning. The distinction between this cycle and the dot-com era—namely, the profitability of current AI companies—represents a structural difference that supports the “healthy correction” narrative. Market data from early 2026 shows relative stability across major indices, with the S&P 500 and NASDAQ holding modest gains while small-cap indices outperform, potentially reflecting rotation into opportunities identified during the correction period [0].

The investment implications center on differentiation within the AI sector, with an emphasis on identifying winners with demonstrated earnings and cash flow rather than speculative names dependent on continued capital inflows. Indirect beneficiaries, particularly in energy infrastructure, may offer exposure to AI demand without direct AI sector concentration risk. The upcoming earnings season will provide critical data points to assess the validity of van Eck’s profitability claims and phase two adoption thesis [1].