Vanguard VGT May Underserve AI Powerhouses

Vanguard's Information Technology ETF (VGT) has long been a core allocation for institutional portfolios seeking broad exposure to U.S. technology equities, but market structure and sector classification mean it may not provide the concentrated exposure to AI leaders that some investors expect. On May 9, 2026, Yahoo Finance published a note highlighting that several of the largest companies driving generative AI adoption—Alphabet (GOOG), Meta Platforms (META) and Amazon.com (AMZN)—are classified outside the Information Technology sector and therefore are excluded from VGT's index (Yahoo Finance, May 9, 2026). That sector exclusion is consequential: together these companies accounted for more than $3.1 trillion of combined market capitalization in the first quarter of 2026 (MarketCap Aggregates, Q1 2026), leaving a sizable portion of AI-related market value outside a technology-only ETF. For institutional investors seeking targeted AI exposure, the mechanics of index construction — not only stock selection — matter materially to portfolio outcomes.

Context

VGT tracks an MSCI-derived information technology universe designed to capture companies classified within the Information Technology sector (Vanguard factsheet, Apr 30, 2026). The construction rule is explicit: inclusion is determined by sector classification frameworks (GICS/MSCI), so firms whose primary business is categorized in Communication Services, Consumer Discretionary or Industrials are not eligible even if they are prominent AI developers. That distinction is not academic. Alphabet and Meta lead in large language model research and infrastructure; Amazon supplies cloud GPU capacity and AI services via AWS, yet those corporate labels sit outside the IT sector and therefore outside VGT's scope (SEC filings and company 10-Qs, 2025–Q1 2026).

Historically, sector-based ETFs have tracked industrial realities rather than technology capability. The Global Industry Classification Standard (GICS) reclassification in 2018 consolidated several digital-first businesses into Communication Services and Consumer Discretionary, a structural change that still influences index construction today. For funds tied to those classifications, the result is a divergence between market perception of an "AI leader" and index eligibility. As of May 2026, Vanguard's published methodology remains aligned to the GICS/MSCI taxonomy (Vanguard methodology page, Apr 30, 2026), and there has been no announced shift to reconstitute VGT's eligibility rules to capture AI leaders outside Information Technology.

This structural issue is of particular relevance now because corporate spending on AI infrastructure leapt in 2024–2025 and continued into early 2026: public filings show that Alphabet, Meta and Amazon each raised AI-related capex by double-digit percentages year-over-year in 2025 (company 10-Ks, FY2025). The market has reacted; those three names delivered outsized returns relative to broad tech indexes during 2024–2025, materially altering the contribution of AI-specific companies to total market performance. That creates a strategic mismatch for investors using sector-labelled ETFs as a proxy for AI exposure.

Data Deep Dive

Three specific data points illuminate the scope of the disconnect. First, Yahoo Finance flagged on May 9, 2026 that VGT’s sector definition excludes Alphabet, Meta and Amazon despite their central roles in generative AI development (Yahoo Finance, May 9, 2026). Second, combined market capitalization for those three firms exceeded $3.1 trillion as of March 31, 2026 — representing roughly 6–7% of total U.S. equity market cap depending on the weighting methodology (MarketCap Aggregates, Q1 2026). Third, Vanguard’s own factsheet (Apr 30, 2026) shows that VGT’s top 10 holdings concentrated approximately 52% of the fund’s assets under management, indicating both concentration risk and the impact of classification on which large-cap names anchor the ETF.

Comparing performance lenses further clarifies the divergence. Year-to-date through April 30, 2026, an AI-focused basket of the leading generative-AI integrators (a composite of NVDA, META, GOOG, AMZN and MSFT) outperformed VGT by approximately 8 percentage points (index returns, Jan–Apr 2026). NVDA and MSFT are entirely eligible for VGT due to their classification in Information Technology, which explains part of VGT's indirect exposure to AI. However, the absence of Meta, Alphabet and Amazon from VGT's holdings creates a measurable underweight versus a full-market view of AI leadership. Institutional investors comparing VGT to bespoke AI baskets or multi-sector strategies should therefore adjust expectations accordingly.

Sources for these datapoints include the Yahoo Finance piece (May 9, 2026), Vanguard factsheet (Apr 30, 2026) and consolidated market-cap datasets through Q1 2026. Where possible, practitioners should cross-check with primary filings: company 10-Ks and 10-Qs for capex and segment disclosure, and MSCI/GICS documentation for sector classification rules (GICS/ MSCI releases, 2018 and updates through 2025).

Sector Implications

The sector-classification gap has direct implications for portfolio construction and active vs passive allocation decisions. For portfolio managers building AI tilts, relying solely on a sector ETF like VGT will bias the portfolio toward companies whose primary SIC/GICS codes are technology — for example, semiconductors and enterprise software — while de-emphasizing platform owners and cloud infrastructure providers that sit in other sectors. That matters because many AI revenue streams and margin expansions accrue through cloud services (AWS, Google Cloud), advertising and social platforms (Meta, Alphabet) and large-scale retail/commercial deployments (Amazon). Excluding those revenue streams reduces the representativeness of VGT as an "AI ETF."

The peer set comparison is instructive. The Invesco QQQ Trust (QQQ) and S&P 500 (SPX) include Alphabet, Meta and Amazon and therefore capture a broader slice of AI-related market cap. Over rolling 12-month windows in 2024–2026, QQQ and SPX exhibited a higher correlation to an AI composite index than VGT did, primarily because of the presence of platform companies. Conversely, VGT may outperform in environments where semiconductors and enterprise software lead — for example, during capital-cycle recoveries where hardware demand re-accelerates. These divergent exposures mean institutions should decompose "AI exposure" into sub-themes (semiconductors, cloud providers, platform monetization, AI services) and choose vehicle(s) accordingly.

From an index-provider vantage point, there are potential product responses: create a cross-sector "AI leaders" index, launch thematic ETFs that cut across GICS buckets, or adjust sector definitions — each with trade-offs in investability, turnover and regulatory status. Vanguard has historically prioritized low-cost, rules-based strategies; any thematic pivot would require altering prospectuses and potentially increasing turnover and tracking error. Institutional allocators must therefore weigh the governance and cost implications of substituting or supplementing VGT with alternative vehicles that capture multi-sector AI exposure. For further reading on how sector and thematic ETFs differ operationally, see Fazen Markets resources and our longer methodology primer on ETF construction.topic

Risk Assessment

Using VGT as a proxy for AI exposure creates two primary risk vectors: omission risk and concentration risk. Omission risk arises because large-cap platform companies that are central to AI monetization are excluded by construction. That omission introduces model risk for systematic strategies that treat VGT as representative of the technology ecosystem. Concentration risk stems from VGT's top-heavy composition: when 50%+ of assets reside in the top 10 positions (Vanguard factsheet, Apr 30, 2026), idiosyncratic moves in those names drive fund performance more than the broader technology cycle.

Another risk to consider is reclassification and turnover. If firms reclassify or if index providers adjust sector boundaries, tracking error can spike during rebalances. Historical reclassification events (GICS changes in 2018) produced meaningful turnover in affected ETFs; similar events could disrupt strategies that assume static sector membership. Liquidity risk is more muted for VGT given its large AUM and tight spreads, but for large institutional trades, the intraday liquidity of underlying constituents — particularly less-liquid mid-cap software firms — remains a practical constraint.

Finally, there is a valuation-risk overlay. Many AI-relevant firms have experienced rapid multiple expansion: a concentrated holding in a high-multiple software or semiconductor name can lead to heightened portfolio volatility. For investors focused on factor exposures (growth vs value, momentum), substituting or augmenting VGT with other vehicles will change factor tilts materially. These are measurable and should be stress-tested in scenario analysis before reallocation.

Fazen Markets Perspective

A non-obvious but practical implication is that sector labels have become less informative for thematic investing as corporate business models diversify. For institutional investors, the efficient path to targeted AI exposure is seldom a single-sector ETF. Instead, a blended approach — combining a tech-sector core (e.g., VGT for semiconductor and software exposure) with a cross-sector AI overlay that explicitly includes Alphabet, Meta and Amazon — delivers a truer representation of the AI ecosystem while preserving cost control and governance. This hybrid approach is not about "timing" but about precision: allocating by economic exposure rather than by historical sector taxonomy.

We also note that active managers who can navigate multi-sector exposure with disciplined risk controls may offer better trade-offs for capturing AI-driven alpha than a reconstituted passive instrument. That said, active mandates come with higher fees and governance complexities. For large allocations, a staged implementation — small overlay positions in targeted ETFs or futures, combined with cash-flow-aware rebalancing — often outperforms abrupt shifts in benchmarked allocations. See our implementation checklist on thematic overlays at Fazen Markets institutional guides.

Outlook

Over the next 12–18 months, expect the divergence between sector-based ETFs and AI market leadership to persist unless index providers materially revise classification rules. Corporate capex patterns and AI revenue recognition will continue to shift market leadership; companies traditionally outside the Information Technology classification will likely capture a larger share of AI economic profits. For those reasons, passive sector allocations will need supplementary strategies to maintain exposure to the fastest-growing nodes of AI adoption.

Institutional investors should also monitor potential product innovation from major ETF issuers. The demand signal for cross-sector AI exposures is clear, and we expect new thematic ETFs or factor-enhanced strategies to compete on fee and tracking characteristics. Regulatory scrutiny on thematic marketing and index construction will intensify as products proliferate, creating a window where careful due diligence — on index methodology, turnover, and liquidity — can yield implementation advantages.

FAQ

Q: If VGT excludes Alphabet, Meta and Amazon, which ETFs capture those names?

A: Broad market ETFs such as the S&P 500 (SPY) and Nasdaq-100 (QQQ) include Alphabet, Meta and Amazon, as do sector-agnostic large-cap funds. Several thematic ETFs launched since 2023 also include cross-sector AI leaders; institutional investors should review prospectus methodologies for eligible sectors and rebalance mechanics to ensure alignment with objectives.

Q: How should institutions measure "true" AI exposure?

A: Best practice is to define exposure by economic activity (revenue sources, AI-related capex, cloud/GPU supply chains) rather than by sector labels alone. Create baskets that allocate to sub-themes — semiconductors (NVDA, ASML), cloud infrastructure (AMZN, GOOG), platform monetization (META) and enterprise software (MSFT, SAP) — and then test correlations to your target AI composite over multiple market regimes.

Bottom Line

VGT remains a low-cost, rules-based vehicle for Information Technology exposure, but its sector-based construction systematically omits several of the largest commercial AI players; institutions aiming for comprehensive AI coverage should supplement VGT with cross-sector instruments or bespoke overlays.

Disclaimer: This article is for informational purposes only and does not constitute investment advice.