Thematic Thursday: Nvidia’s Earnings Reset the AI Valuation Debate—but Is the Selloff Over?

Thematic Thursday: Nvidia’s Earnings Reset the AI Valuation Debate—but Is the Selloff Over?

Extraordinary earnings growth, sharply compressed valuations, and Anthropic’s record IPO are strengthening the case for accumulating oversold semiconductor equities. Yet massive infrastructure commitments, persistent financing risks, and uneven ETF flows suggest investors should remain selective.

Nvidia’s Earnings Challenge the Bearish AI Narrative

Nvidia’s (NVDA) third-quarter earnings delivered a powerful rebuttal to concerns that AI infrastructure demand is approaching a cyclical peak. Revenue increased 103% year over year to $115.8 billion, with Data Center revenue advancing 110% to $107.6 billion. Non-GAAP EPS reached $2.68, while management guided fourth-quarter revenue to approximately $130 billion and raised its fiscal 2028 revenue-growth outlook to 75% from 70%. Those are extraordinary growth rates for a company of Nvidia’s size, and they contrast sharply with the collapse in its valuation. According to FactSet, NVDA now trades at approximately 10.5 times next-twelve-month earnings, compared with a five-year average of 33 times. Its EV/EBITDA multiple has compressed to 8.1 times from a historical average of 27 times. Following the report, consensus fourth-quarter revenue and EPS estimates increased 2.4% and 3.0%, respectively, while longer-term forecasts also moved higher.

The valuation reset suggests investors are discounting a substantial slowdown in future profitability even as near-term operating fundamentals remain exceptional. Nvidia’s results are particularly important because the recent Technology correction has been driven less by evidence of deteriorating computing demand than by concerns about financing costs, capital intensity, export restrictions, and returns on infrastructure investment. The central question is whether Nvidia’s compressed multiple represents an opportunity to accumulate a highly profitable industry leader or whether today’s earnings forecasts are extrapolating an unusually favorable phase of the AI investment cycle. We favor selective accumulation, but the distinction between an attractive valuation and a confirmed market bottom remains important.

Semiconductor Flows Reveal Buying Into Weakness, but Not Yet a Durable Bottom

The latest thematic ETF data show substantial accumulation during the semiconductor correction, followed by significant withdrawals over the most recent week. SOXX and SMH have attracted approximately $7.54 billion in combined one-month inflows despite declines approaching 15%, but their latest weekly flows totaled roughly $4.01 billion in net redemptions. That divergence suggests investors have been willing to establish positions at lower prices without demonstrating the sustained conviction normally associated with a durable recovery.

ETF / Theme 1W return 1W flows 1M return 1M flows
SOXX — Semiconductors -8.5% -$4,754M -13.8% +$2,546M
SMH — Semiconductors -8.4% +$743M -14.9% +$4,998M
XSD — Semiconductors -9.1% +$68M -10.3% -$100M
IGV — Software -4.5% -$410M -7.1% -$1,663M
HACK — Cybersecurity +2.2% +$11M +2.1% +$73M
VFLO — Free Cash Flow -5.2% +$286M -6.5% +$1,230M
SCHD — Dividend Equity -3.4% +$866M +1.0% +$3,949M

Source: FactSet Research Systems Inc., November 19, 2026. ETF returns and net flows through November 18. Flows rounded to the nearest million.

Technical conditions reinforce the possibility of a near-term rebound. SMH’s relative strength index is near 29.5, SOXX’s near 32.8, and the broader Technology ETF VGT’s near 28.5. Semiconductor valuations have contracted substantially, positioning has become less crowded, and Nvidia’s earnings surprise provides a credible catalyst for reassessing the industry’s growth outlook. Nevertheless, oversold conditions and positive rolling monthly flows do not establish that institutional accumulation has resumed. The continued withdrawals from Software and concurrent inflows into dividend and free-cash-flow strategies indicate investors remain focused on financial resilience rather than embracing speculative Growth indiscriminately. SCHD has attracted nearly $4 billion over one month, while VFLO received $1.23 billion, illustrating the continuing appeal of established cash generation during the AI correction.

Anthropic’s IPO Demonstrates Capital Availability—and the Scale of Future Financing Requirements

Anthropic’s (ANT) record IPO provides a second major catalyst. The company priced approximately 980 million shares at $83, above the original offering range, raising $81.34 billion in the largest public listing on record. Including a concurrent $26 billion private placement, the transaction generated approximately $107 billion in gross proceeds. The offering places Anthropic’s fully diluted valuation near $1.5 trillion and demonstrates that investors remain willing to commit extraordinary amounts of capital to leading AI developers despite the recent market correction.

However, the company’s prospectus also illustrates why financing remains central to the investment debate. Anthropic generated approximately $17 billion in third-quarter revenue but recorded a $567 million operating loss, reflecting substantial computing and infrastructure expenditures. More consequentially, the company disclosed approximately $900.9 billion in future computing and infrastructure commitments, including $70.7 billion due within twelve months of September 30, 2026. The IPO materially strengthens its liquidity, but those proceeds represent only a fraction of the commitments associated with its planned expansion. At approximately 18.5 times its reported October annualized revenue run rate of $81 billion, Anthropic’s valuation also embeds considerable expectations for sustained growth and future profitability.

The contrast with Nvidia is striking. Anthropic is attracting capital to finance a business whose profitability remains dependent on rapidly expanding demand, improving computing efficiency, and increasingly competitive AI pricing. Nvidia already generates substantial operating profits and cash flow from supplying that infrastructure. The comparison is not directly captured by their headline valuation multiples—one is revenue-based and the other earnings-based—but it highlights where investors are currently assuming the greatest financial and execution risks. The offering also includes strategic participants from the AI ecosystem, underscoring that capital is circulating among suppliers, customers, and developers rather than representing exclusively independent outside demand.

Anthropic’s IPO is a constructive financing event, not proof that the entire AI infrastructure complex can fund its expansion on attractive terms. Its subsequent trading performance and the conversion of new funding into sustainable earnings will provide more meaningful evidence of investor conviction.

The Opportunity in Semiconductors Is Real, but So Are the Earnings Risks

Several developments support accumulating oversold semiconductor leaders. Nvidia’s extraordinary revenue growth, higher forward guidance, and continuing demand for advanced computing challenge the assumption that AI infrastructure spending is about to contract sharply. The company’s $48.3 billion in quarterly operating cash flow demonstrates substantial internal financing capacity, while the upcoming Vera Rubin product transition and multiyear customer relationships provide visibility into future demand. Valuation compression further improves the prospective risk/reward if earnings forecasts remain achievable. Established suppliers such as Broadcom (AVGO), Applied Materials (AMAT), and selected memory manufacturers may also offer opportunities where current earnings, competitive positions, and cash generation support their valuations.

The bearish case is more complicated than a simple slowdown in chip purchases. Nvidia disclosed $518 billion in future commitments, including $390 billion primarily associated with supply and capacity procurement and $102 billion in cloud agreements and data-center leases. These are multiyear commitments rather than immediately payable debt, but they substantially increase the importance of utilization, future demand, and supply-chain economics. Separately, the company disclosed $32 billion of contingent neocloud obligations and up to $118 billion in maximum gross credit guarantees. Management has emphasized that some arrangements depend on customers obtaining financing and can be reduced or eliminated if financing requirements are not met. That discipline protects Nvidia, but potentially reduces capital availability elsewhere in the ecosystem.

Profitability also warrants scrutiny. Nvidia’s third-quarter non-GAAP gross margin reached 74.9%, but management guided fourth-quarter margins toward 71%–72%, with improvement toward 72%–73% anticipated during fiscal 2028. Those remain extraordinary levels, although memory procurement costs, product transitions, competitive pressure from custom silicon and AMD accelerators, and possible restrictions on Chinese demand could alter the earnings trajectory. The leadership transition introduces additional uncertainty regarding customer financing, employee retention, and the execution of Nvidia’s longer-term roadmap.

The macroeconomic environment is another constraint. The 10-year Treasury real yield remained near 2.88% on November 18, while U.S. high-yield credit spreads widened from 323 basis points on November 13 to 343 basis points on November 17. Higher real rates compress the valuation of future earnings, while wider credit spreads increase the actual cost of financing capital-intensive infrastructure. These conditions can offset improvements in semiconductor fundamentals, particularly if developers begin postponing or reducing investments.

The central vulnerability is therefore not necessarily that Nvidia’s earnings are about to decline. It is that today’s exceptionally strong forecasts depend on an infrastructure expansion whose future financing and profitability remain uncertain. What appears inexpensive on forward earnings can become substantially less attractive if those earnings estimates are revised downward.

Investment Outlook: Begin Accumulating, but Preserve Portfolio Flexibility

We believe the correction has created a legitimate valuation opportunity in established semiconductor businesses, particularly where market prices have fallen considerably faster than earnings expectations. NVDA, AVGO, and AMAT are among our preferred research candidates for selective accumulation, while Micron (MU) also warrants attention given structural demand for advanced memory. However, memory and storage companies require additional discipline because exceptionally favorable pricing can produce temporarily depressed earnings multiples near cyclical profitability peaks.

For diversified implementation, SOXX provides relatively broad semiconductor exposure, while SMH offers greater concentration in leading AI companies. XSD supplies a more equal-weighted alternative but introduces greater exposure to smaller manufacturers and potentially weaker balance sheets. We would favor staged purchases rather than aggressively increasing exposure following a single earnings surprise or a technically oversold reading. Stabilizing weekly fund flows, improving semiconductor-relative performance, and continued positive earnings revisions would provide confirmation that accumulation is becoming more durable.

Beyond semiconductors, Cybersecurity (HACK) offers selective exposure to enterprise AI adoption and security needs without the same direct dependence on physical data-center expansion. Nevertheless, our broader portfolio preference remains a diversified combination of Value, Quality, and Lower Volatility exposures, complemented by measured Technology allocations. The recent recovery in financial and defensive factors should not be abandoned simply because AI valuations have become more compelling.

Our conviction in a sustained AI recovery would strengthen materially if inflation moderated, real Treasury yields declined, and investment-grade and high-yield credit spreads tightened. That combination would improve both equity valuation support and the economics of financing future infrastructure. Until those conditions align, an AI rebound may remain vulnerable to further financing disruptions and changes in investor risk appetite.

Conclusion: Attractive Valuations, but Financial Conditions Still Matter

Nvidia’s earnings demonstrate that the AI investment cycle remains fundamentally powerful, while its valuation reset provides a credible basis for accumulating selected semiconductor leaders. Anthropic’s record IPO confirms that large amounts of capital remain available for established AI franchises, but its enormous infrastructure commitments highlight how much future growth depends on continued investor financing.

We favor selective accumulation of oversold semiconductor leaders rather than an indiscriminate increase in Technology exposure. The opportunity is strongest where companies combine earnings visibility, durable competitive positions, operating cash flow, and manageable capital requirements. The principal risk is that financing constraints eventually undermine the infrastructure expansion supporting today’s extraordinary earnings forecasts.

The market has already priced considerable pessimism into Nvidia and parts of the semiconductor industry. That makes the opportunity more compelling, but the conditions for a sustained recovery remain incomplete.

Nvidia has demonstrated that AI demand is not the immediate problem. The next durable advance requires evidence that the industry’s financing arrangements and future profitability can support that demand at today’s valuations.

 

Sources and References

Market data: FactSet Research Systems Inc., November 19, 2026 thematic ETF returns and fund flows; Nvidia third-quarter earnings brief, FactSet consensus estimates, valuation data, and sell-side research.

Primary corporate disclosures: Nvidia Third-Quarter Fiscal 2027 Results; Nvidia SEC-Filed CFO Commentary; Anthropic Final IPO Prospectus; Anthropic IPO Pricing Filing.

Economic and credit data: U.S. Treasury Daily Real Yield Curve Rates; ICE BofA U.S. High Yield Option-Adjusted Spread through FRED; Bureau of Labor Statistics, October CPI.

Additional context: Reuters and Financial Times reporting, November 17–19, 2026, concerning AI financing, semiconductor demand, export restrictions, and Anthropic’s public offering.

Disclaimer:  This publication is provided by ETFSector.com for informational and research purposes only and does not constitute individualized investment advice or a recommendation to purchase or sell securities. ETF data originate from FactSet Research Systems Inc. Forward earnings multiples depend on estimates that may change materially. Fund flows and historical valuations do not guarantee future performance. Semiconductor and thematic investments involve concentration, valuation, financing, technological, regulatory, and market risks.

 

Patrick Torbert

Editor | Chief Strategist

Patrick Torbert is a veteran financial market analyst who is currently the Editor and Chief at ETF Insight a NY based full-service content, TV, video podcast and digital marketing firm that represents several ETF issuers. Patrick brings 20+ years of experience from Fidelity Asset Management where he most recently served as an equity and multi-asset analyst.
Scroll to Top

Subscribe to our Newsletter

Stay updated with the latests analysis and insights from etfsector.com

If you haven’t received your newsletter email, check your spam/junk folder and add us to your contacts to ensure delivery.