Euphoria, energy and the coming volatility regime.
The May 2026 research note reviews extreme concentration in US equities, record retail options leverage, the unfolding energy shock and its inflation pass-through, the unwinding of the AI subsidy model, and the resulting portfolio positioning across gold, Swiss equities, Chinese equities, Brazilian equities and US fixed income.
Download the full research note (PDF, 4 pages) →
Executive Summary
The S&P 500 closed at a fresh all-time high of 7,444.70 on 13 May 2026, up 8.75% YTD on a price basis, but the surface narrative obscures an unhealthy structure underneath: extreme concentration in Nvidia and Alphabet, record retail-driven 0DTE options leverage, and a market capitalization-to-GDP ratio of 232%, the highest in history. Meanwhile, an energy shock from the closed Strait of Hormuz is now feeding directly into both CPI (3.8% YoY) and PPI (6.0% YoY), and AI operating companies remain structurally unprofitable while their subsidy-led adoption model unwinds. We expect higher US interest rates, materially higher US equity volatility through 2026, and continued outperformance from gold, Swiss equities, Brazilian equities and selectively Chinese equities. Fixed income remains unattractive at current entry points.
1. A Market Running on Two Stocks and Same-Day Options
Behind the headline 8.75% S&P 500 YTD gain (closing 7,444.70 on 13 May, up from 6,845.50 on 31 December 2025), the engine driving the index is narrower than at any point in modern history. Through 13 May, Alphabet is up 27.1% YTD and Nvidia is up 15.4% YTD. Both posted fresh closing all-time highs this week. Their combined market capitalization stands at $10.24 trillion (Nvidia $5.36 trillion, Alphabet $4.88 trillion), or roughly 13.9% of total US equity market value. A pair of stocks now represents more market value than the entire DAX 40, FTSE 100 and CAC 40 combined.
The 2025 picture, for reference, was already dominated by the same two names: per S&P Dow Jones Indices, Nvidia contributed 15.5 and Alphabet 13.5 percentage points to the index's 17.88% total return, meaning two stocks delivered 29% of the gain. Strip out the Magnificent Seven and 2025 collapses from 17.9% to 10.4%. In 2026, the rotation within that concentration has if anything intensified: Alphabet is now within striking distance of overtaking Nvidia as the world's largest company, with the market-cap gap closing from over $500 billion three months ago to roughly $480 billion at this week's close.
The plumbing: 0DTE options and retail leverage
How this rally is being expressed structurally is even more revealing than what is rallying. Per Cboe Global Markets, in 2025 zero-day-to-expiration (0DTE) options on the S&P 500 averaged 2.3 million contracts per day, 59% of total SPX options volume, with monthly peaks above 62%. Total US listed options volume reached a record 15.2 billion contracts in 2025, up 26% year on year and the sixth consecutive record year. Retail traders now account for an estimated 50% of all US options activity and 53 to 54% of 0DTE flow. Citadel Securities reports that nearly 40% of total retail options volume on its platform is in zero-day contracts. The rally is being underwritten by leverage on intraday convexity from a retail cohort that has scaled into the most fragile instruments in the listed market.
Concrete evidence of the fragility: on 1 April 2026, a documented short gamma squeeze, with dealers carrying an estimated minus $7.5 billion in net gamma exposure, drove the S&P 500 over the 6,500 level in a single session. The same mechanism that powers the rally on the way up will accelerate any drawdown on the way down.
The Buffett indicator: 232% of GDP, an all-time record
The total US stock market capitalization-to-GDP ratio, popularly known as the Buffett indicator, stood at 232.1% as of 13 May 2026 (GuruFocus). This is the highest reading in the history of the series, approximately 2.4 standard deviations above the long-run trend. For context, the dot-com peak in March 2000 reached 146%, and the pre-GFC peak in October 2007 was 109%. With Q1 2026 GDP at $31.86 trillion in current dollars, the US equity market is valued at approximately $74 trillion. The structural implication is unambiguous: the United States has, mathematically, become an economy whose private wealth, consumer confidence, corporate financing conditions and tax base are all critically dependent on the stock market rather than on real-economy sectors. Manufacturing, tourism, energy production and traditional services no longer dominate the marginal pricing of US prosperity. A handful of AI-adjacent equities do.
2. The Missing Input: Energy and the AI Cost Stack
The bull case for the AI buildout focuses almost exclusively on output: chip sales, data center revenue, cloud backlog. This framing systematically ignores the input side, where the largest input cost is energy.
As warned in our previous Bloomberg interview at the outset of the US-Iran war, even an immediate diplomatic resolution would not restore normal energy flows for at least six months. That timeline is now corroborated by Saudi Aramco CEO Amin Nasser, who stated on 11 May 2026 that "if the Strait of Hormuz opens today, it will still take months for the market to rebalance, and if its opening is delayed by a few more weeks, normalization will last into 2027". DHL Global Forwarding has separately advised clients to plan for four to six months of shipping disruption. More than 600 tankers remain trapped in the Persian Gulf, with 240 more idling outside.
The IEA has described the disruption as "the largest supply disruption in the history of the global oil market". Global oil supply fell 10.1 million barrels per day in March 2026. Brent peaked at $138 per barrel on 7 April, averaged $117 in April, and closed near $106 yesterday. WTI settled at $102.18 on 12 May. Saudi production has dropped to its lowest level since 1990. The UAE formally exited OPEC on 1 May 2026.
The energy shock is now visibly translating into the official inflation data released this week, and we believe this is only the beginning:
| Indicator | April 2026 | Context |
|---|---|---|
| Headline CPI YoY | 3.8% | Highest since May 2023; +0.5pp from March |
| Core CPI YoY | 2.8% | Well above Fed 2% target |
| Energy CPI YoY | +17.9% | Gasoline +28.4% YoY; over 40% of headline gain |
| Headline PPI MoM | +1.4% | Largest monthly gain since March 2022 |
| Headline PPI YoY | +6.0% | Biggest annual rise since December 2022 |
| PPI Final Demand Energy | +7.8% MoM | Gasoline +15.6% MoM at producer level |
| Core PCE Q1 2026 (annualized) | +4.3% | Sharpest quarterly gain since Q1 2023 |
Sources: BLS, BEA. CPI released 12 May 2026, PPI released 13 May 2026.
Markets have already repriced. Per CME FedWatch, the probability of any Fed rate cut in 2026 has been priced out; the implied probability of a Fed rate hike by December 2026 now sits near 39%, with over 70% probability of a hike by April 2027. The 30-year Treasury auction yesterday cleared near 5% for the first time since 2007. Kevin Warsh was confirmed as Fed Chair by the Senate yesterday afternoon (54-45, the slimmest margin in Fed history); he inherits a $6.7 trillion balance sheet, an energy-led inflation shock and an administration politically dependent on lower rates.
AI's energy bill is unpriced in current equity valuations
This matters acutely for AI. Data centers are among the most energy-intensive infrastructure ever deployed. Microsoft signaled this exposure when its stock dropped 12% in a single session on 29 January 2026, erasing $440 billion in market value, after disclosing the scale of its AI capex commitments. The chipmakers, Nvidia in particular, continue to monetize the buildout. The operating AI companies face a very different reality.
3. The Subsidy Trap: AI Adoption is Not Profitable
The economics of foundation model providers are starkly worse than the equity market is currently pricing in.
- OpenAI: Projected to lose $14 billion in 2026 on roughly $20 billion in annualized revenue, with operating losses widening to $74 billion in 2028 before meaningful profitability is targeted for 2029-2030. Per leaked Microsoft revenue-share data, OpenAI burns $2 for every $1 earned on inference alone, before R&D, sales and marketing. Only 5.5% of ChatGPT's roughly 900 million weekly active users pay for a subscription.
- Anthropic: More disciplined cost trajectory, but still cash-burning. Plans to spend approximately $12 billion training models and $7 billion on inference in 2026, against an $18 billion revenue target. Projects positive cash flow by 2027- 2028.
- Wrappers and integrators: Cursor (Anysphere) sends essentially 100% of its revenue back to Anthropic as compute cost. Perplexity spent 164% of its 2024 revenue on AWS, Anthropic and OpenAI compute.
Developer adoption, the central thesis behind the AI revenue ramp, has been driven by aggressive subsidization. A developer paying $100 to $200 per month for an AI coding subscription has, by the providers' own admission, been consuming many multiples of that figure in compute. Anthropic's own public statement on the rate-limit rollout (28 July 2025) acknowledged that "a small fraction of users were consuming thousands of dollars' worth of compute under low-cost subscriptions". The party is now ending. Anthropic introduced weekly usage caps on 28 August 2025 and has since tightened them. Pro subscribers ($20/month) get 40 to 80 hours of Sonnet 4 per week; $100 Max plans 140 to 280 hours; $200 Max plans 240 to 480 hours, with surplus usage billed at API rates.
The implication is consequential. Once subsidies are removed and usage is priced at marginal cost, the economic case for replacing a human developer with AI tooling narrows materially. At full inference pricing, including the 1 million token context window beta tier at $10/$37.50 per million tokens for Opus, a heavy user can easily consume $1,000 to $5,000 in monthly compute. A junior developer at that cost point is once again competitive, even after factoring in productivity gains.
This is the fundamental fragility in the current AI equity narrative. Hardware providers are selling shovels into a gold rush whose miners are still net loss-making. When the cost of the dig becomes visible, demand for shovels normalizes.
4. Portfolio Positioning
Gold remains our largest single allocation
Gold currently represents 25% of our model portfolio, our largest position. The spot price closed at $4,696 per ounce on 13 May 2026, up 8.4% YTD. Year-to-date, gold and the S&P 500 are running essentially neck-and-neck (gold +8.4%, S&P 500 +8.75%), but over longer horizons gold's structural outperformance is unambiguous:
| Horizon | Gold (spot) | S&P 500 (price) |
|---|---|---|
| YTD 2026 | +8.4% | +8.75% |
| 1 year (May 2025 to May 2026) | +45.3% | +26.5% |
| 3 years (May 2023 to May 2026) | +133.5% | +80.5% |
| 5 years (May 2021 to May 2026) | +157.5% | +81.0% |
Sources: spot gold (TradingEconomics, Fortune), S&P 500 (FRED, Yahoo Finance, S&P Dow Jones Indices). Price returns shown; total return including dividends would lift S&P 500 by roughly 1.3% to 1.5% per annum.
Drivers remain structurally intact: central bank accumulation (Poland led February 2026 purchases at 20 tons; China continues to add), de-dollarization, persistent inflation surprises, and a fiscal trajectory the Congressional Budget Office projects will widen the structural deficit by approximately 600 basis points of GDP. Morgan Stanley, JP Morgan, Deutsche Bank and Citi all have year-end targets ranging from $5,000 to $6,300 per ounce.
Equity exposure: away from the US
- Swiss equities: SMI +9.8% YTD (from 11,949 at year-end 2025 to 13,119 on 12 May 2026), narrowly outpacing the S&P 500 in CHF terms while offering defensive cash-generative exposure through Nestlé, Roche and Novartis. Swiss franc strength acts as a parallel hedge against US fiscal deterioration. The August 2025 US tariff shock on Swiss exports created an attractive re-entry point in selected names that has now partially closed.
- Chinese equities: Hang Seng +3.7% YTD (from 25,630 to 26,576) and Shanghai Composite +5.7% YTD (from approximately 3,974 to 4,199, near its highest level since July 2015). The Hang Seng trades at a P/E of 12.9, less than half the US market's multiple. The 10-11 May 2026 US-China tariff truce (US duties on Chinese imports cut from 145% to 30% for 90 days) materially improves the near-term outlook. Selective exposure with active management.
- Brazil: Ibovespa +9.9% YTD (from 161,125 to 177,098), having pulled back about 11% from the 14 April all-time high at 199,355. It trades at a trailing P/E of approximately 11 and forward P/E of 9, a deep discount to most developed and emerging peers. Brazil benefits as a major commodity exporter (oil, soft commodities, iron ore) precisely when the energy and food cost cycle is shifting in producers' favor. The strong real (BRL appreciated 11.2% against USD in 2025) adds USD-translated returns for foreign investors.
Fixed income: avoid
With Fed funds at 3.50% to 3.75%, headline CPI at 3.8%, PPI at 6.0%, the Strait of Hormuz situation still unresolved, the 30-year Treasury auctioning at near 5% yesterday and the market now pricing meaningful probability of further hikes rather than cuts, US fixed income duration is asymmetrically unattractive at current entry points. The risk-reward favors waiting.
Volatility outlook: meaningfully higher through 2026
The conditions for sustained higher US equity volatility are now structurally in place: extreme concentration in two names, record options leverage, 0DTE-driven gamma fragility, an energy-led inflation impulse that the Fed cannot fight without breaking growth, a newly confirmed Fed Chair committed to shrinking a $6.7 trillion balance sheet, and the unwinding of cross-border carry positions financing US risk assets. We expect materially higher realized and implied volatility in US equities through the balance of 2026, with sharp drawdowns increasingly likely as the gap between price action and earnings durability widens.
5. Conclusion
The current US equity rally has the hallmarks of late-cycle euphoria: extreme concentration, retail leverage on the most fragile derivatives instruments, a market capitalization-to-GDP ratio at an all-time high, a hardware boom whose end users have yet to achieve profitability, a subsidy model that is now being unwound, and an energy shock that is just beginning to show up in headline inflation. The bull case requires every single one of these to resolve favorably. We do not believe that is the most probable outcome.
Our positioning reflects this view: an oversized gold allocation, selective equity exposure to markets with reasonable valuations and structural tailwinds (Switzerland, Brazil, China), avoidance of US fixed income duration at current entry points, and a clear expectation of higher US equity volatility through 2026.
Important Information
This document is a research note prepared by Ryan Lemand, Founder and CEO of the Neovision Group, for institutional and professional audiences. It does not constitute investment advice, an offer to sell, or a solicitation to buy any financial instrument. All market data are sourced from publicly available data as of the close on 13 May 2026, including but not limited to S&P Dow Jones Indices, FRED, Cboe Global Markets, BLS, BEA, GuruFocus, World Gold Council, EIA, IEA, TradingEconomics, Yahoo Finance, and direct disclosures by Anthropic, OpenAI and Saudi Aramco. Past performance is not indicative of future results. Neovision Wealth Management Limited is regulated by the FSRA in ADGM. NIFM is regulated by the CMA in onshore UAE. Héritage Riviera SA is authorized by FINMA in Switzerland.
This document is prepared for institutional and professional audiences. It does not constitute investment advice, an offer to sell, or a solicitation to buy any financial instrument. Past performance is not indicative of future results.