CoinMarketCap research head Alice Liu distinguishes between tokenized Nvidia Corp (NASDAQ: NVDA) products and standalone AI tokens, noting that the former offers direct exposure to the chipmaker’s equity performance.

Liu, Head of Research at CoinMarketCap, told Benzinga that tokenized stocks provide an "honest way" for global retail investors to access Nvidia with fractional sizing and leverage outside traditional market hours. She emphasized that an AI token represents a separate investment bet, despite often sharing thematic branding.

Tracking Performance vs Thematic Speculation

The distinction is critical for understanding return profiles. A tokenized Nvidia product is designed to mirror the underlying stock. CoinMarketCap’s NVDAX was roughly flat over the month covered by the interview, broadly matching Nvidia’s actual share price movement.

In contrast, AI tokens exhibit high volatility and divergent returns driven by crypto market dynamics rather than corporate fundamentals. As of Sept. 10, TAO gained 24% over 30 days, while Render fell 63% over the year.

Market Scale for Tokenized Equities

Tokenized equities are gaining traction as an access layer for traditional assets. CoinMarketCap tracked 1,851 instruments in this category, representing approximately $2.3 billion in market capitalization and roughly $2 billion in daily volume.

What the Numbers Show

The data reveals a sharp divergence in risk-return profiles between asset classes. While NVDAX remained flat—mirroring the underlying equity—TAO surged 24% in just 30 days. This highlights that AI tokens trade on speculative momentum rather than operational performance, offering returns completely decoupled from Nvidia’s financial results.

Altimeter Capital founder Brad Gerstner stated that leading AI labs must raise their combined annualized revenue from roughly $100 billion to at least $180 billion by year-end to maintain the current AI investment thesis.

Gerstner, whose firm held nearly $1.9 billion of Nvidia Corp. (NASDAQ: NVDA) shares as of June 30, identified AI lab revenue as the critical market data point during the All-In Summit. He argued that this revenue growth is essential to support the infrastructure spending of major technology firms.

Revenue Targets and Market Expectations

Gerstner estimated that Anthropic, OpenAI, and SpaceX, which owns xAI, had a combined run rate of roughly $100 billion based on July figures. He projected these companies need to reach $180 billion by year-end, an increase of about 80% from his earlier estimate, just to keep the AI trade intact.

He noted that investors are already trading on these figures. Anthropic's revenue run rate reportedly reached $65 billion in July, up from $47 billion in May. Gerstner attributed the spring rally to Anthropic's monthly revenue but noted stocks consolidated after the $65 billion figure fell short of the roughly $75 billion investors had expected.

Capex Justification

Gerstner linked the required revenue growth to the massive capital expenditure by Microsoft Corp. (NASDAQ: MSFT) and Alphabet Inc. (NASDAQ: GOOGL). He stated that if these companies are building $1.5 trillion a year in capex, someone must pay for it.

He clarified that Microsoft and Google are building computing capacity to rent out to AI labs like OpenAI and Anthropic, not paying for it themselves. The labs need sufficient revenue from consumers and businesses to cover these costs. Otherwise, he argued, such large-scale capex cannot be sustained.

Portfolio Exposure and Market Context

Altimeter held 9.41 million Nvidia shares worth $1.88 billion at June 30, representing about 19% of its reported U.S. equity portfolio. Its filing also listed $303.5 million of SpaceX shares. Altimeter also holds stakes in Anthropic and OpenAI.

Gerstner described the boom as the largest super cycle in technology history, noting that semiconductors have generated 70% of the Nasdaq's return this year. On Polymarket, traders assign Nvidia a 74% chance of finishing 2026 as the world's largest company, with roughly $7.4 million in trading volume.

Nvidia shares were trading around $220 on Friday morning after gaining 2.5% on Thursday.

What the Numbers Show

The divergence between Anthropic's reported $65 billion run rate and the $75 billion investor expectation highlights the sensitivity of valuations to specific data points. Gerstner's requirement for an 80% increase in combined lab revenue within months underscores the aggressive growth trajectory needed to validate the $1.5 trillion annual capex cycle described by hyperscalers.