Sep 16, 2026, 11:20 PM ETConference Presentation, Other

Brad Gerstner: No AI Bubble, Semis Eat the Nasdaq & AI's Take Off Problem

The market has risen 15% year-to-date and 39% since January 2024, driven primarily by the AI infrastructure supercycle despite concerns regarding tariffs, geopolitics, and regulation.
Unlike the 2000 dot-com bubble, current valuations reflect earnings growth rather than multiple expansion; earnings are up 26% while P/E multiples on the NASDAQ, S&P 500, and SOX have contracted.
Semiconductors account for 70% of the Nasdaq's returns, with hardware makers like NVIDIA (trading at 14x forward earnings) and hyperscalers driving the gains.
Major indices including the MAG7, NASDAQ, S&P, and SOX are trading at or below their historical average multiples.
Consumer discretionary, software, and financial sectors have lagged, indicating a market narrowly driven by CapEx build-outs rather than broad-based participation.

Anthropic's monthly revenue trajectory served as the catalyst for the April/May rally, rising from $2 billion in January to $11 billion in March.
Analysts project the combined run rate for the top three labs (Anthropic, OpenAI, and SpaceX) to reach $100 billion by July, requiring growth to $180 billion by year-end to sustain the current trade.
The "rental" model dictates that hyperscalers (Microsoft, Google, Amazon) are investing in infrastructure to rent out capacity, meaning massive "of-take" revenue from labs is required to validate trillions in capital expenditure.
A gap analysis suggests the market needs annual run-rate revenue to grow from the current ~$200 billion to $450 billion, then $800 billion, and eventually $1 trillion to sustain projected CapEx levels.
Cumulative compute growth forecasts show a massive scale increase, with 19 gigawatts of compute expected to be added in 2026, half of which is anticipated for the two leading labs.
Analyst Dylan Patel forecasts 43 gigawatts of compute capacity additions for 2026, a figure the speaker views as aggressive.
The speaker estimates actual feasible capacity additions closer to 25 gigawatts, citing permitting delays, grid interconnection issues, and skilled labor shortages.
Despite capacity constraints, the speaker believes 5 gigawatts of additional compute could generate another $100 billion in revenue for top labs, rendering the 43-gigawatt forecast unnecessary for revenue targets.

Regulation: The speaker warns against activist-driven "over-regulation" that could stifle innovation, citing the historical precedent of shutting down 67 US nuclear reactors due to safety concerns.
Power Infrastructure: Building the required 43 gigawatts faces significant hurdles including local opposition, grid interconnection delays, and sold-out power equipment.
Interest Rates: There is a >90% probability of interest rate hikes, which increases the hurdle rate for the borrowed capital funding new data centers.
Equity Risk: Elevated risk-free rates (e.g., 5.5% on the 10-year Treasury) create a competitive barrier for equity investments compared to fixed income.
Oil Prices: Rising oil prices are correlated with higher interest rates, adding another layer of pressure to the equity market valuation.

Revenue Sensitivity: The trajectory for the rest of the year hinges on monthly AI lab revenues; reaching an $8 billion monthly run rate could trigger takeoff and an IPO in 2026.
Portfolio Positioning: The speaker maintains a "medium" position, emphasizing mental flexibility over aggressive leverage ("don't YOLO").
Conditional Adjustments:
- Bull Case: If AI revenues exceed expectations and oil prices retreat, the portfolio will increase exposure.
- Bear Case: If revenues disappoint or macro conditions worsen, the portfolio will reduce exposure further.
Market Shift: The investment thesis has evolved from a broad "AI supercycle" bet (2023-2025) to a specific, fact-driven assessment of revenue execution and infrastructure feasibility in 2026.
Productivity Dividend: Companies are projected to achieve margin expansion (moving from 38 bps to 100 bps) by replacing headcount growth with AI efficiency rather than outright mass layoffs.
Consumer Agents: The speaker anticipates a new trillion-dollar category emerging from consumer AI agents capable of booking travel and managing tasks.

Brad Gerstner is leading a push to establish an investment account for every American child (70 million under 18) to promote capitalism as an antidote to socialism.
Gerstner promotes "CAC scans" (cardiac CT) as a high-ROI preventative health measure, estimating it could save 50,000 lives annually if adopted as widely as mammograms.
These initiatives are described as "moonshots" but are secondary to the day's primary market analysis.