Move Over, Magnificent 7, Now There's the AI 11

The discussion centered on the frothy valuations and cyclical risks in the AI infrastructure basket dubbed the "AI 11," with contributors noting the group generated negative aggregate free cash flow in 2023 before improving to $123B in 2025. In athletic apparel, On Holding was viewed as the strongest operator, with gross margin up more than 4 percentage points and gross margins approaching 65%, while Under Armour was described as a restructuring story and Adidas as solid but less compelling. The mailbag highlighted rising competitive and regulatory pressure on DraftKings and Flutter from prediction markets, which account for about 85% of bets on those platforms.

Analysis

The cleanest signal here is not “AI is strong” but that the profit pool is shifting upstream and becoming more cyclical than the market wants to admit. The infrastructure winners with real pricing power are the ones that can monetize scarcity in the near term; the rest are effectively levered to capex sentiment, which can unwind fast if hyperscaler budgets normalize. That makes the group’s current cohesion deceptive: when the cycle turns, semis, memory, and equipment won’t de-rate evenly — the highest-beta memory names and the most execution-sensitive smaller beneficiaries should fall first.

The athletic apparel setup is more interesting as a pricing-power test than a consumer-demand story. The brands that can pass through tariff and freight pressure without losing velocity will take share from the promotional players, and that share shift tends to persist for 4-6 quarters because wholesale partners preserve the strongest doors for the highest-throughput brands. That argues for avoiding laggards that need discounting to clear inventory; once the channel learns to expect promotions, margin repair becomes structurally harder.

On gambling, the market is likely underestimating how fragile the economics are if prediction markets remain legally viable. The key second-order effect is customer acquisition cost inflation: if a more tax-efficient alternative can offer better economics to bettors, the incumbent sportsbook model loses the ability to monetize scale, and the first casualty is margin rather than handle. If the regulatory pendulum swings back, these stocks can rebound violently, but until then they look like short-duration options on litigation, not durable compounders.

Key Decisions for Investors

• Long ONON / short NKE as a 3-6 month relative-value trade: ONON has better pricing power and margin expansion potential; NKE remains trapped in a promotional recovery that can lag even if demand stabilizes.
• Use pullbacks to add AMD and AVGO only on weakness, not strength, over the next 1-2 quarters: both are high-quality AI beneficiaries, but at current valuations the asymmetry is poor if AI capex growth slows.
• Short UAA on any relief rally for a 1-2 quarter trade: the restructuring path looks longer than consensus, and tariff/discounting pressure makes operating leverage fragile.
• Avoid initiating new long DKNG/FLUT positions until regulatory clarity improves; if forced, prefer a small options-defined call spread over common stock to express a reversal scenario without underwriting downside.
• Pair long ASML / short WDC or SNDK for the next 6-12 months: equipment exposure should be more resilient than memory if AI spend stays firm but de-rates from peak intensity.