Mag 7 Stocks Rally on AI Infrastructure Demand; $249M+ Call Buying Signals Conviction

The artificial intelligence trade continues to accelerate across the Mag 7 as institutional conviction shows up in both equity positioning and options markets. In a span of just two days last month, the CEOs of Microsoft, Meta, Google, Amazon, and Apple all delivered the same signal on their earnings calls: memory is constrained and the constraint is not ending soon. That bottleneck is now feeding a bull case for sustained capex cycles that extends well beyond model training into inference capacity and networking infrastructure. The market is repricing memory and chipmakers accordingly, with call buyers particularly aggressive on NVDA, TSLA, and AAPL, which accounted for roughly 46% of $249M+ in bullish call premium purchased today alone.

Cisco's recent earnings provided another data point supporting this thesis. The company signaled that AI networking demand is not just confined to data center training clusters but is widening into switches, optics, and distributed networking gear. This broadening buildout suggests the capex cycle has room to accelerate beyond NVIDIA's GPUs and into adjacent semiconductor and network-infrastructure vendors. Meanwhile, NVDA itself continues to push toward new highs on reports that the U.S. government approved H200 chip sales to ten Chinese companies, a signal that even amid trade tensions, the demand for advanced semiconductors remains strong enough to clear regulatory hurdles.

The cross-asset implications are significant. Broadening AI capex favors semiconductor sub-sectors (AMD, AVGO) that have historically lagged pure-play GPU names. It also supports the case for infrastructure-heavy utilities and data-center real-estate names, though the primary near-term beneficiary remains Big Tech itself. The options market is pricing little room for disappointment; any miss on forward capex guidanceCompany-issued forecasts of future financial performance. or memory-constraint timeline could trigger sharp pullbacks, particularly in names that have already re-rated aggressively.

Sceptics point out that markets are now pricing in a multi-year capex supercycle with little margin for error. If the memory bottleneck resolves faster than expected, or if cloud providers pull back on spending due to margin pressure, the thesis unravels quickly. Additionally, the focus on capex alone ignores the critical question of monetization: when and how much of those AI infrastructure investments translate into earnings growth remain uncertain.