NVDA, MSFT, GOOGL Hit Record Highs as AI Memory Demand Extends Capex Cycle

The confluence of CEO commentary and record gains in mega-cap technology stocks underscore a pivotal market narrative: the AI capex cycle is not peaking but rather accelerating into 2026. In earnings calls over the past week, the five largest technology companies all noted the same bottleneck: memory constraints are binding. This is not a transient supply hiccup but a structural mismatch between booming inference demand and limited capacity.

Cisco's latest earnings also validated this thesis beyond the hyperscaler cohort. The company reported stronger-than-expected AI infrastructure orders, with demand extending into switches, optics, and networking layers. NVDA CEO Jensen Huang's appearance in Beijing alongside President Trump and Secretary of Commerce Scott Bessent signals continued geopolitical negotiation around chip export approvals. The US government approved H200 chip sales to 10 Chinese companies, a move interpreted as coordinating commercial interests with diplomatic outreach.

The market repricing is swift. NVDA has added roughly $1 trillion in market cap over the past week. Call options on Mag 7 names saw over $249 million in bullish single-leg premiums bought, with NVDA, TSLA, and AAPL accounting for 46% of all call buying. GOOGL has added $1.5 trillion in market cap over the past six weeks alone. But the question haunting traders is not whether AI demand is real; it is whether the price now embedded in these valuations assumes a maintenance capex phase or a step change in sustained infrastructure investment.

Sceptics point to the fact that memory stocks like MU are pricing in a slower-growth scenario despite CEO warnings from the big five. If capex remains elevated and memory utilisation stays tight, MU's 7x earnings multiple looks cheap. Conversely, if hyperscalers reach saturation sooner than current consensus assumes, the breadth of the rally outside the core AI beneficiaries will compress, leaving the Mag 7 vulnerable to rotation.