Alphabet Rivals Nvidia in AI Chips, Adds New Variable to Capex Thesis

What happened

A Motley Fool report indicates Alphabet is ramping up efforts to compete with Nvidia in the AI accelerator market, potentially reducing its reliance on the chip giant. The news arrives as Alphabet executes a massive $180–190B FY2026 capex plan, much of which is directed at AI infrastructure including custom TPUs. While in-house chips could improve margin structure and supply assurance over time, they also introduce new execution and design risk, and represent a longer-term strategic shift that does not address near-term capacity constraints or the regulatory overhang from the EU Digital Markets Act (DMA) decision due July 27, 2026. The base case in our master report assumes Alphabet’s infrastructure spend converts into deployable capacity, but custom chip development adds another dimension of uncertainty to that conversion timeline and success probability. This announcement, while directionally positive for Alphabet's vertical integration narrative, does not alter the fundamental near-term dependency on Nvidia-supplied GPUs and the physical constraints (power, land, chips) that currently cap cloud growth.

Implication

If Alphabet successfully develops competitive AI accelerators, it could reduce long-term cost dependency on Nvidia and strengthen its cloud AI margin structure. However, this is a multi-year effort with high execution risk. The current valuation (P/E 30) already prices in significant AI upside; investors should await evidence of capacity relief and regulatory clarity before adding exposure.

Thesis delta

The thesis shift is minor but noteworthy: Alphabet's move to compete with Nvidia in AI accelerators adds a potential long-term margin tailwind and reduces supply-chain risk, but it also adds execution complexity to an already ambitious capex program. The near-term thesis remains driven by cloud capacity conversion and EU regulatory outcomes. This development increases the range of possible outcomes modestly, but does not change our base-case WAIT rating or the key catalysts we track.