Broadcom's $230 billion AI chip outlook outpaces Nvidia's 70% growth

Both companies are benefiting from the same AI hardware boom, but their latest quarters show the different shapes of their businesses. Nvidia's quarter ended July 26, 2026, brought total revenue of $96.2 billion, up 106% from a year earlier, with $89 billion from data center. Broadcom's quarter ended Aug. 2 produced $29.6 billion in revenue, up 86% year over year, of which $16.7 billion came from AI semiconductors — more than half of the total and up 221% from a year earlier.

On Nvidia's earnings call, chief financial officer Colette Kress divided its data center business into hyperscalers, which brought in $49 billion and grew 13% from the prior quarter, and the category Nvidia calls "ACIE," which includes industrial and regular enterprises building their own models and reached $40 billion, up 25%. That split illustrates the breadth of Nvidia's customer base. Broadcom's outlook, by contrast, rests on a smaller number of massive customers building custom chips at scale, a model that can deliver exceptional profits but carries greater risk if a major buyer pulls back or changes direction.

The narrow question of which company has the faster AI-revenue outlook from here points to Broadcom: its forecast implies one doubling followed by another, a steeper trajectory than Nvidia's roughly 70% growth. Still, the headline growth rate does not tell the whole story. Whether Broadcom's concentrated bet proves as durable as Nvidia's broader base may depend on whether those few large custom-chip customers keep expanding their orders, while Nvidia's ability to ship enough hardware to meet demand will shape how its own outlook plays out.