A semiconductor analyst warns that custom chips developed by hyperscalers pose a significant risk to the dominance of NVIDIA [1, 2].
This shift represents a potential disruption of the AI chip market, as the world's largest cloud providers—known as hyperscalers—begin to develop their own same-purpose hardware to reduce dependency on a single supplier.
Jay Goldberg of Seaport Research said that there is the potential for this trend to entirely disrupt NVIDIA [3]. Goldberg's warnings were part of a segment on the Marketplace Morning Report [4].
Market volatility has already begun to reflect these concerns. NVIDIA shares dropped 2.23% [5] on Friday, coinciding with with broader market shifts in the AI sector.
While NVIDIA has maintained a tight grip on the AI hardware market, the move toward custom silicon is a strategic pivot by cloud providers. By developing their own chips, these companies can optimize performance for their own specific workloads, and potentially lower long-term costs.
Goldberg's assessment highlights a tension between NVIDIA's current revenue streams and the long-term sustainability of its market share. If hyperscalers transition away from general-purpose GPUs, the demand for NVIDIA's flagship products may decrease over time.
“Custom chips developed by hyperscalers ”
The transition from general-purpose AI accelerators to custom silicon (ASICs) represents a structural shift in the semiconductor industry. While NVIDIA currently leads in the ecosystem, the trend of vertical integration by cloud providers suggests that the largest customers are now attempting to mitigate the same risk of over-reliance on a single vendor, which fundamentally alters the long-term demand forecast for the AI hardware market.
