Nvidia's stock fell by 0.3% to $202.14 in premarket trading on Friday following the announcement that Meta Platforms is set to begin production of its proprietary AI chip, named "Iris," starting in September. This development is part of Meta's broader strategy to reduce its reliance on external suppliers such as Nvidia and AMD while aiming to lower computing costs.
The "Iris" chip, developed under the Meta Training and Inference Accelerator (MTIA) program with support from Broadcom, is expected to be manufactured by Taiwan Semiconductor Manufacturing Company (TSMC). Meta's internal communications revealed that the initial bug testing phase took only six weeks without revealing major complications, indicating a significantly improved timeline for a project that has faced challenges since its inception over five years ago.
Meta's strategy involves releasing new chip iterations approximately every six months until 2027, which is more frequent than the typical annual cycle seen in the industry. However, it is important to note that the "Iris" chip is not meant to entirely replace Nvidia's GPUs; instead, it is designed to complement the substantial GPU volumes that Meta currently procures. The company acknowledged in its memo that integrating the latest GPUs at its scale presents difficulties, which have resulted in delays.
Market analysts, including Benchmark Research's Cody Acree, have expressed that the emergence of Meta's chip may not necessarily spell doom for Nvidia. Despite a potential dip in Nvidia's relative market share within Meta's expanding infrastructure, total spending by hyperscalers is predicted to more than double, which could sustain Nvidia's actual revenue growth from these customers. Furthermore, Meta intends to invest up to $145 billion in AI infrastructure this year, a notable portion of a larger tech industry investment estimated to exceed $700 billion overall.
This announcement from Meta aligns with a broader trend where giants like Google and Amazon are also gearing up to introduce their AI chip offerings Google's Tensor Processing Units and Amazon's Trainium processors aimed at external clients. These chips may introduce additional competition for Nvidia, particularly in specific workloads where cost-efficient performance is prioritized. As a result, Nvidia's stock has been underperforming relative to other semiconductor companies recently, such as AMD, which rose by 5.67% on the day.
This situation reflects Nvidia's lingering vulnerability amid ongoing competition as futures associated with the Nasdaq 100 index also dipped by 0.2%, suggesting a larger market trend, though the news surrounding Nvidia specifically has contributed additional pressure.
This article is informational and does not constitute financial advice.
