NVIDIA's CEO Jensen Huang has categorically denied market speculation regarding a potential delay to the company's next-generation AI accelerator system, Vera Rubin. Speaking to journalists at an event in Tokyo, Japan, Huang stated that Vera Rubin hardware is already in production and is on schedule for customer delivery, progressing towards "massive" production volumes.
Recently, concerns have surfaced in the market about a possible postponement of NVIDIA's upcoming product. Reports suggested that the Kyber NVL144 rack system, intended for the Rubin Ultra GPU and originally slated for a 2027 shipment, might be pushed back to 2028 due to production issues. Additionally, some analysts have noted that the Rubin platform could face a "minor delay" related to thermal challenges, though they believe the issue has been resolved and should not pose a significant obstacle.
Huang's remarks align with NVIDIA's previously reiterated stance that its "roadmap is intact." According to earlier reports, NVIDIA has begun delivering Vera Rubin platform samples to select customers, with plans to formally commence mass production and shipments in the second half of 2026. The market widely expects the platform to begin shipping to major AI clients starting in July.
While addressing the delay rumors, Huang also painted a picture of the company's robust growth outlook. During an investor conference last week, he revealed that quarterly revenue is approaching the $100 billion milestone, with growth rates still accelerating. Analysts suggest that the core challenge for NVIDIA is no longer the existence of AI demand, but rather how to convert its massive order backlog into actual revenue amid multiple constraints such as power supply, networking, memory, and data center space.
Comments