TLDRs;
- Nvidia’s next-generation Kyber AI server platform has reportedly been delayed until 2028 due to manufacturing complications.
- The delay could also affect the rollout of the larger NVL576 AI system built around Kyber architecture.
- Nvidia reportedly abandoned an alternative hardware design after receiving negative feedback from major cloud customers.
- The setback may provide AMD and Google with an opportunity to strengthen their positions in the AI infrastructure market.
Nvidia (NASDAQ: NVDA) stock is drawing renewed attention after reports suggested the company’s highly anticipated Kyber AI server architecture has been pushed back to 2028, raising fresh concerns about the pace of its next-generation artificial intelligence infrastructure roadmap.
According to research firm SemiAnalysis, the delay stems from manufacturing issues involving a critical printed circuit board (PCB), forcing Nvidia to postpone the launch of the rack-scale system originally expected to accompany its Rubin Ultra GPU platform in 2027.
The reported setback comes as Nvidia continues to dominate the AI chip industry, where demand for increasingly powerful computing systems has surged alongside the rapid adoption of generative AI technologies.
Manufacturing Challenges Surface
Kyber represents one of Nvidia’s most ambitious AI infrastructure projects to date. Rather than focusing on a single graphics processor, the platform is designed as a complete rack-scale computing solution capable of integrating 144 GPUs into one server cabinet.
The architecture is intended to deliver enormous computational power for hyperscale cloud providers, AI laboratories, and enterprise customers training increasingly sophisticated machine learning models.
However, SemiAnalysis reported that technical difficulties related to a key printed circuit board have disrupted production timelines, ultimately delaying commercial availability until 2028.
The delay reportedly extends beyond the base Kyber rack itself. Nvidia’s larger NVL576 platform, which connects eight Kyber racks through high-speed optical networking to create an even more powerful AI computing cluster, could also experience delays or initially ship in only limited quantities.
Such challenges illustrate the growing engineering complexity involved in building next-generation AI systems that combine hundreds of processors into tightly integrated computing environments.
Customer Feedback Changed Plans
The report also revealed that Nvidia had explored an alternative backup approach intended to keep its rollout schedule on track.Rather than waiting for Kyber‘s new architecture, the company reportedly considered creating a solution by combining two existing-generation AI racks into a larger configuration.
Nvidia's next-gen AI rack system delayed to 2028 on manufacturing snags, SemiAnalysis says https://t.co/aBcxNrzSsv
— CNBC (@CNBC) July 6, 2026
That contingency plan was ultimately abandoned after major cloud service providers and hyperscale customers expressed concerns over its operational complexity and deployment requirements.
Large cloud operators prioritize efficiency, power consumption, cooling requirements, and ease of maintenance when deploying AI infrastructure across thousands of servers. According to SemiAnalysis, customer objections convinced Nvidia to discontinue the fallback strategy in favor of refining the original Kyber platform.
The decision highlights the close collaboration between Nvidia and its largest customers, many of whom influence product development through large-scale purchasing commitments.
Competitive Pressure May Increase
Although Nvidia remains the clear leader in AI accelerators, any delay to its future product roadmap may offer competitors an opportunity to narrow the gap.
SemiAnalysis suggested the postponement could create openings for companies including Advanced Micro Devices (AMD) and Google, both of which continue investing heavily in AI hardware designed for hyperscale computing environments.
As AI infrastructure spending continues accelerating worldwide, cloud providers increasingly seek multiple hardware suppliers to reduce dependency on a single vendor and diversify their computing capacity.Even so, Nvidia continues to benefit from an extensive software ecosystem, including CUDA, networking technologies, and AI development tools that remain deeply integrated into enterprise workflows.
4th of July Flash Sale – 50% OFF!
Celebrate Independence Day by investing in your future. For a limited time, get 50% OFF a Knockout Stocks membership and unlock our latest high-conviction stock picks, powered by our proprietary KO Score algorithm.
You'll also get access to our long-term investment ideas and shorter-term trade opportunities, helping you identify potential opportunities before the crowd.
Sign up to Knockout Stocks today and get 50% OFF to unlock the full list of premium stock picks.
Use coupon code SPECIAL50 for your exclusive discount.
Offer ends soon. Don't miss out!
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
