Did Nvidia's 2028 Rack Delay Under Jensen Huang Open a Door for AMD and Google?

One of the biggest evolutions of Nvidia's (NASDAQ: NVDA) business model is that it has gone from just selling graphics processing units (GPUs) to designing full server rack systems to handle specific AI tasks.

This has been a key part of CEO Jensen Huang's strategy; however, it could have hit a snag when technology research outfit SemiAnalysis reported that its Kyber NVL144 next-gen AI rack system could be delayed until 2028 due to troubles with manufacturing a crucial circuit board.

Nvidia has come out and said its roadmap remains intact, although questions linger.

Image source: The Motley Fool.

Nvidia has been able to largely keep competitors out of the high end of the market through an aggressive technology roadmap, but physical manufacturing limits could finally be catching up to it.

Kyber is designed to be at the heart of AI data centers to help push AI models forward, and it also has different variants for specific AI tasks like inference and agentic AI. A delay could help open the door for Advanced Micro Devices (NASDAQ: AMD) and Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG) to gain a foothold in the high end of the AI market.

AMD's biggest challenge is that there has always been a large software ecosystem gap between it and Nvidia. However, AMD has greatly improved its ROCm platform over the past few years, and the move of programmers working higher up the software stack and using open-source AI frameworks like OpenAI's Triton have helped close the gap, especially for inference.

At the same time, AMD's chiplet designs, which can package more memory onto its chips, and its recent acquisition of memory optimization software platform Mext, position it well to offer a high-end server solution designed specifically for inference.

As for Alphabet, its Tensor Processing Units (TPUs) have become highly regarded AI chips, and with its next generation, it will have chips optimized for both training and inference. A delay in Kyber, meanwhile, could make its cost-efficient TPU offering look more attractive to customers that want a completely optimized system without the risk of delay or the cost of Nvidia's premium platform.

Both AMD and Alphabet stocks look attractive

Neither AMD nor Alphabet has to displace Nvidia as the king of AI infrastructure to be winners, as just getting a piece of the high-end market should be a big boost. AMD has already captured some big GPU deals, and it also looks set to ride the wave with its data center central processing units (CPUs), which become more important with agentic AI. Alphabet, meanwhile, is the most complete AI play, with both world-class chips and AI models, giving it a cost edge.

Taking some additional share from Nvidia would just be a bonus that undoubtedly would lift their stocks. I own both AI stocks and think their futures look bright.

Geoffrey Seiler has positions in Advanced Micro Devices and Alphabet. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, and Nvidia. The Motley Fool has a disclosure policy.