How More Pieces Are Falling Into Place For Google's Nvidia AI Chip Challenge

AMD's market cap surpassed $1 trillion for the first time, driven by a 10% surge in shares to $615.52. The company's data center revenue doubled year-over-year to $6.7 billion, with CEO Lisa Su calling it an 'excellent quarter.' AMD is the fourth U.S. semiconductor firm to reach this milestone, following NVIDIA, Broadcom, and Micron.

Alibaba's stock rose 4% after unveiling the Zhenwu V900 AI chip, claiming 3x performance gains and clusters of up to 500,000 units. The chip targets the gap left by NVIDIA due to export controls. Baidu's stock remained flat, suggesting the market views this as an Alibaba-specific win. Mass production is scheduled for Q1 2027, with all claims vendor-reported.

Apple's new Mac Minis and Mac Studios, priced up to $20,000, aim to compete with Nvidia and Microsoft in AI tasks. Apple claims its devices handle AI locally, reducing cloud costs. Apple has 4.6% of the enterprise desktop market vs. Windows' 91.3%. Microsoft and Nvidia are also targeting AI desktop markets. Apple's hardware chief highlights cost savings and performance.

Nvidia (NVDA) and AMD (AMD) rely on ASML Holding (ASML) for advanced EUV lithography machines, crucial for manufacturing AI chips. ASML expects 25% sales growth in 2023 and 75% by 2026, driven by demand for AI and memory chips. ASML's monopoly in EUV lithography positions it to benefit from increasing AI infrastructure spending, with EPS growth expected to exceed 30% annually.

NVIDIA launched PAIR, a beta software that turns home PCs into an AI cluster. It supports Windows, macOS, and Linux, and works with GeForce RTX 20-series GPUs. In a demo, PAIR reduced task time from 18 minutes to 8 minutes and 48 seconds across three devices. Ollama and LM Studio are supported at launch, and data remains local. The source code is available on GitHub.

Nvidia's Vera Rubin NVL72, the successor to Blackwell, has entered full production. The NVL72 rack features 72 Rubin GPUs and 36 Vera CPUs, offering significant performance improvements. Nvidia reported that the platform is being manufactured at scale by top server makers and is already deployed in cloud data centers. The company expects Vera Rubin-based systems to be available from partners in the second half of 2026.