NVIDIA's GB300 Blackwell Ultra AI Servers Set to Revolutionize AI Computing in Late 2025

NVIDIA's new GB300 Blackwell Ultra AI servers to begin shipping in second half of 2025

As an Amazon Associate, we earn from qualifying purchases. TweakTown may also earn commissions from other affiliate partners at no extra cost to you. NVIDIA's new GB300 "Blackwell Ultra" AI servers are expected to begin shipping in the second half of 2025, with the new GB300 AI servers to be the most powerful AI servers on the planet. Taiwanese media outlet UDN reports that Taiwanese contract manufacturing giant Foxconn has secured the biggest number of orders for NVIDIA's next-gen GB300 AI servers. Foxconn isn't the only one, with contract manufacturers including Quanta, Wistron, Wiwynn, and Inventec all itching to pump out GB300 AI servers. According to supply chain sources, NVIDIA's new GB300 AI server manufacturing will have precedence over Apple's upcoming iPhone, as Foxconn manufactures the iPhone for Apple, and is leading the charge with GB300 AI servers. The new NVIDIA GB300 has similarities with GB200, with Foxconn expecting to find a few difficulties when shifting to GB300 AI server mass production. Foxconn is also expecting NVIDIA GB300 AI servers to make up over half of its server revenue moving forward, too. Quanta Computer started shipping NVIDIA GB200 AI servers in Q2 2025, and is currently testing and verifying its new GB300 AI servers with its customers, with GB300 AI servers made by Quanta expected to begin shipping in September, with Quanta management believing that the more powerful GB300 AI servers will see high-end AI companies waiting for them because of pressing model training needs. We recently spotted NVIDIA's new GB300 NVL72 AI servers at Computex 2025 earlier this year, looking mighty fine in the flesh indeed.

NVIDIA's Latest GB300 AI Servers Will Start Shipping In H2, Say Supply Chain Sources

This is not investment advice. The author has no position in any of the stocks mentioned. Wccftech.com has a disclosure and ethics policy. NVIDIA's Blackwell GB300 AI servers are expected to start shipping in the second half of this year, with analysts expecting them to take precedence over Apple's upcoming iPhone. The GB300 AI servers are expected to be the most powerful AI servers in the world, and Taiwanese contract manufacturers are rushing to win over their orders. Among the contract manufacturers expected to work with NVIDIA are Foxconn, Quanta, Wiwynn and Inventec, which are all expected to ship in the second half. According to the details, Taiwanese contract manufacturing giant Foxconn has managed to secure the largest number of orders for NVIDIA's next-generation GB300 AI servers. The GB300 will feature 72 Blackwell AI GPUs for its highest-end variant to make it the most powerful AI platform in the world. Along with Foxconn, contract manufacturers Quanta, Wistron, Wiwynn and Inventec are also eager to ship the new servers. The rush to deliver GB300 servers due to their integral nature in the global AI supply chain has also given precedence to the servers over Apple's upcoming iPhone, believe supply chain sources. Foxconn continues to lead NVIDIA's server assembly supply chain. Due to the GB300's similarities with its predecessor, the GB200, Foxconn expects to encounter a few difficulties in shifting to GB300 mass production. Additionally, the booming demand for AI servers is impacting Foxconn's income statement, and AI servers are expected to account for more than half of the firm's server revenue. Contract manufacturer Quanta Computer started to ship GB200 servers in the second quarter and it is currently testing and verifying the GB300 servers by working with its customers. The GB300 servers from Quanta are expected to ship in September, and Quanta'a management believes that despite the greater power punch offered by the GB300 products, high-end AI companies are unlikely to wait for them due to pressing model training needs. Like Quanta's servers, products from Winwynn and Inventec are also expected to ship by September. Shipment volumes are expected to grow significantly in the second half of the year, and the firm's management believes that it could gain market share with the latest NVIDIA servers. Supply constraints and high prices have created large market demand for alternatives to NVIDIA's GPUs. These alternatives include in-house chips developed by big tech players such as Amazon and Alphabet and orders made to designers such as Broadcom and Marvell. NVIDIA CEO Jensen Huang has dismissed speculation that alternatives or custom chips could impact the firm's market as he believes NVIDIA's products offer superior performance. However, other reports have suggested that OpenAI is allegedly shifting some of its AI computing needs to Google's tensor processing units (TPUs) as it is wary of the high costs of using NVIDIA's products.