SANTA CLARA, CALIFORNIA / RankWire.AI / – Nvidia is planning to implement price increases exceeding 15% for numerous AI server models scheduled for delivery in early 2027. These adjustments pertain to configurations built on the company’s Vera Rubin and Grace Blackwell platforms. The extent of each price hike varies depending on the chip generation and memory configuration. Nvidia has not officially announced a blanket price raise across all server lines, but manufacturers have already conveyed the updated pricing details to leading data center clients.
Major server manufacturers have shared these new pricing figures with key data center operators, including Microsoft, Google, and Oracle. These companies purchase large quantities of AI infrastructure to support their cloud and data center services. The primary driver behind the reported increases in server prices is rising memory costs, which constitute a significant component of overall expenses. AI workloads demand large volumes of high-bandwidth and server memory, alongside GPUs, CPUs, networking hardware, and storage devices. During 2026, the high demand for AI deployments has kept several memory segments in tight supply.
According to TrendForce, the typical contract prices for conventional DRAM are expected to increase by 13% to 18% in the third quarter of 2026. Additionally, NAND Flash contract prices are forecasted to grow by 10% to 15% within the same period. The research firm highlighted that server DRAM remains in short supply because suppliers are prioritizing capacity for AI-related applications. Although some long-term supply agreements have helped cap certain price hikes, overall server memory costs remain elevated due to ongoing demand pressures.
Memory cost increases as AI server utilization stays robust
In May, Nvidia announced that Vera Rubin had entered full production alongside server manufacturers and supply chain partners worldwide. The company indicated that products based on Rubin would become available during the latter half of 2026. The Vera Rubin platform integrates the Vera CPU, Rubin GPU, NVLink 6, ConnectX-9, BlueField-4, and Spectrum-6 networking solutions. It succeeds Grace Blackwell as Nvidia’s latest rack-scale architecture optimized for large-scale AI workloads. System developers are creating Rubin-based products for cloud providers, AI research labs, and hyperscale data centers.
Meanwhile, Grace Blackwell continues to be a significant element of Nvidia’s current data center offerings. The GB200 NVL72 design links 36 Grace CPUs with 72 Blackwell GPUs within a single liquid-cooled rack. Nvidia engineered this system to function as a unified NVLink-enabled computing domain tailored for extensive AI models. The reported price increases for servers differ depending on configuration, with memory capacity and chip generation playing key roles in pricing. Therefore, the 15% figure does not imply a uniform increase across all Nvidia server models.
Expansion of Vera Rubin production impacts Nvidia’s entire server lineup
Suppliers of memory have shifted more production capacity toward server and high-performance products to meet AI-driven demand. TrendForce noted that this shift has reduced the supply of PC and consumer DRAM products. Despite robust server shipment volumes through 2026, buyers have been accumulating data center memory inventories. The market analyst expects server DRAM supply to remain constrained into 2027 as demand growth continues to surpass new supply, creating the cost environment influencing Nvidia’s announced price adjustments.
Nvidia enters this pricing phase with record-breaking data center revenue. In its fiscal first quarter ending April 26, the company reported a total revenue of $81.6 billion. Revenue from the Data Center segment hit an all-time high of $75.2 billion, representing a 92% increase compared to the previous year. Nvidia projected second-quarter revenue of approximately $91 billion, with a margin of error of plus or minus 2%, when releasing those results in May. The company plans to publish its fiscal second-quarter earnings on August 26.
}<|endofjson|>