Photos recently shared on the Chinese tech forum Chiphell show what looks like an empty PCB from Nvidia’s upcoming GeForce RTX 5090 graphics card. This layout gives us a sneak peek into the specs and features of this powerful next-gen GPU.
Technical Specifications
The PCB images display an unsoldered board featuring a large BGA package designed for the GB202 GPU, the core component expected to power both the RTX 5090 and potentially higher-tier variants. While lacking Nvidia branding, markings on the PCB’s bottom right suggest this is a PNY model based on Nvidia’s PG145 board design.
The memory configuration shows 16 solder pads surrounding the main GPU core, supporting 32GB of GDDR7 memory. The possibility exists for memory configurations up to 48GB or 64GB, depending on the modules chosen. GDDR7 technology brings improvements in data transfer speeds and power efficiency compared to current GDDR6X memory.
Power delivery comes through a single 12V-2×6 connector rated for 600W, replacing the previous 12VHPWR design that caused issues with RTX 40-series cards. Third-party manufacturers may opt for dual connectors in their overclocked models.
Advanced Features and Launch Timeline
The RTX 5090 comes with DisplayPort 2.1a, offering impressive display options like 8K at 240Hz or 16K at 60Hz with DSC. This is a big step up from the DP1.4a standard that’s been around since 2016.
PCIe 5.0 support debuts in this consumer GPU series, doubling external bandwidth to 64GT/s. While gaming benefits may be minimal, this advancement particularly benefits AI workloads and multi-GPU configurations.
Industry sources indicate the RTX 5090 will debut during CES 2025, launching after the RTX 5080. Additional models including the RTX 5070 Ti and RTX 5070 will likely join the lineup. The enhanced specifications position this GPU for high-end gaming, content creation, and AI applications.
Read this also: Apple Partners With Sony to Introduce PlayStation VR Controllers for Vision Pro
