Bolt Graphics unveils Zeus GPU with dev kits set for 2026

Bolt Graphics disrupts the GPU market with Zeus—a robust processor aiming to transform gaming, high-performance computing, and professional visualization with next-level memory and networking.

Bolt Graphics has entered the competitive GPU market with its new Zeus GPU, targeting a broad spectrum of demanding applications including path tracing in gaming, computer-aided design workloads, high-performance computing simulations, and media production for film and television. The Zeus platform sets itself apart with an innovative memory architecture, offering configurations of 32, 64, or 128 GB soldered memory and expandable capacity via two or four DDR5 SO-DIMM slots for a total of up to 384 GB. Additionally, the GPU integrates a high-speed 400 GbE QSFP-DD port, enabling direct networking without a network interface card. This configuration is tailored for large-scale render and simulation farms where thousands of these GPUs can interconnect efficiently, laying groundwork for next-generation computational clusters.

On the performance front, Bolt Graphics positions Zeus as a direct competitor to industry leaders like the NVIDIA GeForce RTX 5090, particularly in gaming scenarios. Traditional challenges in real-time ray tracing and path tracing—where compromises between visuals and performance have been the norm—are addressed by Zeus´s ability to deliver full-quality, 4K path-traced graphics at 120 frames per second, with a minimum of 25 samples per pixel. This claims to eliminate the need for developers to trade speed for fidelity. For scientific and engineering computations, Zeus delivers up to 20 FP64 TeraFLOPS at under 400 watts, positioning it as a potent tool for research and industrial simulations. The GPU demonstrates a dramatic leap in specific workloads, handling electromagnetic wave simulations up to 300 times faster than previous solutions while maintaining full precision.

Bolt Graphics plans to release developer kits for Zeus in 2026, followed by full-scale production in 2027. The anticipation surrounding Zeus is high, largely due to its unconventional approach to expandability and interconnectivity. As real-world performance evaluations await the arrival of developer kits, the industry watches closely to see if Bolt´s bold claims hold true on both the gaming and high-performance computing fronts. The arrival of Zeus could signal a new phase in GPU advancement for creators, scientists, and engineers alike as the boundaries of memory, networking, and computational throughput are pushed further.

72

Impact Score

NVIDIA maps broader RTX Spark rollout

NVIDIA says it has secured enough supply of the 3 nm N1X silicon from TSMC to support RTX Spark growth in PCs. The company is positioning the platform for mainstream and performance systems while already planning follow-on generations.

NVIDIA NemoClaw targets autonomous engineering workflows

NVIDIA is positioning NemoClaw as a secure blueprint for long-running autonomous engineering agents across design, simulation and manufacturing. Major software vendors and startups are using it to automate workflows that have traditionally required extensive manual coordination.

NVIDIA and Microsoft expand stack for agentic Artificial Intelligence

NVIDIA and Microsoft are extending their partnership across Windows devices, Azure, local deployments and data infrastructure to support agentic Artificial Intelligence workloads. The effort combines new hardware, model availability, faster data services and secure runtimes for autonomous agents.

Jensen Huang set to lead Computex in Taipei

Nvidia chief Jensen Huang is poised to dominate Computex with a major speech centered on Artificial Intelligence chips, software and systems. The appearance is expected to highlight Taiwan’s strategic importance to Nvidia’s plans and the broader Artificial Intelligence supply chain.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.