Vectris Says Waveform Raises AI Inference Throughput on Existing GPUs

Aug 20, 2026
Vectris says its Waveform control plane recovered unused AI inference capacity inside deployed GPUs, with tests showing 30 to 73% higher throughput and lower energy use. The product is scheduled to launch October 1, 2026, with limited design partner access.

Vectris announced in a press release that it has built Waveform, a control plane for recovering unused AI inference capacity inside deployed GPUs. The company said tests across NVIDIA H100, H200, and B200 infrastructure showed 30 to 73% higher throughput, 51 to 56% lower energy use, and 22 to 42% faster workload completion.

Waveform is designed to increase productive AI output without retraining models, changing model weights, or modifying GPU kernels. Vectris describes its metric as Compute Yield, meaning the amount of quality equivalent accepted AI output produced from existing compute infrastructure.

The company said the system is based on deterministic structural patterns it found inside AI inference workloads. Vectris said Waveform has also been tested across silicon from NVIDIA, AMD, and Intel.

Waveform is scheduled to launch on October 1, 2026, with initial access limited to design partners.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.