#United States #San Francisco #AMD #TensorWave #Zyphra

Zyphra and AMD Announce Zyphra Cloud

SAN FRANCISCO, May 4, 2026 — Zyphra has unveiled its new AI platform, Zyphra Cloud, designed to streamline the integration of advanced AI technologies for developers and enterprises. This state-of-the-art platform is powered by AMD's Instinct™ MI355X GPUs and utilizes TensorWave's high-performance infrastructure. Zyphra Cloud represents a significant leap forward in model serving, agent infrastructure, and computing power, combining all these elements into a single cohesive environment for deploying advanced AI solutions.

The Core Features of Zyphra Cloud

Zyphra Cloud has debuted with a particular focus on serverless inference capabilities through its flagship Zyphra Inference service. This service grants access to cutting-edge open-weight models, including DeepSeek V3.2, Kimi K2.6, and GLM 5.1. By incorporating custom kernels and advanced algorithms tailored for long-context inference, the platform is engineered to deliver rapid performance critical for real-world applications like agentic coding and extensive workflow automation.

Krithik Puthalath, the Founder and CEO of Zyphra, remarked, “Zyphra Cloud is the natural extension of our research. We've spent years perfecting our AI systems on AMD infrastructure, and now we're excited to provide a comprehensive platform for developers and enterprises.” He further emphasized that this platform presents an opportunity for teams to effectively develop and implement advanced AI systems that meet the stringent demands of real-world applications.

Collaboration with AMD and TensorWave

AMD has played a crucial role in the success of Zyphra Cloud. Negin Oliver, AMD's Corporate Vice President for Business Development in AI, noted that the combination of optimized software and AMD’s cutting-edge architecture enables unparalleled inference performance for demanding AI models. TensorWave, with its focus on providing high-performance cloud solutions, is the backbone of this partnership. Jeff Tatarchuk, Co-Founder and Chief Growth Officer at TensorWave, expressed that outfitting Zyphra Inference with their MI355X infrastructure aligns with their goal to help AI-native firms deliver production-ready solutions.

Future Expansions

Looking ahead, Zyphra Cloud aims to extend beyond inference services to encompass a broader range of integrated capabilities. Planned future enhancements include distributed post-training services, such as reinforcement learning and fine-tuning, as well as sandboxed development environments supported by AMD EPYC™ CPUs. This expansion will provide users with a comprehensive suite of tools for creating, training, and executing AI systems in a unified environment.

Availability

Zyphra Cloud is available for users today, with further information accessible at Zyphra Cloud or directly at cloud.zyphra.com.

About Zyphra

Based in San Francisco, Zyphra is a pioneering research and product company dedicated to developing human-aligned AI systems that empower individuals and organizations to reach their full potential. For more details, visit www.zyphra.com.

About TensorWave

TensorWave is a cloud provider specializing in AI and HPC solutions powered exclusively by AMD Instinct™ Series GPUs, focused on delivering performance-driven infrastructure optimized for demanding AI workloads. For more information, go to tensorwave.com.

About AMD

AMD (NASDAQ: AMD) is a global leader in semiconductors, delivering products and services innovative enough to solve some of the world’s most pressing challenges. Discover more at www.amd.com.