Fermyon Breaks New Ground in Serverless Hyperscaling with WebAssembly on Google Kubernetes Engine

Fermyon Platform for Kubernetes is the only solution capable of scaling in lock step with GKE’s new dynamic autoscaling virtual machine technology

/EIN News/ -- LONGMONT, Colo., April 09, 2025 (GLOBE NEWSWIRE) -- Fermyon™ Technologies, Inc., the serverless WebAssembly company, today demonstrated a breakthrough for serverless hyperscaling as Fermyon Platform for Kubernetes reaches unparalleled scale on Google Kubernetes Engine (GKE) clusters, achieving 7x faster pod scheduling and execution speeds 50x faster than today’s leading serverless cloud technologies.

At Containers @ Google NEXT ‘25, the Google Cloud team debuted the first high-scale demonstration of next generation GKE Autopilot capabilities. Google Cloud has just announced new performance improvements to GKE Autopilot, including faster pod scheduling, scaling reaction time, and capacity right-sizing. The on-site demo features a GKE Autopilot cluster that hosts 1,000 typical lightweight serverless functions on a single four core node, capable of handling unanticipated bursts of traffic to any function doubling the amount of traffic you can serve within seconds without over-provisioning or introducing latency. The unprecedented workload density is made possible by Fermyon’s specialized high-density runtime, Fermyon Platform for Kubernetes.

This further cements WebAssembly’s leading-edge stature as the highest-performing compute standard and Fermyon’s leadership within the WebAssembly (Wasm) ecosystem for its selection by Google Cloud to demonstrate this level of scale. The combination of WebAssembly’s fast application scaling and Autopilot’s resource scaling ushers in a new era of cloud elasticity capable of supporting highly responsive “bursty” applications delivered faster than the blink of an eye (<100 milliseconds) while simplifying cluster operations and reducing cost for consumers.

Media, ecommerce, financial services, and other high-volume digital experience companies know that even 100-millisecond delays can impact customer engagement and online revenue — especially under unpredictable crushing-load situations such as a viral moment or an unexpected high response to a product launch. Current Kubernetes architectures compound the problem by forcing companies to pre-scale to massive compute levels and leave those computers more than 83% idle which puts pressure on both corporate wallets and global emissions.

"The synergy between Google Kubernetes Engine (GKE) and Fermyon's WebAssembly platform offers remarkable application and infrastructure elasticity. Fermyon's use of WebAssembly enables serverless applications to achieve startup times in the sub-millisecond range, facilitating rapid scaling in response to sudden demand spikes," stated Paul Nashawaty, Practice Lead and Principal Analyst at theCUBE Research. "GKE enhances this by providing automated, rapid scaling of virtual machine resources through features like Autopilot mode, which manages node provisioning and scaling based on traffic. This dual-layer elasticity ensures that applications can handle unexpected surges, such as viral events, with near-instantaneous responsiveness, highlighting the growing momentum for serverless WebAssembly and GKE adoption."

Scaling with Fermyon Serverless

Fermyon Platform for Kubernetes (self-hosted, single-tenant) and Fermyon Wasm Functions (hosted, multi-tenant) are both WebAssembly serverless execution engines which achieve blazingly fast cold-starts (<1 millisecond), incredibly high workload density (>50x more than typical Kubernetes app density), and instant scale. Both execute Spin serverless functions. Spin is a Fermyon-contributed CNCF Sandbox project focused on delivering superior developer experience for writing event-driven serverless functions, enabling a developer to go from blinking cursor to deployed serverless function in two minutes or less.

Pairing this high performance with Google Cloud’s new next-generation Autopilot, featuring a new container-optimized compute platform alongside performance improvements like faster pod scheduling and scaling reaction times, means nodes in a Kubernetes cluster can dynamically resize in a few seconds, capable of hosting thousands of Spin serverless functions and hundreds of thousands of requests per second. This avoids the painful preparatory work cloud operations teams have to do to imagine how large viral spikes could be, then plan for and put in place — at an expense — enough capacity to handle those loads.

Starting today, Google Cloud customers can develop Spin serverless functions and deploy them on GKE with the container-optimized compute platform enabling them to:

• Avoid painful blockages, blackouts, and shortages during high-volume bursts that cause customers to be locked out and walk away, thus retaining eyeballs and revenue.
• Enjoy a serverless developer experience on Kubernetes that rids developers of painful overhead to prepare their code for Kubernetes deployments.
• Write a serverless function in almost any language with SDK support for Javascript/TypeScript, Rust, Python, .NET/C#, and Go.
• Cut cold start time of HTTP-centric applications to a mere fraction of a millisecond.
• Avoid over-spending on Kubernetes clusters for future, unknown events.
• Avoid the stress that comes with under-guessing.

Google Cloud customers will be able to directly procure Fermyon Platform for Kubernetes and/or Fermyon Wasm Functions from Fermyon on next-generation GKE Autopilot. And starting in Q3, Autopilot’s container-optimized compute platform will also be available to standard GKE clusters, without requiring a specific cluster configuration.

“Bringing the world’s fastest serverless platform to the world’s first on-demand expandable compute, backed by intelligent capacity provisioning that supports fast pod scheduling, is a big win for any organization interested in avoiding those painful viral-moment blackouts,” said Fermyon CEO Matt Butcher. “With cold starts under half a millisecond, robust service APIs, AI and GPU integration and support for a broad array of programming languages, Fermyon brings Google customers next-generation compute that is not just industry leading, but redefining.”

“Companies worldwide are constantly seeking superior performance with less risk and cost. We’ve responded to that with the next generation GKE Autopilot offering featuring a new container-optimized compute platform, along with performance improvements like fast pod scheduling and capacity right sizing, powered by unique Google Cloud hardware. This gets rid of the uncertainty, doubt and cost that comes with pre-planning Kubernetes deployments. Fermyon simply can help customers reach new heights when it comes to scale and speed, utilizing WebAssembly on GKE,” said Gabe Monroy, VP/GM of Cloud Runtimes at Google.

Google Cloud and Fermyon will be showcasing the Fermyon Platform for Kubernetes on GKE at Google Cloud Next in Las Vegas from April 9th to 11th.

Additional Resources

• Learn about Fermyon Platform for Kubernetes
• Learn about Fermyon Wasm Functions

About Fermyon™ Technologies, Inc.

Fermyon is leading the next wave of cloud computing with the first cloud-native WebAssembly FaaS that lets developers build better serverless apps faster. Fermyon is focused on empowering cloud developers to quickly realize the things they are thinking about creating and focus on the code that brings value instead of the obligatory foundation code. Fermyon was founded by the Deis Labs team at Microsoft Azure and is backed by Insight Partners and Amplify Partners. For more information, go to https://www.fermyon.com or follow @fermyontech.

Media Contact
constantia@fermyon.com