At KubeCon North America 2026, Spectro Cloud and AMD today announced the AMD Instinct™ Coder 355, a turnkey, local-first AI coding appliance powered by Dell PowerEdge XE9785 servers and using AMD Instinct™ MI355X GPUs. Delivered by Spectro Cloud through Dell Technologies OEM Solutions, AMD Instinct™ Coder 355 expands the AMD Instinct™ Coder portfolio with a higher-performance option for larger teams.

Since its launch in August 2026, AMD Instinct™ Coder 325 has demonstrated strong demand for running AI coding inference on infrastructure organizations own and control. AMD Instinct™ Coder 355 builds on that momentum. Its AMD Instinct™ MI355X GPUs can serve larger open-weight models, longer context windows, and more concurrent developers and agents.

The AMD Instinct™ Coder 355 appliance combines:

• Spectro Cloud PaletteAI Inference Launchpad: the software layer that includes local-first inference, prepackaged and pre-optimized open-weight models, bring-your-own-model, intelligent model routing, metering, quotas, KV cache optimization, on-prem and hosted deployment options, and governance.
• AMD Instinct™ MI355X GPUs: 288 GB of HBM3E memory and up to 8 TB/s of memory bandwidth per GPU for large-model AI inference.
• Dell PowerEdge XE9785 servers: enterprise-grade, production-ready servers engineered to handle high-density AI machine learning and HPC workloads.

Choose for the right business outcome

AMD Instinct™ Coder 355 gives customers more choice in how they deploy local-first AI coding. One pre-validated appliance supports approximately 50 concurrent developers and delivers strong cost efficiencies for larger teams. This product line expansion enables organizations to choose based on team size, performance needs, and the business outcomes they are striving to accomplish.

Organizations that want to deploy and centrally manage a larger estate of appliances consistently at scale can extend the solution with the Spectro Cloud PaletteAI control plane, available with an additional PaletteAI software subscription.

Local-first, with model choice and security built in

Today’s open-weight models can tackle the majority of AI coding tasks. The AMD Instinct™ Coder 355 appliance supports a range of leading open-weight models, including regional and sovereign options, so organizations can choose models that meet their policy and compliance requirements. Policy-based routing sends each request to the right model based on workload, capability and cost, and tasks that call for a frontier model can still be routed to one. According to Gartner®, “Without a governed engineering operating model, costs can escalate faster than the productivity gains these tools are designed to deliver.”¹ Built-in metering, quotas and multi-tenant controls let platform teams scale AI coding with confidence.

“AMD Instinct™ Coder 325 proved that enterprises want a local-first path to AI coding, and today’s open-weight models can handle most of that work,” said Tenry Fu, Co-founder and CEO of Spectro Cloud. “With AMD Instinct™ Coder 355, customers can choose the appliance that fits their specific business requirements, on the same platform and developer experience. Powered by Dell PowerEdge servers and AMD Instinct MI355X GPUs, it brings that choice to larger teams on infrastructure enterprises already trust.”

“With AMD Instinct™ Coder 355, we are bringing AMD Instinct™ MI355X GPUs into the Instinct Coder portfolio,” said Kumaran Siva, Corporate Vice President, Enterprise AI at AMD. “This gives enterprises access to a wider choice of hardware that supports more intensive use, with greater concurrency for larger developer teams, all on the open AMD ROCm™ software ecosystem.”

Availability

AMD Instinct™ Coder 355 can be ordered through Spectro Cloud and its reseller and distribution partners. Attendees at KubeCon can see live demos at the Spectro Cloud booth #125.