This press release was originally published in the PaletteAI website on January 5th and can be found here.

PaletteAI makes full-stack AI easy to integrate, fast to deploy, and efficient to manage—supporting optimal resource utilization and ongoing innovation across the AI ecosystem.

As AI factories move from pilots to production, reference architectures have become the preferred path to accelerate time to value and reduce risk.

NVIDIA Enterprise AI Factory validated design brings together best practices for AI infrastructure and software, so IT teams can operationalize AI quickly and predictably. In turn AI and data science teams can benefit from self-service access to standardized accelerated environments for a faster, more consistent workflow, from experimentation and training to deploying inferencing workloads at scale across on-premises and edge.  

Spectro Cloud’s PaletteAI and PaletteAI Secure are full-stack AI platforms that are now included in the NVIDIA Enterprise AI Factory validated design, and support the end-to-end provisioning, lifecycle management, policy enforcement, and ongoing automation required to adopt AI factories at scale  from the model to the metal.

With this support,  PaletteAI enables organizations to deploy and operate AI factories built on NVIDIA accelerated computing, high-performance networking, and NVIDIA AI Enterprise software  alongside a rich ecosystem of partner solutions,  all aligned with NVIDIA’s validated design guidance.

The result? Platform teams can deliver standardized, full-stack AI environments with the speed, repeatability, and operational control needed to support their enterprise AI practitioner teams’ critical initiatives.

Building and Operating Enterprise AI Factories with PaletteAI

Below are some of the key technologies and capabilities that PaletteAI supports and deploys:

• NVIDIA Blackwell accelerated computing and
• Integration with the NVIDIA AI Enterprise software platform, withNVIDIA NeMo providing the tools for custom model training and refinement and NVIDIA NIM microservices delivering optimized, secure inference endpoints for running AI models at scale
• NVIDIA BlueField acceleration, cybersecurity, and lifecycle automation using the NVIDIA DOCA software framework.
• Streamlined deployment of NVIDIA Spectrum-X Ethernet networking and GPU-accelerated infrastructure from the edge to the data center.

For organizations with FIPS 140-3 and sovereignty requirements, PaletteAI Secure delivers a hardened, compliant version of PaletteAI for deploying AI factories across air-gapped data centers, classified networks, and sovereign clouds.

PaletteAI includes PaletteAI Studio - a library of deployable reference architectures based around the NVIDIA Enterprise AI Factory validated design, and supports component choices such as NVIDIABlueField and Spectrum-X, NVIDIA DOCA,  NVIDIA AI Data Platform, NVIDIA AI Enterprise , NVIDIA NIM microservices, NVIDIA Run:ai,  and a number of the most popular 3rd party and open source projects.  PaletteAI Studio blueprint profiles  have been tested on the most common system partnerservers and the PaletteAI Studio content is  continually updated to provide easy access to the latest versions and security fixes available.  

PaletteAI Studio turns the NVIDIA Enterprise AI Factory validated design into deployable, reusable profiles that reduce the need for deep Kubernetes and infrastructure knowledge. This allows customers to stand up AI factories faster and more secure, and concentrate on delivering AI workloads at scale, delivering value and increasing ROI.

Spectro Cloud’s Solution Support provides customers with a single support entry point for their Enterprise AI Factory, with collaborative triage across all of the vendors in the PaletteAI Studio stack - including NVIDIA AI Enterprise. This gives customers a single, accountable partner, ensuring clear ownership, faster resolution, and a smoother support experience from discovery to fix.Learn more about our work with NVIDIA to make AI factories real.

Spectro Cloud is an NVIDIA Preferred Partner and has been helping organizations in every industry deploy AI workloads at scale — our support for AI factories is a natural extension of our ongoing work to streamline production deployment of NVIDIA accelerated computing and NVIDIA AI Enterprise software across telco, enterprise and edge environments.