Spectro Cloud

Spectro Cloud's Palette is an enterprise Kubernetes management platform, and PaletteAI extends it into GPU and AI-factory operations. It gives platform, IT operations and AI infrastructure teams one place to design, deploy, patch and govern full-stack clusters across cloud, data center, bare metal and edge, replacing hand-assembled AI stacks with versioned, repeatable blueprints.

Spectro Cloud is a San Jose, California company founded in 2019. Its Palette platform manages the full Kubernetes lifecycle, from the operating system and Kubernetes distribution through networking, storage, add-ons and applications. It covers public clouds, VMware vSphere and Nutanix data centers, bare metal and edge devices, and runs VM and container workloads side by side. The core building block is the Cluster Profile, a versioned blueprint for everything in a cluster. The architecture is decentralized: management intelligence runs inside each cluster, so policy is enforced locally and upgrades roll out in parallel. The AI side arrived with PaletteAI, which became generally available on March 16, 2026. It launched with an ecosystem that builds in NVIDIA AI Enterprise and adds integrations with Supermicro, Canonical, WEKA, F5, ClearML and others. PaletteAI finds GPU hardware and deploys complete AI stacks from a curated catalog called PaletteAI Studio. Each stack is packaged as a Profile Bundle: infrastructure plus applications such as Run:ai and ClearML. It also offers GPU-as-a-Service and Model-as-a-Service to AI practitioners, within tenant quotas and role-based access controls. On July 21, 2026 the company added PaletteAI Inference Launchpad. It routes requests between local and frontier models, meters token usage and enforces governance and quota policies on NVIDIA or AMD GPUs. The vendor claims up to 70% lower token costs, depending on workload mix. Palette VerteX and PaletteAI VerteX are built for air-gapped, heavily regulated and sovereign-AI deployments, with FIPS 140-3 validated cryptography. The company closed a $75M Series C led by Growth Equity at Goldman Sachs Alternatives in November 2024. It also contributed Kairos, an operating system for edge Kubernetes that is now a CNCF Sandbox project, to open source. Customers it names include GE HealthCare, T-Mobile, Airbus and Yum! Brands.

Category

Infrastructure & Cloud

Pricing

Usage-based, Subscription, Contact for pricing

Target Market

CIOs, CTOs, Platform Engineering Teams, IT Operations, AI Infrastructure Teams

Deployment

Multi-cloud, Hybrid, Self-hosted

Founded

2019

Headquarters

San Jose, California, USA

• ✓
  Cluster Profiles

  Versioned blueprints that cover the OS, Kubernetes distribution, networking, storage, add-ons and applications, so every cluster is built and upgraded the same way.

• ✓
  PaletteAI Studio and Profile Bundles

  A curated catalog of validated AI stacks. Each bundle packages infrastructure plus applications such as Run:ai and ClearML so teams can redeploy it the same way every time instead of assembling it by hand.

• ✓
  GPU discovery and GPU-as-a-Service

  Finds GPUs during setup, pools and allocates them under per-tenant quotas, and lets AI practitioners serve themselves within limits the platform team sets.

• ✓
  PaletteAI Inference Launchpad

  Locally managed inference that routes requests between local and frontier models, meters token usage, and enforces governance and quota policies to control spend.

• ✓
  Decentralized management architecture

  Management intelligence runs inside each cluster, so policy is enforced locally and upgrades roll out in parallel across large fleets.

• ✓
  VerteX editions for regulated environments

  Palette VerteX and PaletteAI VerteX use FIPS 140-3 validated cryptography and are built for air-gapped, heavily regulated and sovereign-AI deployments.

• ✓
  Kairos open-source edge OS

  Spectro Cloud contributed Kairos to open source. It is an Apache-2.0, CNCF Sandbox immutable Linux distribution for edge Kubernetes with atomic A/B upgrades.

• •
  Standing up an enterprise AI factory

  Take GPU servers from bare metal to production AI stacks with NVIDIA AI Enterprise built in, using validated partner integrations instead of custom glue code.

• •
  Cutting inference token spend

  Send suitable requests to locally hosted models through Inference Launchpad, metering tokens and enforcing quotas to reduce reliance on paid frontier model APIs.

• •
  Running large edge fleets

  Run one standardized Kubernetes stack across thousands of sites; the vendor says Yum! Brands runs one edge stack across more than 40,000 restaurants.

• •
  Patching fleets at scale

  Patch large cluster fleets quickly and consistently; the vendor says GE HealthCare patched all 100 of its clusters in about four hours.

• •
  Air-gapped and sovereign AI

  Run AI and Kubernetes infrastructure in disconnected government or regulated environments with FIPS 140-3 validated cryptography, using the VerteX editions.

• •
  Consolidating VMs and containers

  Run virtual machines and containers side by side on one Kubernetes-managed platform; the vendor also markets a PaletteAI VM Launchpad for migration.

Palette is metered in kilocore-hours (worker CPU cores multiplied by time) for cloud and virtualized environments, and per node for edge and bare metal. PaletteAI costs a flat fee per GPU under management. The vendor site does not publish list prices. The AWS Marketplace listing shows 1-month tiers from $4,000 (100-499 kCh) to $200,000 (5,000+ kCh), $40 per kCh with no commitment, and up to 50% off on 12- to 36-month contracts. The vendor advertises free access to its SaaS platform but does not publish the limits.

This page was written from 11 sources, 5 on domains other than spectrocloud.com.

1. 1.spectrocloud.com — spectrocloud.comvendor
2. 2.spectrocloud.com — paletteaivendor
3. 3.spectrocloud.com — editionsvendor
4. 4.spectrocloud.com — spectro cloud announces expanded paletteai ecosystemvendor
5. 5.spectrocloud.com — spectro cloud launches paletteai inference launchpadvendor
6. 6.spectrocloud.com — free tiervendor
7. 7.docs.spectrocloud.com — saml sso
8. 8.am.gs.com — growth equity spectro cloud
9. 9.aws.amazon.com — B08J8383Q5
10. 10.hn.algolia.com — search
11. 11.github.com — kairos