Foundry Technologies
AI NativeFoundry is now Mithril — the AI omnicloud
Foundry is now Mithril — the AI omnicloud
Published Oct 15, 2026
| Starting Price | Varies by GPU type and configuration; vendor publishes spot and reserved instance-hour pricing that differs by GPU and instance size | Varies by configuration; vendor published instance-hour pricing for H100 and other GPUs | Varies by configuration; on-demand GPU pricing is published for multiple instance types |
| Best For | Teams needing multi-provider flexibility, short reservations, and market-based pricing | Enterprise-scale GPU workloads requiring a single-vendor GPU cloud with a defined instance catalog | Teams seeking transparent on-demand pricing and self-serve GPU instances and cluster products |
| Key Strength | Aggregation and abstraction across multiple providers; published market-based pricing; short commitments | Broad GPU catalog and enterprise focus; established infrastructure footprint | Transparent pricing tables; broad availability of instance sizes and cluster options |
| Key Weakness | Limited public reviews and less-documented enterprise features (fine-grained RBAC/SSO) | Higher pricing in some configurations; enterprise-sales motion may be required | Not positioned as a multi-provider orchestrator; availability and queueing can vary |
Where Mithril is differentiated:
Where competitors may have advantages:
Adoption trend context (market): The GPU neocloud market is expanding, with notable activity from established GPU cloud providers; Mithril’s partnership model suggests a strategy of leveraging third-party supply rather than owning all infrastructure.