Companies / Modular / Subprocessors
Modular provides an AI inference platform that runs open and custom AI models across GPUs, CPUs, and other accelerators, offered as a managed cloud service or deployed inside a customer's own VPC. It serves developers and teams that need to run AI models in production, with options such as shared and dedicated endpoints and per-usage or per-minute pricing.
Summarized from its homepage on 10 Nov 2026.
Subprocessors of Modular
Modular declares 8 subprocessors.
| Subprocessor | Purpose |
|---|---|
| Cloud Infrastructure | |
| Cirrascale | GPU-specific resources |
| DNS, WAF etc | |
| Full-Stack Observability for Cloud-Scale applications and Infrastructure. | |
| Cloud-based source control software | |
| TPU-specific resources | |
| HydraHost | GPU-specific resources |
| TensorWave | GPU-specific resources |
Recent changes
| Date | Change |
|---|---|
| 17 Nov 2026 | removed BentoML |
| 17 Nov 2026 | added TensorWave |
| 17 Nov 2026 | added Google Cloud Platform |
| 17 Nov 2026 | added Cirrascale |
| 17 Nov 2026 | no longer shows Google LLC |
Vendor's own statement, retrieved from the source URL on the date shown. StrataDB does not verify the underlying claim. Data sources.
See Modular's full stack
Every vendor with its source and date, and an alert when Modular adds or drops one. The same lookup works for any company.
- Every company and vendor, with its public source
- 100 API calls a month, plus MCP and the Chrome extension
- Work email only. No card, no sales call