singletenant.ai

Vendors · Comparison

Azure OpenAI Provisioned Throughput (PTU) vs Databricks Mosaic AI Model Serving (now Databricks Model Serving)

Only Azure OpenAI PTU documents C5. Pricing, compliance, regions and residency compared, every fact sourced and dated.

Last verified (older of the two profiles) How we verify

From the data

Key differences

  • Pricing Azure OpenAI Provisioned Throughput (PTU): Per PTU per hour; Azure Reservations discount with 1-month or 1-year terms (published pricing). Databricks Model Serving: Per DBU per hour by GPU instance size (e.g. A10G 20 DBU/hour); per-token Foundation Model APIs (published pricing).
  • Category Azure OpenAI Provisioned Throughput (PTU) is a Hyperscaler dedicated offering; Databricks Model Serving is Managed self-host.
  • Compliance Both document SOC 2 Type II, ISO 27001, FedRAMP, HIPAA BAA and GDPR DPA. Only Azure OpenAI Provisioned Throughput (PTU) documents C5.
  • Data residency Azure OpenAI Provisioned Throughput (PTU) commits to data residency in its contract terms; Databricks Model Serving states data residency on its own pages (not contractual).
  • Jurisdiction Both parent companies are under United States jurisdiction.

Generated from the verified vendor data below; sources and dates on each vendor profile.

Azure OpenAI Provisioned Throughput (PTU)

Provisioned throughput is now a Microsoft Foundry (formerly Azure AI Foundry) deployment type billed hourly per PTU, with 1-month or 1-year Azure Reservations. Global, Data Zone (US, EU or APAC), and Regional options set where prompts are processed; stored data stays in the customer-designated geography. FedRAMP scope names Azure OpenAI; SOC 2, ISO 27001, and C5 attest the Azure platform.

Databricks Mosaic AI Model Serving (now Databricks Model Serving)

Model Serving deploys custom MLflow models and fine-tuned foundation models on Databricks-managed serverless compute; provisioned throughput allocates dedicated inference capacity. Trust pages list ISO 27001:2022 (including AWS single-tenant), SOC 2 Type II, HIPAA options and FedRAMP Class C and D (formerly Moderate and High). A standard BAA and a DPA with SCCs are published.

Side by side

Capabilities compared

Azure OpenAI Provisioned Throughput (PTU) Databricks Mosaic AI Model Serving (now Databricks Model Serving)
Pricing model Per PTU per hour; Azure Reservations discount with 1-month or 1-year terms Per DBU per hour by GPU instance size (e.g. A10G 20 DBU/hour); per-token Foundation Model APIs
Public pricing Yes Yes
Category Hyperscaler dedicated Managed self-host
Deployment models Global provisionedData-zone provisionedRegional provisioned Custom model servingProvisioned throughputPay-per-token endpoints
Regions Global, United States, EU/EEA, Asia-Pacific United States, United Kingdom, EU/EEA, Asia-Pacific
Compliance SOC 2 Type IIISO 27001C5FedRAMPHIPAA BAAGDPR DPA SOC 2 Type IIISO 27001HIPAA BAAFedRAMPGDPR DPA
Data residency evidence Contractual Stated
Parent jurisdiction United States United States
Sovereign for Not recorded Not recorded
Analyst note Provisioned throughput is now a Microsoft Foundry (formerly Azure AI Foundry) deployment type billed hourly per PTU, with 1-month or 1-year Azure Reservations. Global, Data Zone (US, EU or APAC), and Regional options set where prompts are processed; stored data stays in the customer-designated geography. FedRAMP scope names Azure OpenAI; SOC 2, ISO 27001, and C5 attest the Azure platform. Model Serving deploys custom MLflow models and fine-tuned foundation models on Databricks-managed serverless compute; provisioned throughput allocates dedicated inference capacity. Trust pages list ISO 27001:2022 (including AWS single-tenant), SOC 2 Type II, HIPAA options and FedRAMP Class C and D (formerly Moderate and High). A standard BAA and a DPA with SCCs are published.

Where they diverge

Deployment differentiation

Only Azure OpenAI Provisioned Throughput (PTU)

Global provisionedData-zone provisionedRegional provisioned

Both

No overlap in deployment models.

Only Databricks Model Serving

Custom model servingProvisioned throughputPay-per-token endpoints