---
title: "Baseten | Ethical Stack"
url: https://ethicalstack.ai/providers/baseten/
catalog_version: f6baff2119d0e9158520da7fbe4232011dac79f25bd08757436d7a1dbd593db2
---

# Baseten

A managed platform for deploying and serving customer-selected machine-learning models through hosted inference APIs.

Canonical profile: https://ethicalstack.ai/providers/baseten/

Scope: Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Published overall: 4.6; 5/6 dimensions rated.

Round the equal-weight mean of available numerical ratings to one decimal; require at least four of six for an overall score. BDS status stays outside the mean; for eligible profiles, a boycott target caps the final score at 2/10 and a pressure target subtracts one point (minimum 1). Not listed and not checked do not adjust the score.

## Energy practices

Finding: 2/10

Assessment: https://ethicalstack.ai/providers/baseten/#energy

Baseten has identifiable product documentation, but no measured energy footprint or clean-electricity sourcing evidence for the assessed route.

Reason: 1 + 1 reporting + 0 electricity + 0 controls = 2/10. This grades documented practices and transparency.

Argument: The official sources describe how Baseten is delivered and used. They provide no dated, quantitative energy or water report for this product and no evidence tying its requests to clean electricity. The pages reviewed also name no energy-saving workload or compute control.

Type: argument; confidence: medium

Scope (dimension): Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Per-request electricity and water use, regional power mix and independently verified clean-electricity sourcing for Baseten remain unpublished.

```json
{
  "method": "practices-transparency-v1",
  "scope": "Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.",
  "components": {
    "disclosure": {
      "score": 1,
      "reason": "The Baseten documentation identifies the service route, but provides no dated, quantitative energy or water report for this product.",
      "sourceUrls": [
        "https://www.baseten.co/",
        "https://www.baseten.co/security-practices/",
        "https://www.baseten.co/legal-hub/"
      ]
    },
    "electricity": {
      "score": 0,
      "reason": "No reviewed official source ties this workload to metered, regional or hourly clean electricity; this is an evidence gap.",
      "sourceUrls": [
        "https://www.baseten.co/",
        "https://www.baseten.co/security-practices/",
        "https://www.baseten.co/legal-hub/"
      ]
    },
    "demand": {
      "score": 0,
      "reason": "The reviewed Baseten documentation names no energy-saving compute or workload setting for this route; no physical savings are inferred from speed, pricing or general configuration choices.",
      "sourceUrls": [
        "https://www.baseten.co/",
        "https://www.baseten.co/security-practices/",
        "https://www.baseten.co/legal-hub/"
      ]
    }
  },
  "actualImpactMeasured": false
}
```

- [Baseten inference platform](https://www.baseten.co/) (primary; checked 2026-10-28)

- [Baseten Security Practices (updated September 2026)](https://www.baseten.co/security-practices/) (primary; checked 2026-10-28)

- [Baseten legal hub](https://www.baseten.co/legal-hub/) (primary; checked 2026-10-28)

## Training data

Finding: Depends on the model

Assessment: https://ethicalstack.ai/providers/baseten/#training-rights

Training-data rights vary with the model selected for Baseten; check that model’s source and license.

Reason: Select the exact model and provider used by this feature to assess the training-data permissions for that release.

Argument: Baseten can use different model routes. Its product policy describes how requests are handled, while training-data provenance and permissions depend on the selected model. Review that release’s source disclosure, license and creator controls separately.

Type: argument; confidence: medium

Scope (profile): Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: The selected model name, version and provider can vary by feature or setting. Those details are needed for a model-specific training-data review.

```json
{
  "method": "training-rights-evidence-v1",
  "scope": "The stated training release or selected model",
  "components": null,
  "routeDependency": "Select the exact model and provider used by this feature to assess the training-data permissions for that release.",
  "modelReference": null
}
```

- [Baseten inference platform](https://www.baseten.co/) (primary; checked 2026-10-28)

- [Baseten Security Practices (updated September 2026)](https://www.baseten.co/security-practices/) (primary; checked 2026-10-28)

- [Baseten legal hub](https://www.baseten.co/legal-hub/) (primary; checked 2026-10-28)

## Your data & privacy

Finding: 8/10

Assessment: https://ethicalstack.ai/providers/baseten/#privacy

Baseten promises zero persistent retention for Model APIs and Dedicated Inference; asynchronous inputs can remain queued for up to 72 hours.

Reason: Baseten’s September 2026 security terms say it does not use Customer Content to train models and applies zero-retention handling to Model APIs and Dedicated Inference. Asynchronous inputs may be queued for up to 72 hours, and customer-selected logging, runtimes or third-party products can create other data paths.

Argument: Baseten’s September 2026 security terms say it does not use Customer Content to train models and applies zero-retention handling to Model APIs and Dedicated Inference. Asynchronous inputs may be queued for up to 72 hours, and customer-selected logging, runtimes or third-party products can create other data paths.

Type: argument; confidence: medium

Scope (profile): Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Confirm the product, customer configuration, backups and log settings in the signed order and technical documentation; this service-level commitment does not control software selected by the customer.

- [Baseten inference platform](https://www.baseten.co/) (primary; checked 2026-10-28)

- [Baseten Security Practices (updated September 2026)](https://www.baseten.co/security-practices/) (primary; checked 2026-10-28)

- [Baseten legal hub](https://www.baseten.co/legal-hub/) (primary; checked 2026-10-28)

## Openness

Finding: 4/10

Assessment: https://ethicalstack.ai/providers/baseten/#openness

Baseten documents deployment options, security controls and customer VPC deployment, but its managed serving platform is not offered under an open-source license; customer model weights remain separately owned.

Reason: Baseten documents deployment options, security controls and customer VPC deployment, but its managed serving platform is not offered under an open-source license; customer model weights remain separately owned.

Argument: Baseten documents deployment options, security controls and customer VPC deployment, but its managed serving platform is not offered under an open-source license; customer model weights remain separately owned.

Type: argument; confidence: medium

Scope (profile): Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: The cited product pages do not provide a complete source release, reproducible model pipeline and license package for the full assessed service. Individual model licenses and artifacts can differ.

- [Baseten inference platform](https://www.baseten.co/) (primary; checked 2026-10-28)

- [Baseten Security Practices (updated September 2026)](https://www.baseten.co/security-practices/) (primary; checked 2026-10-28)

- [Baseten legal hub](https://www.baseten.co/legal-hub/) (primary; checked 2026-10-28)

## Worker protections

Finding: 1/10

Assessment: https://ethicalstack.ai/providers/baseten/#labour

Public sources for Baseten describe the product, but do not report pay, worker voice or independent safeguards for the people who build and support it.

Reason: 1 + 0 + 0 + 0 = 1/10. Documented worker protections; missing disclosure is an evidence gap.

Argument: The official pages reviewed for Baseten do not report worker pay and conditions, representation or independent follow-through across employees, contractors and AI supply-chain workers. The low evidence score records what those sources leave unanswered; it is not a finding of mistreatment.

Type: argument; confidence: medium

Scope (dimension): Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Pay, contractor protections, grievance outcomes and independent checks for Baseten's wider workforce remain unreported in the cited product documentation.

```json
{
  "method": "worker-safeguards-evidence-v1",
  "entity": "Baseten",
  "company": "Baseten",
  "scope": "Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.",
  "components": {
    "conditions": {
      "score": 0,
      "reason": "The reviewed product and legal pages give no qualifying pay or conditions disclosure for AI supply-chain workers.",
      "sourceUrls": [
        "https://www.baseten.co/",
        "https://www.baseten.co/security-practices/",
        "https://www.baseten.co/legal-hub/"
      ]
    },
    "voice": {
      "score": 0,
      "reason": "No worker representation, grievance or remedy mechanism was established in these cited official materials.",
      "sourceUrls": [
        "https://www.baseten.co/",
        "https://www.baseten.co/security-practices/",
        "https://www.baseten.co/legal-hub/"
      ]
    },
    "assurance": {
      "score": 0,
      "reason": "No independent worker-level assurance covering the product supply chain was established.",
      "sourceUrls": [
        "https://www.baseten.co/",
        "https://www.baseten.co/security-practices/",
        "https://www.baseten.co/legal-hub/"
      ]
    }
  },
  "disclosureLimited": true
}
```

- [Baseten inference platform](https://www.baseten.co/) (primary; checked 2026-10-28)

- [Baseten Security Practices (updated September 2026)](https://www.baseten.co/security-practices/) (primary; checked 2026-10-28)

- [Baseten legal hub](https://www.baseten.co/legal-hub/) (primary; checked 2026-10-28)

## Military ties

Finding: 8/10

Assessment: https://ethicalstack.ai/providers/baseten/#military-contracts

No military relationship was found; the terms restrict specified weapons and intelligence uses.

Reason: 7 base + 1 safeguard adjustment = 8/10 (3 of 9 safeguard points; every 3 adds 1). No military relationship was established in the cited sources. This ranks above known defense ties without claiming a verified clean record.

Argument: Baseten requires customer agreements to bar specified weapons work and government intelligence or reconnaissance uses. Customers enforce these terms; Baseten can terminate for material breach. These are bounded contractual safeguards, with no independent enforcement record in the reviewed sources.

Type: argument; confidence: medium

Scope (dimension): Baseten corporate relationship; no specific deployment of Baseten inferred

Research reviewed: not recorded; evidence date: not recorded

Unknowns: This source review does not cover undisclosed customers, contracts, subcontractors or deployments.

```json
{
  "method": "military-ties-priority-v2",
  "entity": "Baseten",
  "company": "Baseten",
  "scope": "Baseten corporate relationship; no specific deployment of Baseten inferred",
  "components": {
    "restrictions": {
      "score": 2,
      "reason": "Section 3.3 requires end-user restrictions on nuclear devices, missiles, chemical or biological weapons and intelligence organizations; it is not a general military ban.",
      "sourceUrls": [
        "https://www.baseten.co/terms-and-conditions/"
      ]
    },
    "oversight": {
      "score": 1,
      "reason": "Customers enforce end-user terms; section 7.2 permits termination for material breach. No independent outcomes are reported.",
      "sourceUrls": [
        "https://www.baseten.co/terms-and-conditions/"
      ]
    },
    "transparency": {
      "score": 0,
      "reason": "The reviewed sources provide no defense-relationship disclosure.",
      "sourceUrls": [
        "https://www.baseten.co/terms-and-conditions/"
      ]
    }
  },
  "disclosureLimited": true,
  "relationship": "not-established",
  "relationshipReason": "The official pages reviewed describe the product, service and user terms; they do not identify a defense relationship or deployment. No identity, nationality or unrelated-company inference is used.",
  "score": 8,
  "baseScore": 7,
  "safeguardPoints": 3,
  "safeguardAdjustment": 1,
  "rawScore": 8,
  "cap": null
}
```

- [Baseten inference platform](https://www.baseten.co/) (primary; checked 2026-10-28)

- [Baseten Security Practices (updated September 2026)](https://www.baseten.co/security-practices/) (primary; checked 2026-10-28)

- [Baseten legal hub](https://www.baseten.co/legal-hub/) (primary; checked 2026-10-28)

- [Baseten Terms and Conditions, sections 3.3 and 7.2](https://www.baseten.co/terms-and-conditions/) (primary; checked 2026-10-29)

## BDS targets

Finding: Not listed

Assessment: https://ethicalstack.ai/providers/baseten/#israel-palestine

Baseten is not named in the two official BNC campaign lists checked on 2026-10-28; this is not an endorsement.

Reason: Campaign category, excluded from numerical scoring.

Argument: The check is limited to the BNC consumer-boycott guide and No Tech campaign. Neither list names Baseten; the result says nothing about other campaigns or suppliers.

Type: argument; confidence: medium

Scope (profile): Baseten Model APIs, dedicated inference and asynchronous inference; customer-hosted deployments have separate customer-controlled logging.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: This check does not cover every BDS campaign, local campaign, supplier, investment or subsequent list change; it is not an endorsement or clean-record conclusion.

- [BNC live Guide to BDS boycott targets](https://bdsmovement.net/Guide-to-BDS-Boycott) (campaign; checked 2026-10-28)

- [BNC No Tech campaign](https://bdsmovement.net/no-tech-oppression-apartheid-or-genocide) (campaign; checked 2026-10-28)

## Reading the evidence

- Assessments are provisional editorial judgments. Cite the profile and the underlying sources, and retain confidence, scope and unknowns.

- Unknown, not-applicable and model-dependent are different states. Missing evidence does not establish good or bad conduct.

- Company-level evidence does not establish use of a particular product or model. Account plan, model, host and route can change an assessment.

- Energy scores describe documented practices and transparency, not measured request emissions. Zero component points can mean an evidence gap.

- Published overall scores require at least four of six numerical dimensions. Preserve coverage and BDS adjustments; BDS itself is categorical.

- No named BDS target found is limited to the reviewed campaign sources and is not approval. Personalized priorities do not change the published overall score.

- Review and source-check dates describe the research, not this deployment. Fetch current primary sources when the decision requires current verification.

Methodology: https://ethicalstack.ai/methodology/