Seeded record — this provider has not yet been through the independent verification pass. Treat every claimed/unverified field below as unconfirmed.

Identity

Category	TEE play
Tier	1
Record status	Unverified
First tracked	2026-08-24
Last checked	2026-10-12
Last changed	2026-10-12
Website	https://phala.network

Phala Network runs TEE-backed confidential AI infrastructure and serves inference through OpenRouter as one of its provider listings, hosting open-weight model families such as GLM- and DeepSeek-class models. Confidential execution is the project's core pitch: operators are claimed to be unable to see the prompts and outputs running through their hardware. Throughput is described only qualitatively in the roster sweep ("high throughput claimed"), with no specific figure. Neither the attestation nor a direct API probe has been run for this seed batch.

API

Field	Value	Grade	Note
OpenAI-compatible	—	Unverified	Served via OpenRouter as a provider; a direct API surface has not been independently reviewed.
Endpoint	—	Unverified	—
Status	live	Claimed	—
Models	GLM-class, DeepSeek-class	Claimed	Model family names as given in the roster sweep; not enumerated directly.
Access	openrouter	Claimed	—

Privacy

Field	Value	Grade	Note
Mechanism	TEE (confidential compute)	Claimed	TEE-backed confidential execution claimed; operators claimed 'blind' to prompts, not independently attested.
Attestation independently verifiable	false	Unverified	—
Zero-retention	—	Unverified	—

Scale

Field	Value	Grade	Note
Tokens/day	—	Claimed	Roster describes 'high throughput claimed' without a specific figure.
ARR estimate	—	Unverified	—
Pricing published	—	Unverified	—

On-chain

Field	Value	Grade	Note
Token	PHA	Claimed	Relationship between the network token and inference revenue not yet queried on-chain.
Self-dealing flag	—	N/A	—

Change log

2026-08-24 — record: seeded from provider roster v1.0 (Unverified) (source)
2026-09-14 — scale.tokens_per_day: X claim (c_scale): Phala Confidential AI: 78.09 billion tokens over the past 24 hours. (Claimed) (source)
2026-09-21 — scale.tokens_per_day: X claim (c_scale): Phala Confidential AI: 178.40 billion tokens over the past 24 hours. Top models: GLM-5.3 47.87% deepseek-v4-pro-0813 12.79% GLM-5.3 Flash 10.08% https://phala.com/models/z-ai/glm-5.3 TEE-backed AI infrastructure. (Claimed) (source)
2026-10-05 — privacy.mechanism: X claim (e_privacy): Your coding agent can read your entire codebase. Private AI Proxy verifies the AI service before that context leaves your machine, using hardware attestation, TLS key pinning, and fail-closed forwarding. Get v0.3.0 for macOS, Windows, or Linux: https://github.com/Dstack-TEE/private-ai-gateway/releases/tag/desktop-v0.3.0 (Claimed) (source)
2026-10-05 — record: X claim (f_funding): SOON Network's Strategic Investment in Phala's TEE GPU Cluster Powers Private, Verifiable Compute for AI Agents (Claimed) (source)
2026-10-12 — privacy.mechanism: X claim (e_privacy): @PhalaNetwork lets developers run AI models in protected CPU/GPU environments that isolate prompts from infrastructure operators. Phala's attestation lets clients check hardware and deployed software. Bind the encrypted connection to that verified environment. (Claimed) (source)

The change log is append-only: corrections are recorded as new entries, never rewrites. Full feed via /api/v1/changes.

Sources

[roster] docs/decentralized-inference-provider-roster.md (2026-08-24)
[website] https://phala.network (2026-08-24)
[x-account] https://x.com/PhalaNetwork (2026-08-24)