@PhalaNetwork

Empowering 10,000 builders and companies to build scalable, private, and safe intelligence.

 San Francisco, CA
phala.com
Joined August 2019
  • Tweets6.3K
  • Following835
  • Followers141K
OPPO × Phala published a joint paper on Kubernetes Pod-level remote attestation for confidential AI. Goal: prove the hardware, container image, and Pod identity before sensitive data enters the workload. arxiv.org/abs/2606.03323
19
31
10
183
54,009
Proud to build private, verifiable compute for AI agents with @soon_svm. 🤝
SOON - Solana Optimistic Network (Mainnet Arc)
@soon_svm
11h
SOON Network's Strategic Investment in Phala's TEE GPU Cluster Powers Private, Verifiable Compute for AI Agents The core infrastructure developer behind SOON is a major funder of Phala Network’s TEE GPU cluster. AI agents in the SOON ecosystem can access this capacity and run custom models in a private, trusted environment. Why it matters AI agents entering capital markets need high-performance execution and private, verifiable compute. SOON now brings both together. • Private strategies, verifiable execution. For quant agents, custom models and trading strategies are the core asset. Running inside Phala’s TEE keeps models and data sealed, while remote attestation lets anyone verify that each execution ran as intended. • Dedicated SVM chains for any team. SOON Stack is currently the only SVM stack that launches dedicated chains for third parties across multiple L1s, including Ethereum, BNB Chain and Base. Every chain gets the throughput and low latency of parallel execution. The vision Our goal is for every agent to run its own chain with Solana-grade performance: paying for its own blockspace and compute, operating 24/7 without human ops, while its owner always retains final control. The SOON AI Capital Market will serve as the execution and settlement layer for these agents. What's next • The revenue-sharing framework for this compute, and how it connects to the SOON ecosystem, will be detailed in an upcoming announcement. • Onboarding plans for the first AI Capital Market projects will follow shortly. "AI capital markets need more than speed — they need private, verifiable compute, and that’s exactly what Phala brings to SOON." — Marvin Tong, Phala Network This announcement contains forward-looking statements. Plans may change
5
21
3
151
9,797
Important: our official Discord invite link has been updated. The previous link (discord.com/invite/phala-net…) has expired. Please use the new one instead: discord.gg/saGuzKq5Wm CoinMarketCap and CoinGecko will be updated shortly. Be careful of fake links and scams.

Join the Phala Discord Server!

Check out the Phala community on Discord - hang out with 610 other members and enjoy free voice and text chat.

discord.com
7
9
118
6,578
Phala Confidential AI: 112.21 billion tokens over the past 24 hours. Top models: GLM-5.3 Flash 65.11% GLM-5.3 8.81% qwen3.5-397b-a17b 8.40% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
3
16
115
7,963
Phala Confidential AI: 202.74 billion tokens over the past 24 hours. Top models: GLM-5.3 Flash 42.98% GLM-5.3 24.90% deepseek-v4.1-flash 13.25% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
5
16
113
7,743
1/ SayGM built its privacy-first AI inference gateway on Phala Cloud. Every request routes through a hardware-attested confidential VM running an approved, pinned image, making the workload behind each request independently verifiable. saygm.com
7
26
3
126
11,536
2/ The attestation server and data-plane proxy run inside the same measured workload. The registry checks the image hash before traffic starts, while operator credentials stay sealed inside the confidential VM. dstack: github.com/Dstack-TEE/dstack

GitHub - Dstack-TEE/dstack: Open framework for confidential AI

Open framework for confidential AI. Contribute to Dstack-TEE/dstack development by creating an account on GitHub.

github.com
3
4
43
2,980
3/ Live today: 60 coldkeys deployed, 18 operators active, and 57 confidential VMs running. Independent operators can serve AI traffic under one verifiable runtime. Full case study: phala.com/posts/saygmprivacy…

Phala x SayGM: Privacy-first AI inference gateway built on Phala Cloud | Phala

Phala Cloud delivers confidential AI on TEE-protected GPUs — private LLM inference, sealed agents, and verifiable compute on Intel TDX + NVIDIA H100/H200.

phala.com
1
17
853
Phala Confidential AI: 173.64 billion tokens over the past 24 hours. Top models: GLM-5.3 51.89% GLM-5.3 Flash 25.07% deepseek-v4.1-flash 10.61% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
5
13
3
114
7,938
Phala Confidential AI: 254.51 billion tokens over the past 24 hours. Top models: GLM-5.3 70.49% Kimi K3 10.73% GLM-5.3 Flash 8.18% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
12
20
1
128
8,444
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% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
6
13
105
7,243
Phala Confidential AI: 168.69 billion tokens over the past 24 hours. Top models: GLM-5.3 42.87% deepseek-v4.1-flash 19.24% GLM-5.3 Flash 12.00% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
7
18
1
112
8,861
Phala Confidential AI: 55.18 billion tokens over the past 24 hours. Top models: GLM-5.3 Flash 49.10% GLM-5.3 14.71% Kimi K3 12.62% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
5
10
93
6,283
where's @marvin_tong and @bgmshana ? how's my confidential GPU built?
Marvin Tong
@marvin_tong
Sep 9
How to build sovereign AI from first principles? where we disappear:
1
8
85
7,554
Phala Confidential AI: 78.09 billion tokens over the past 24 hours. Top models: GLM-5.3 43.54% GLM-5.3 Flash 26.96% DeepSeek V4 Flash 11.46% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
6
10
2
93
6,694
If chain-of-thought monitoring weakens as agents gain tools, safety cases must move beyond inferred intent: constrain permissions, verify actions at runtime, and test outcomes under adversarial conditions.
Jakub Pachocki
@merettm
Sep 6
I wrote about the state of AI, why I’m concerned about the next few years, and the choices we need to make to keep the future in humanity’s hands. An Alien Mind: openai.com/index/an-alien-mi…

An Alien Mind

Jakub Pachocki reflects on increasingly capable AI and the challenge of keeping it aligned. He calls for stronger safeguards and international coordination.

openai.com
5
6
75
7,410
Qwen3 8B is live on Phala—served inside a TDX-attested GPU TEE. Private, verifiable inference for coding, long-horizon agents, reasoning. phala.com/models/qwen/qwen3-…
2
13
85
6,972
41K context. 8.2B parameter causal language model from the Qwen3 series. Reasoning, tool use, structured outputs. $0.11/M input · $0.45/M output. docs.phala.com/phala-cloud/c…

Phala Cloud Documentation — Confidential AI on TEE

Deploy confidential AI applications on Phala Cloud with hardware-level privacy via Intel TDX, NVIDIA H100 GPU TEE, and verifiable attestation.

docs.phala.com
1
23
2,724
Phala Confidential AI: 50.72 billion billed input + output tokens over the past 24 hours. Top models: 1) z-ai/glm-5.3-flash 54.28% 2) deepseek/deepseek-v4-flash 17.77% 3) z-ai/glm-5.3 7.01% phala.com/models/z-ai/glm-5.… TEE-backed AI infrastructure.
3
10
99
6,522
Phala Confidential AI: 38.86 billion billed input + output tokens over the past 24 hours. Top models: 1. DeepSeek V4 Flash — 28.59% 2. GLM-5.2 — 13.61% 3. GLM-5.3 — 9.53% Explore #1: phala.com/models/deepseek/de… TEE-backed AI infrastructure.
5
6
86
6,593
ViMax transforms ideas, novels, or scripts into multi-scene video plans and final generated clips. The Phala template uses a source_verifier runtime. cloud.phala.com/templates/vi…
1
5
62
5,325
more replies
By default, this template does not run video generation. Its /demo constructs upstream Pydantic models for a scene, shot brief, shot description, camera, and event. github.com/Phala-Network/pha…
10
1,750
Load more