Phala Network

Phala’s rolling deep into TEE at EthCC 🇫🇷

Join us on July 2 for the Afternoon TEE Party with our friends @iEx_ec, @OasisProtocol & @AutomataNetwork. Expect a lot of cocktails, sandcastles, and Web3 privacy.

⏰ July 2, 6:00 PM CEST (UTC+2)
📍 RSVP: https://lu.ma/oo9j90j1

Phala Cloud saw growth in the past week with more users, more VMs, and higher vCPU demand.

Sure, GPU TEE usage dipped 23.6% — but on the bright side, LLM token usage for Qwen 2.5 7B Instruct went through the roof since early June. Smarter AI, sleeker compute.

Phala Cloud saw growth in the past week with more users, more VMs, and higher vCPU demand.

Sure, GPU TEE usage dipped 23.6% — but on the bright side, LLM token usage for Qwen 2.5 7B Instruct went through the roof since early June. Smarter AI, sleeker compute.

Phala Cloud saw growth in June with more users, more VMs, and higher vCPU demand.

Sure, GPU TEE usage dipped 23.6% — but on the bright side, LLM token usage for Qwen 2.5 7B Instruct went through the roof since early June. Smarter AI, sleeker compute.

🤖 AI runs on GPUs, NOT CPUs.

Modern AI workloads—think LLMs, real-time inference—need massive parallelism. GPUs like NVIDIA’s H100 do that beautifully. But here's the harsh truth: once your data hits GPU memory, it's COMPLETELY exposed.

GPU TEEs change the game.

🧵

☁️🔥The stats are in.

The past week (June 11–18) was a strong week for Phala. We saw growth across all areas—users, VMs, and vCPUs—but Phala's GPU TEE usage on OpenRouter stood out, surging from 296M to 657M: a 122% increase🔥 in secure compute demand (

🔒 Case Study: Encifher x Phala Network 🔒

@encifherio taps into Phala’s TEE-powered Dstack to launch a privacy-preserving co-processor enabling secure, verifiable, off-chain computation on existing blockchains.

🔥 Bonus: TEE-secured threshold decryption guards against MPC collusion.

Phala Cloud grew to 4,849 users this week (+4.5%), while activity across compute and usage saw a noticeable dip.

Growth continues, but patterns are shifting. 🫡

🧊 Here's the hard truth about AI in finance: It's hard for people to trust AI when there's real money involved.

And for good reason. Would YOU give an AI access to your wallet when it's running on some random cloud server?

Hell no. 🔥

Not unless that cloud is TEE-based. 🧵

😶‍🌫️ Phala Cloud’s been putting in work — Here’s a quick snapshot of last week’s growth.

Most key areas saw solid growth📈. The only dip was in GPU TEE usage, partly because Nous: Hermes 3 was removed from Phala’s LLM offering on OpenRouter (https://openrouter.ai/provider/phala).

Phala × PublicAI: Building Trustworthy AI from the Ground Up

Phala and @PublicAI_ are teaming up to fix AI’s biggest problems: ❌trust, ❌privacy, and ❌transparency in training data.

Here’s how TEE + decentralized data annotation changes everything.

🧵

🧵 Phala × PublicAI: Building Trustworthy AI from the Ground Up

Phala and PublicAI are teaming up to fix AI’s biggest problems: ❌trust, ❌privacy, and ❌transparency in training data. Here’s how TEE + decentralized data annotation changes everything.

🧵

🤝 Partnership Announcement: Phala x PublicAI

Phala is partnering up with @PublicAI_ to set a new standard for ethical AI. By combining Phala's TEE-secured infrastructure with decentralized data annotation, we’re enabling AI training verifiable, private, and tamper-proof — from dataset to model.

Phala Cloud saw steady growth last week (May 21-28), with a 19.0% increase in users, 21.1% in VMs, and 11.4% in vCPUs.

🔥But the real headline? Phala's TEE GPU usage surged 131.3% after adding Nous Hermes 3 and Qwen 2.5 to our LLM lineup on OpenRouter (

Just dropped - Phala Cloud's Flash VRF Template, a turnkey Verifiable Random Function (VRF) service that lets devs deploy VRFs in MINUTES, not days.

It gives you:
⚡ TEE-secured randomness
⚡ 2-4s latency
⚡ Cryptographically verifiable
⚡ One-click deployment

A 🧵.

You chase dollar per hour.
I chase security per dollar.

We’re not the same.

Partnership Announcement: Phala 🤝 OODA AI

We’re joining forces with @OODA_AI_, a Nasdaq-listed AI powerhouse trading under ticker $OODA, focused on AI Apps, API:s & Agents, GDPR-grade data security, and decentralized AI computing.

Together, we’ll advance safe AGI via TEE-secured AI for both CPU & GPU compute.

Phala Cloud is a place where you build products, not just run workloads. We turn builders into shippers - from running code to running products.

Your brand. Your launch. Your Cloud.

Start Now: https://cloud.phala.network

🤝 New Partnership: Phala x Caila

Phala is excited to team up with @Caila_AI by @nubilanetwork! With Phala's TEE-secured infra, Caila agents can now process sensitive weather & location data in real-time - knowing that their LLM inference, user data and model integrity are protected.

Phala was here before TEE was cool.