Today’s “cloud” is concentrated in a handful of mega data centers. Acurast is flipping that model by turning smartphones into a confidential, verifiable compute network—made for Web3 and AI. Founder Dr. Christian Killer explains how they verify hardware, encrypt workloads end-to-end, and deliver trust-minimized compute for price feeds, bridging, LLM inference, web scraping, residential proxy/VPN, and even distributed agent indexing (MIT NANDA).
Timestamps
[00:00] Intro & the problem: compute is centralized in ~5 companies
[00:02] Christian’s path (gaming → Bitcoin Lightning NFC → Web3 infra)
[00:04] Why decentralize compute? Cost, control, and data misuse risk
[00:05] Phones vs servers: performance-per-watt, economics, upcycling
[00:07] The hard bit: verifying hardware + trusted execution environments
[00:08] How Acurast works (supply/demand, encrypted jobs, scheduling)
[00:09] Web3 demand: price feeds & bridging done trust-minimized
[00:10] Web2 demand: scraping + residential IP/VPN for AI startups
[00:11] AI focus: inference today, confidential LLMs, agent guardrails
[00:13] “Agentic AI” hype vs practice—what actually matters
[00:16] Distributed agent indexing (MIT NANDA) explained simply
[00:18] What most people miss about LLMs (probabilistic behavior)
[00:20] Quantum computing: progress, hype, and reality
[00:21] Biggest technical challenges ahead (clustering, MPC, DKG)
[00:24] Moore’s Law → performance per watt (phones keep winning)
[00:25] DePIN’s real challenge: organic demand & dev UX
[00:27] North star metrics & real adoption vs spam
[00:28] Roadmap: Cargo (containers), clustering, futures for compute
[00:30] The ask: strategic partners & builders; grants/hackathons coming
Connect
Disclaimer
Nothing mentioned in this podcast is investment advice and please do your own research. Finally, it would mean a lot if you can leave a review of this podcast on Apple Podcasts or Spotify and share this podcast with a friend.
Be a guest on the podcast or contact us - https://www.web3pod.xyz/
