Akash Accelerate 2025
2025-06-23
Brooklyn, United States
as the organisers put it
Akash Accelerate is a full-day summit that will bring together the entire ecosystem for a packed agenda of talks, panels, and discussions exploring the future of decentralized compute and open-source AI.
| title | author | length |
|---|---|---|
| Nous Research | Dillon Rolnick | 16m |
Nous Research describes a working open alternative to closed labs: a Psyche network that trains a 40B-parameter model on ~500 billion tokens via a global… | ||
| Panel: Investing in Intelligence | Cheng Wang, Maja Vujinovic, Conor Moore, Mark Palmer, Christian Lopez | 53m |
Panel drills into concrete moves: asset‑backed crypto debt built from GPU farms and GPU-backed ABS; project financing versus equity for data centers; observed… | ||
| Passage | Lex, Ariel Avellino | 10m |
Passage delivers full Unreal Engine 5 worlds streamed to any browser or phone with zero client processing, running on aos and claiming a 70% GPU cost… | ||
| The Roadmap to a User-Owned AI Model | Illia Polosukhin | 17m |
Proposes a practical stack for user-owned AI that uses trusted execution environments (Intel TDX, NVIDIA Hopper/Blackwell) to run models confidentially, with… | ||
| Highlights | (no speaker listed) | 02m |
Akash frames a push for a decentralized cloud where nodes can join the platform and workloads run on a peer network, arguing power is becoming the primary… | ||
| Is DePIN Real? | Tommy Eastman, Haseeb Qureshi, Mahesh Ramakrishnan, Abhay Kumar | 32m |
Discussion gives concrete signals: total DePIN market cap sits around $3–10B with on-chain revenue is tiny; Helium boosts adoption by remove crypto from… | ||
| Akash Accelerate 2025: Nirmal Krishman from Chakra | Nirmal Krishman | 10m |
Chakra builds an open-data protocol with a nextG data warehouse that goes live within a month of incorporation and claims 80% of Snowflake at 20% of the cost… | ||
| Beyond Centralized Models | Casey Caruso, Jake Brukhman, Benjamin Parr, Chad Fowler | 35m |
Panelists give concrete wins and gaps: a decentralized inference provider beats Google on price by 90% cheaper by aggregating idle GPUs; Plurales demonstrates… | ||
| AI Content Generation | Zack Abrams | 12m |
Open-source workflows demonstrate granular control: using Comfy UI with Hydream the presenter runs the sampler for 40 out of 100 steps, swaps prompts… | ||
| Fireside chat | Javier Villamizar, Greg Osuri | 24m |
Announces Stargate $500 billion and a claimed $1 trillion Arizona‑Illinois AI robotics hub with robotics for rack assembly and liquid/immersion cooling to… | ||
| Decentralized Energy | Jason Badeaux | 13m |
Daylight Energy lays out a concrete plan to tokenise distributed solar + battery systems, using on‑chain incentives to aggregate behind‑the‑meter assets and… | ||
| Keynote | Greg Osuri | 22m |
Claims a roadmap to scale decentralized AI with concrete milestones: Akash star cluster phases (2025 data‑center grade racks, 2026 telco edge, 2027… | ||
| Star Bonds | (no speaker listed) | 02m |
A fractional on‑chain ownership model called Starbox lets anyone reserve a stake while a node keeper provisions and operates nodes that run AI workloads and… | ||
| AkashML | (no speaker listed) | 02m |
Akash ML is a distributed AI runtime that turns idle hardware into production-grade infrastructure, built GPU native and container-first to run models across… | ||
| Starcluster | Greg Osuri | 02m |
A planetary‑scale mesh of decentralized AI compute claims to let independent node keepers run data‑center grade racks powered by renewables and join a global… | ||
| nothing matches | ||