Salad, a distributed cloud for AI (like Airbnb for GPUs)
Hey HN! Bob here, Founder and CEO of Salad (www.salad.com), a distributed cloud for AI inference at scale. *Why Salad?* Because there’s 400 Million consumer GPUs in the world (~100 Million AI-enabled RTX GPUS) and most of them lie unused for 18-22 hrs a day. These consumer GPUs offer similar or better cost-performance than high-end AI GPUs for many AI/ML use cases (Speech-to-text, batch jobs, text-to-image, computer vision to name a few). With Salad, we have activated this latent compute in everyday PCs to power AI/ML workloads. *How big is our network?* More than 1,900,000…
In plain words
Salad is a distributed cloud platform that connects consumer GPUs from personal computers to AI companies needing inference compute power. It leverages the estimated 400 million idle consumer GPUs worldwide—most unused 18-22 hours daily—to run AI workloads like speech-to-text, image generation, and computer vision. The platform serves AI companies and data scientists seeking cost-effective GPU access while allowing individual PC owners to earn by sharing their unused computing resources. Salad operates a network of over 1.9 million participants with thousands of GPUs actively running workloads.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Hey HN! Bob here, Founder and CEO of Salad (www.salad.com), a distributed cloud for AI inference at scale. *Why Salad?* Because there’s 400 Million consumer GPUs in the world (~100 Million AI-enabled RTX GPUS) and most of them lie unused for 18-22 hrs a day. These consumer GPUs offer similar or better cost-performance than high-end AI GPUs for many AI/ML use cases (Speech-to-text, batch jobs, text-to-image, computer vision to name a few). With Salad, we have activated this latent compute in everyday PCs to power AI/ML workloads. *How big is our network?* More than 1,900,000 individual PC owners on the network. Around 10K+ GPUs (RTX/GTX series) are running AI workloads from companies at any given time. *What’s in it for AI companies & data scientists?* Users get GPUs at the lowest market price with high scalability *What’s in it for people sharing GPU time?* GPU owners get rewards ranging $20 - $180 depending on the GPU type (Higher vRAM GPUs earn more). These rewards can be exchanged in our marketplace for gift cards, games and more. *How are our prices?* GPU prices start from $0.02/hour and across different consumer GPUs, we have the lowest price and largest scale in the market. For ex, an RTX 3080Ti for AI transcription costs $0.0012/hour with a self-managed OS model on Salad. The next best cost in the market is an API ($0.28/hr). *Who can run on SaladCloud today?* Here are the best fits for our infra today: (1) Needs >3 GPUs running 24*7 (2) AI inference workloads or batch processing (3) No compliance-related workloads (HIPAA, etc). As a distributed cloud, we are workload-agnostic. That being said, with our current features & team, we are best poised to support AI/ML workloads. As we grow, we will add more support for other workloads. We’re happy to offer $1000 in free credits if you have a demand for more than 10 GPUs per month running consistently. Just sign up for a demo call and select ‘HackerNews’ in the ‘How did you hear about us?’ field. Book a call to discuss your use case: https://salad.com/get-a-demo Signup for SaladCloud here: https://portal.salad.com/ To earn with your PC, download the Salad app here: https://salad.com/download GitHub Repo: https://github.com/SaladTechnologies Our mission is to democratize cloud computing with a fully people-powered cloud and helping companies break free from the cloud monopoly.
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
