Perplexity and Nvidia Partner for Local-First AI Platform
bakawolf123 · reddit · 2026-08-26
Perplexity is partnering with Nvidia for a local-first AI platform, utilizing Qwen models (likely 27B or 3.8 flash) on DGX Spark hardware. Most workloads will run on-device with occasional cloud access under strict privacy rules. This move addresses rising cloud costs and leverages local hardware.
More from Companies & People
- Understanding trumps prompting skills when hiring colleagues — littmath · 2026-08-26
- NVIDIA applied DL research VP on specialized teacher models for the open ecosystem — arena · 2026-08-26
- JHU ISI Welcomes Erik Rye, member of ISI and JHU IAA, and Assistant Professor at JHU CS, focusing on AI research. — mdredze · 2026-08-26
- A Founders Fund-backed startup slides "Humans replace AI" on keynote — iamfakhrealam · 2026-08-26
- SpaceX launches a new page mapping its launch sites and facilities — XFreeze · 2026-08-26
- x402 follows the Kubernetes playbook: 75M transactions this month — kleffew94 · 2026-08-26