Snap Inc. has officially launched Specs, its long-teased consumer augmented-reality headset, at $2,195. Announced in Beverly Hills on September 16, 2026, the wearable ships alongside Specs Intelligence, an AI assistant that spans the glasses, iPhone and Mac, plus platform tie-ins with NVIDIA, Salesforce and AWS.
Two Snapdragons, two displays and a lot of software
Specs is a fully self-contained computer, not a phone-tethered accessory. Snap put a Snapdragon processor in each temple, dual see-through displays, six cameras and microphones inside a chunky but wearable frame. Battery life and weight are the trade-off for the wearable's independence.
The launch price positions Specs above Meta's $799 Ray-Ban Display models but below dedicated mixed-reality headsets. CEO Evan Spiegel framed the price as an on-ramp for enterprise, saying "the goal with both Salesforce and NVIDIA is to make it really easy for businesses to get the benefits of those agentic platforms".
Specs Intelligence targets agentic AR
Specs Intelligence is Snap's own AI assistant, designed to anticipate what wearers are about to do from time-of-day, calendar signals and app usage across Specs, iPhone and Mac. Snap has not fixed pricing beyond confirming a free tier and paid usage-based tiers. The stack sits alongside a broader push into agentic AI that Salesforce and NVIDIA are jointly commercialising.
Enterprise-first, courtesy of NVIDIA and Salesforce
NVIDIA is contributing an open-source XR AI platform for developers to build spatial applications on top of Specs, echoing its physical-AI push at the DSX factory. Salesforce joins as an agentic platform partner so field techs, insurance adjusters and warehouse workers can pull Agentforce workflows through the display. AWS is also on the platform list, though Snap gave no cloud-service specifics.
A big bet on the "post-smartphone" era
Snap has invested in AR for more than a decade — from Spectacles to Camera Kit to Lens Studio — but Specs is its first true consumer wearable computer. Spiegel is betting that the hardware, plus a developer platform and Specs Intelligence, can carve out enterprise volume even as Meta and Apple push their own wearable AI stacks.
Reporting based on coverage from Android Headlines, CNBC, Engadget and Dataconomy.