Beyond phones: How Meta is building a hardware ecosystem around Muse AI

Meta is spreading Muse across smart glasses, a Neural Band, VR hardware and the new Charm device, distributing AI interaction across hardware rather than concentrating it on a smartphone

Meta has expanded the hardware around its Muse AI agent, with new smart glasses, a compact AI device, virtual reality glasses and more ways to interact with its software without reaching for a smartphone.

 

The announcements at Meta Connect 2026 point to a broader hardware strategy than adding AI features to individual products. Meta is distributing different parts of the interaction with Muse across devices. Cameras on glasses can provide visual context, microphones can capture voice and surrounding audio, the Meta Neural Band can translate hand movements into controls, while the new Muse Charm gives the agent a dedicated mobile device. 

 

What Meta announced at Connect 2026

Meta's new hardware spans several categories, from conventional-looking glasses to a dedicated AI device powered by its Muse AI agent, and Virtual Reality (VR) hardware designed around a separate compute unit.

 

Meta VR Glasses

 

Meta VR Glasses are the company's attempt to move some of the capabilities of its Quest headsets into a lighter glasses form factor. The device weighs around 100g, compared with 515g for the Meta Quest 3. Meta achieved the reduction by separating the system into two pieces. The glasses contain the displays and sensors, while a puck connected through an optical cable houses the processor, battery and storage.

 

The glasses use micro-OLED panels for what Meta calls its 5K Infinite Display, with a resolution density of 37 pixels per degree. Dolby Vision is supported for video, while Dolby Atmos spatial audio is built into the frames.

 

Meta is also integrating Muse directly into the operating system. Users can speak to the agent or use eye and hand tracking to open applications, search, control media and arrange virtual workspaces without physical controllers. The puck uses Qualcomm's Snapdragon Reality Elite processor and provides up to three hours of high-resolution media playback, according to Meta.

Meta is positioning the device for entertainment and computing. It supports multiple virtual screens and can turn a flat surface into a virtual keyboard and touchpad. The company has also partnered with media and sports companies for 3D films and immersive live sports, including more than 100 events annually once the glasses launch.

 

Ray-Ban Meta gets a third generation, alongside camera-free audio glasses

 

Meta is also widening its smart-glasses range.

 

The newly launched Ray-Ban Meta (Gen 3) retains a 12MP camera and can record 3K video, but gets a slimmer design, a customisable action button for Meta AI and up to nine hours of battery life. Meta has also added a six-microphone array designed to reduce background noise during calls and voice commands.

 

Meta has also unveiled new Ray-Ban Meta Audio glasses that take another approach by removing the camera, in part to address privacy concerns around camera-equipped smart glasses and the possibility of recording people without their knowledge. The 43g glasses are built around speakers, microphones and access to Meta's AI, with up to 12 hours of battery life and another 48 hours through the charging case.

Meta said Muse will be available across its AI glasses, giving users access to the same personal agent through voice while away from the phone.

 

Meta Ray-Ban Display is becoming a larger interface

 

Meta Ray-Ban Display, which was introduced last year alongside the Meta Neural Band, is also getting more functions that previously required a phone.

 

The glasses will add cycling and public transport directions, with navigation information appearing directly in the display. Walking navigation can also use visual landmarks from the user's surroundings rather than relying only on street names.

 

Meta is adding an Explore app for experiences designed specifically for the glasses, along with richer calendar controls and mirrored smartphone notifications.

 

Another feature, called Hologram, creates a digital representation of the user for video calls. After an initial face capture through the Meta AI app, the model uses the person's voice to infer facial expressions during calls. Meta plans to begin testing the feature through WhatsApp in the US later this year.

 

The glasses work with Meta's Neural Band, an electromyography or EMG wristband introduced last year. The band reads electrical signals generated by muscle activity in the wrist, allowing small hand movements to become controls for the glasses.

 

Muse gets its own hardware with Charm

 

The more unusual addition is Muse Charm, a small dedicated device built around Meta's recently launched personal AI agent.

 

According to Reuters, Charm is roughly the size of an AirPods case and includes a roughly two-inch touchscreen and its own 5G connection. Meta expects the device to ship around the holiday period in December, though it has not announced a price.

 

The device gives Muse another interface outside the smartphone. Meta launched Muse earlier this month as an agent capable of handling tasks such as sending emails and booking travel. It runs inside a dedicated cloud-based virtual machine and can connect to services authorised by the user.

Meta is breaking the smartphone into smaller pieces

Taken together, Meta's announcements show a clearer division of computing between devices.

 

A smartphone combines cameras, microphones, display, touch input, connectivity, applications and processing in one device. Meta's emerging hardware ecosystem distributes some of those functions according to where they are useful.

 

Glasses can provide Muse with the visual and audio context around the wearer. The Neural Band adds a discreet input system based on hand movements. Display glasses can put directions, notifications and AI responses in the user's field of view. Charm provides a small dedicated interface for the agent, while the VR glasses move heavier visual computing into a separate display-and-compute system.

 

Muse sits across those products as the common software layer.

 

This is less a replacement for the smartphone than a redistribution of what happens around it. Smartphones remain useful because they combine cellular connectivity, storage, processing, mature app ecosystems and a person's accounts and digital identity. Peripheral devices do not need to duplicate all of that hardware if some computing, data and services can be shared with a phone or cloud system.

 

Meta's approach also gives its AI something smartphones struggle to provide continuously: physical context. A phone sitting in a pocket has limited information about what its owner is looking at or doing. Glasses sit close to the eyes and ears, while a wristband can detect intentional movements. Together, those inputs can give an AI agent more information about the immediate environment and how a person wants to interact with it.

Meta is not alone in building hardware around AI

OpenAI is working with former Apple design chief Jony Ive on dedicated AI hardware after acquiring Ive's io Products startup. The first product was originally expected in 2026, but court filings earlier this year indicated that it would not ship before February 2027.

 

The project has since become entangled in a separate dispute with Apple. Apple sued OpenAI and two former employees in July, alleging that trade secrets were taken to support OpenAI's move into consumer hardware. OpenAI has denied the allegations. Apple has also sought an injunction that could restrict work on products allegedly based on its technology.

 

Apple itself is exploring hardware designed to give Siri more information about the user's surroundings. Bloomberg reported in February that the company was developing smart glasses, a pendant that could be worn or pinned to clothing, and AirPods with cameras, with the devices designed around providing visual context to Siri.

 

Bloomberg reported in May that the camera-equipped AirPods had reached advanced testing, with prototypes close to their intended design. The cameras are meant to capture information about the surrounding environment that can be interpreted by Apple's AI systems.

 

Samsung and Google are also developing AI hardware around the Android XR platform. Samsung has already launched the Galaxy XR headset, the first device built on Android XR, with Google's Gemini integrated at the system level. Gemini can understand what the user is seeing and hearing and respond through voice, vision and gesture.

 

The two companies are also working on intelligent eyewear based on Android XR, with Samsung and Google partnering with eyewear brands including Gentle Monster and Warby Parker. Google said the glasses will come in audio-only and display variants, with Gemini providing assistance through voice and, on display models, information in the user's field of view. The glasses are designed to work with a smartphone while allowing users to access directions, messages, translations and other functions without taking the phone out.