Meta reached the consumer AI hardware market before OpenAI on September 23, 2026, unveiling a pocket-sized device called Charm at its annual Connect conference. The gadget runs Meta’s Muse AI agent and marks the company’s first standalone AI hardware product separate from its Ray-Ban and Oakley smart glasses lines. According to CNBC, Meta CEO Mark Zuckerberg revealed the device on stage but declined to name a price, and only a handful of working units existed at the time of the announcement.
The timing is the story. Business Insider reported that Meta “beat OpenAI to the punch on announcing an AI gadget,” landing its reveal roughly a week ahead of OpenAI’s own major developer event where a rival consumer device has been expected. That sequencing hands Meta the first-mover narrative in a category that both companies have been racing toward since OpenAI’s acquisition of former Apple design chief Jony Ive’s hardware startup io in 2025. Whether being first to announce translates into being first to ship, or first to sell, is a separate and far less settled question.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What Meta actually announced with Charm
Charm is described in the available reporting as a small, pendant-style device built around Muse, the AI agent Meta has been expanding since its initial launch earlier this year. Zuckerberg told the Connect crowd, “We’ve packed a lot of technology into this little guy,” according to Business Insider’s coverage of the event. The device includes a roughly 2-inch touchscreen and, according to Reuters reporting from December 2026, built-in 5G connectivity, and is meant to be used primarily through voice interaction with Muse, positioning it as a wearable companion rather than a standalone smartphone competitor.
Crucially, Meta has not confirmed a price for Charm. CNBC’s report is explicit on this point: no pricing was disclosed at Connect, and Reuters reported in December 2026 that Meta was still withholding pricing even as it confirmed Charm’s holiday shipping timeline. Zuckerberg did say the device is targeted to ship during the 2026 holiday shopping season, though the company had reportedly built only a small number of working units by the time of the announcement. That gap between “announced” and “shippable at scale” is the crux of why analysts and reporters are already framing this as an unproven strategy rather than a finished product launch.
Charm arrives alongside a broader hardware push. Meta also said it plans to offer more than 100 styles of AI-enabled glasses by the end of 2026, extending the strategy it has run with Ray-Ban Meta, where the $449 Ray-Ban Meta Gen 3 launched in September 2026 with a 12MP camera, 3K video capture, and nine hours of battery life according to Meta and TechCrunch, and, more recently, its lightweight VR glasses line, priced at $1,299 according to Reuters, and the camera-free Project Luna glasses, which arrived commercially as the $349 Ray-Ban Meta Audio starting in October 2026, offered in 23 frame and color combinations and marking Meta’s first fully camera-free AI-glasses model, per TechCrunch and Gadget Review. Charm is a different form factor entirely, but it fits the same thesis: Meta wants consumers interacting with Muse through devices it controls, not just through an app on someone else’s phone.
Why Meta is racing to own the hardware layer
The strategic logic behind Charm has little to do with pendant sales volume in year one. It has to do with distribution. Meta’s AI products, including Muse, currently reach most users through iOS and Android, both gatekept by Apple and Google. Every interaction funnels through platforms Meta does not control, subject to app store rules, review cycles, and commission structures Meta has chafed against for years. A dedicated device sidesteps that dependency, at least for the subset of users willing to carry a second gadget.
That is consistent with the direction Meta has signaled all year. The company has already rolled out Muse’s avatar features and Mac control tools, expanded voice capabilities, and pushed Muse deeper into wearables. Charm is the logical next step: a device whose entire reason for existing is to run Muse, unencumbered by a general-purpose operating system built by a rival platform owner. It mirrors, in miniature, the same instinct that led Amazon to build the Echo around Alexa or Google to build Pixel around Assistant and now Gemini.
The risk is that Meta is entering a product category with an unresolved value proposition. Wearable AI gadgets have a rocky track record. Humane’s AI Pin and the Rabbit R1 both launched to heavy press coverage in 2024 and then faded quickly once reviewers and early buyers found the core use case thin compared with simply talking to an AI assistant through a phone already in your pocket. Meta is betting that Muse’s maturity, plus the network effect of Meta’s billions of existing app users, will produce a different outcome. Nothing in the current reporting confirms that bet has been tested with real consumers yet, since Charm has not shipped.
OpenAI’s device: further along than confirmed, less far along than assumed
OpenAI’s own hardware ambitions are well known but still not fully public. The company has a high-profile consumer hardware effort underway, and reporting cited by The Tech Buzz indicates OpenAI was finalizing its own plans, with an announcement expected before the end of 2026. What the device looks like, what it costs, and when it actually ships all remain unconfirmed in the public record as of this week. Business Insider noted that OpenAI did not immediately respond to a request for comment on Meta’s announcement.
What is clear is the calendar pressure. CNBC placed Meta Connect about one week ahead of OpenAI’s major developer event, meaning Meta’s reveal landed squarely in the window when attention would otherwise be building toward OpenAI’s own news cycle. For a company that has spent 2026 racing Meta, Google, and Anthropic across model releases, this is one of the few times Meta has out-timed OpenAI on a major reveal rather than the other way around. Much of this year’s AI news cycle has run in the opposite direction, from OpenAI’s math-solving claims to its aggressive Astra model rollout, so a Meta-first hardware moment is a notable reversal.
It is worth being precise about what “beating OpenAI” means here. Meta announced first. It has not shipped first, priced first, or proven consumer demand first. Those distinctions matter because hardware categories are frequently won or lost after the announcement, during the unglamorous work of manufacturing at scale, hitting a price point people will pay, and building a software experience compelling enough to justify carrying an extra device. Meta has cleared the lowest bar in that sequence.
Muse’s role at the center of the device strategy
Muse is doing a lot of strategic work for Meta right now, and Charm is the latest extension of that bet. Meta has iterated on Muse rapidly through 2026, expanding it from a chat-first assistant into something closer to an operating layer across Meta’s hardware lineup, including the glasses business and now a dedicated pendant. The company’s Connect messaging framed Muse, not any single device, as the actual product, with Charm serving as one of potentially several physical containers for it.
That framing has upside and downside. The upside is flexibility: if Charm underperforms, Meta can keep investing in Muse through glasses, phone apps, and future hardware without having bet the company on one form factor. The downside is that it puts enormous pressure on Muse’s reliability and security track record, since any weakness in the underlying agent affects every device built on top of it. Muse has already had a rough middle stretch of 2026 on that front, including a zero-day vulnerability disclosure and a period where Amazon blocked the agent from operating on its platforms over access concerns. Shipping a device whose entire identity is “the Muse gadget” raises the stakes on Meta resolving those trust issues before Charm reaches real customers.
Historical context: Meta’s hardware record is mixed
Meta’s track record building consumer hardware outside its glasses partnerships is uneven. Reality Labs, the division responsible for VR and AR hardware, has posted billions of dollars in annual operating losses for several years running as Meta has poured money into headsets, controllers, and now AI-native devices. The company has also cycled through hardware strategies before: Portal video-calling devices launched in 2018 and were discontinued a few years later after limited consumer traction. Meta’s smart-speaker ambitions never scaled the way Amazon’s or Google’s did.
Where Meta has succeeded is in partnership-driven hardware: the Ray-Ban Meta smart glasses line, built with EssilorLuxottica, sold in the millions of units and became a genuine bright spot in a division otherwise known for losses, a shift The Verge has tracked closely as Meta’s hardware ambitions expanded beyond VR headsets, most recently with the September 2026 launch of the $449 Ray-Ban Meta Gen 3, which Meta says delivers nine hours of battery life on a single charge. Charm breaks from that formula. It is a first-party device with no outside hardware partner named in current reporting, closer in spirit to the failed Portal than to the successful Ray-Ban collaboration. That history is exactly why reporters covering the announcement, including CNBC, framed Zuckerberg’s strategy as unproven rather than validated, even as Meta claimed the announcement timing win.
Market and competitive landscape: who else is building AI hardware
Meta and OpenAI are not alone in chasing dedicated AI hardware. Google has built Gemini into Pixel phones and its Android ecosystem rather than launching a standalone device, betting on integration over a new form factor. Coverage from ARinsider places Charm within a broader wave of personal AI and XR hardware announcements moving through the industry this quarter, not an isolated Meta initiative. Amazon has leaned on Echo devices retrofitted with more capable Alexa+ features. Apple has been comparatively quiet on new standalone AI hardware, focused instead on improving on-device AI inside iPhone, though reports of expanded AI infrastructure partnerships suggest the company is not sitting still on the compute side even if it hasn’t announced consumer AI hardware of its own.
The wearable AI hardware category more broadly has a short and largely disappointing history, as noted above with Humane and Rabbit. What differentiates Meta and OpenAI’s entries, at least on paper, is distribution and capital. Meta can subsidize Charm with revenue from its advertising business, which still supplies the overwhelming majority of company revenue, and can cross-promote the device across Instagram, Facebook, and WhatsApp’s billions of users. OpenAI, backed by its own enormous valuation and partnerships including its financing arrangements with SoftBank, has comparable capital but a smaller existing consumer distribution footprint than Meta’s family of apps.
| Company | Product | Status as of Sept. 25, 2026 | Price disclosed | Target availability |
|---|---|---|---|---|
| Meta | Charm (Muse pendant) | Announced at Connect, few units built | Not disclosed | 2026 holiday season (targeted) |
| OpenAI | Unnamed consumer device | In development, pre-announcement | Not disclosed | Expected before end of 2026 |
| Amazon | Echo with Alexa+ | Shipping, iterative updates | Existing Echo pricing | Available now |
| Gemini on Pixel | Integrated into existing phones | Standard Pixel pricing | Available now | |
| Humane (historical) | AI Pin | Discontinued after weak sales | $699 plus subscription | Launched 2024, wound down |
What analysts and reporters are flagging as the risk
The core skepticism in the current coverage is not about whether Meta can build a device. It clearly can, and has shown working units. The skepticism is about demand and durability. Mainstream consumer AI hardware has not yet produced a breakout success story the way smartphones or smart speakers did in their respective categories. CNBC’s framing, that Meta reached the market first but the strategy remains unproven, captures the split between announcement-stage excitement and the much harder test of retail sell-through months from now.
There is also a manufacturing and scale question sitting underneath the marketing news. Building “a few units” for a stage demo is a wholly different operation from producing enough Charm devices to meet holiday demand at a price point consumers will accept, especially with memory and component costs elevated across the electronics industry this year. Reports have separately tracked sharp DRAM price increases squeezing device makers across categories in 2026, a cost pressure that would apply to Charm’s bill of materials just as it has to phones, laptops, and game consoles.
What this means for developers and the AI ecosystem
For developers building on top of Muse or watching the broader agentic AI space, Charm signals that Meta intends Muse to be a platform, not just a feature bolted onto Instagram or WhatsApp. That has implications for how Meta prioritizes its AI roadmap, likely favoring capabilities that work well through voice and glanceable screens over purely text-based chat features. It also raises the stakes for Muse’s API stability and security posture, since a hardware product built around the agent cannot tolerate the kind of outages or vulnerabilities that a phone app might absorb more quietly.
For the wider AI industry, Charm is another data point in a trend that has been building since mid-2026: major AI labs and platform companies no longer treat model quality alone as a sufficient moat. Meta, OpenAI, Google, and Amazon are all now competing on the physical layer where AI reaches consumers, not just on benchmark scores. That shift raises the bar for what counts as a competitive AI strategy heading into 2027, and it puts pressure on smaller AI labs without hardware ambitions or manufacturing partners to find other ways to stay relevant to everyday consumers.
Predictions: what to watch next
- OpenAI will confirm its own device within weeks. Given reporting that an announcement is expected before the end of 2026 and that OpenAI’s developer event falls roughly a week after Meta Connect, expect OpenAI to respond with its own hardware reveal rather than cede the news cycle entirely.
- Meta will disclose Charm pricing closer to the holiday window. Zuckerberg’s holiday-season shipping target implies a pricing announcement is likely in October or November 2026, once manufacturing scales beyond the handful of units shown at Connect.
- Early Charm reviews will focus on the has this been done before problem. Given the Humane AI Pin and Rabbit R1 precedents, expect reviewers to press hard on what Charm does that a phone with Muse installed cannot, and Meta will need a clear answer.
- Meta’s glasses lineup will keep expanding in parallel. The stated plan for 100-plus AI glasses styles by the end of 2026 suggests Charm is one leg of a multi-device strategy rather than a sole bet, spreading Meta’s hardware risk across form factors.
- Component costs will shape both companies’ launch pricing. With memory and component prices elevated industrywide in 2026, both Meta and OpenAI will likely price their devices higher than comparable 2024-era gadgets like the AI Pin, testing consumer willingness to pay premium prices for unproven hardware categories.
Meta Charm vs. the AI Pin generation: a quick comparison
It is useful to place Charm against the previous generation of standalone AI wearables to see what Meta is trying to do differently. The table below compares publicly known details of Charm with Humane’s AI Pin, the most prominent cautionary tale in this category.
| Feature | Meta Charm (2026) | Humane AI Pin (2024) |
|---|---|---|
| Form factor | Pocket-sized pendant with screen | Clip-on wearable pin, no screen (laser projector) |
| Backing AI agent | Meta Muse | Cosmos (Humane’s proprietary assistant) |
| Parent company distribution | Billions of existing Instagram, Facebook, WhatsApp users | No existing consumer platform |
| Price at announcement | Not disclosed | $699 plus $24/month subscription |
| Outcome | Unproven, targeting 2026 holiday shipping | Discontinued after weak sales in 2024 |
The bigger picture: Meta’s app dependency problem
Strip away the pendant and the demo units, and Charm is best understood as Meta trying to solve a distribution problem it has had since the smartphone era began: it does not own the platform its products run on. Apple and Google set the rules for the App Store and Google Play, take a cut of in-app transactions, and can change policies that affect how Meta’s apps behave overnight. A successful Charm would give Meta a channel to Muse that neither Apple nor Google can gate, mirroring why Amazon built Kindle and Echo, and why Google eventually built Pixel despite already dominating Android software.
Whether that motivation is enough to make Charm a commercial success is a separate question from whether it is strategically rational for Meta to try. Plenty of strategically rational hardware bets have failed at the point of consumer adoption. The Business Insider and CNBC coverage of this announcement both land on roughly the same conclusion: Meta has won the announcement race, but the harder race, actual usage, actual sales, and actual retention, has not started yet.
Frequently asked questions
What is Meta Charm?
Charm is a pocket-sized AI device Meta unveiled at its Connect conference on September 23, 2026. It runs Meta’s Muse AI agent and is designed to be used primarily through voice interaction, with a built-in screen.
How much does Meta Charm cost?
Meta has not announced a price. CNBC reported that Zuckerberg revealed the device on stage without disclosing pricing.
When will Meta Charm be available?
Zuckerberg said the device is targeted to ship during the 2026 holiday shopping season, though only a small number of units had been built at the time of the announcement.
Is OpenAI releasing a competing AI device?
OpenAI has a consumer hardware effort underway, and reports indicate an announcement is expected before the end of 2026. No product name, price, or launch date has been confirmed publicly as of this article.
Does Meta Charm beat OpenAI’s device to market?
Meta beat OpenAI to the announcement, unveiling Charm roughly a week before OpenAI’s major developer event. Neither company has shipped a consumer AI device to the public yet, so it is not yet a race that has been won on sales.
What AI agent powers Meta Charm?
Charm runs Meta’s Muse AI agent, the same assistant Meta has been expanding across its app ecosystem and smart glasses throughout 2026.
Has a device like Charm been tried before?
Yes. Humane’s AI Pin and the Rabbit R1 both launched in 2024 as standalone AI wearables and struggled to find sustained consumer demand, with the AI Pin later discontinued. Reviewers are likely to compare Charm against that history.
Why does Meta want its own AI hardware instead of relying on phone apps?
Meta’s AI products currently reach most users through Apple’s iOS and Google’s Android, both of which set app store rules and take commissions. A dedicated device reduces Meta’s dependence on those platforms for reaching consumers with Muse.
Related Coverage
- Meta Muse Adds Avatars, Mac Control in 15 Days [2026]
- OpenAI Agents Probed Hugging Face 2 Months Early [2026]
- OpenAI Claims 100+ Math Problems Solved in 24 Days [2026]
- OpenAI Admits 6 AI Models Went Rogue: 2.15% Rate [2026]
- Amazon Blocks Meta Muse AI Agent After 12 Days [2026]
