Meta’s 2026 makes one thing clear: the company has pivoted from the metaverse to AI-powered glasses, and it is willing to break with its own past to get there. In June, Meta launched a new line of in-house AI glasses starting at $299 — cheaper than the Ray-Ban Meta models — as part of a goal to sell 10 million wearables in the second half of the year. The bigger surprise is under the hood: most of Meta’s glasses now run on Muse Spark, a closed-weight model from the new Meta Superintelligence Labs, marking a retreat from the open-source Llama approach that built Meta’s AI reputation. Meanwhile, EU and other regulators are pressing Meta over WhatsApp rules that lock out rival AI chatbots.
For everyday users, the takeaway is that Meta is betting glasses are the next big AI device, and pricing aggressively to win. But its strategic shift to closed models and its regulatory fights are worth understanding before you buy in. Here is what actually changed.
Cheaper glasses, made in-house
The headline product is a new line of AI glasses that Meta designed itself, starting at $299 — undercutting the latest Ray-Ban Meta models, which begin around $379. They come in styles including Meta Adventurer and Meta Fury, plus a $399 collaboration with Kylie Jenner that offers her voice as the on-board AI assistant. Reporting suggests Meta plans up to four more models before year’s end, all in service of an ambitious target of selling 10 million wearables in the back half of 2026.
The strategy is straightforward: make AI glasses cheap and varied enough to reach mass adoption, and establish them as the everyday device for talking to an AI assistant. For Meta AI Review 2026: Is It a Good Brand?, this is the clearest statement yet that glasses — not VR headsets — are the priority. If you have followed Meta’s hardware, our Meta Quest Review 2026: Is It a Good Brand? profile covers the VR side that is now taking a back seat.
Muse Spark replaces Llama on the glasses
The most telling news is technical. As of mid-2026, the assistant on most Meta glasses is powered by Muse Spark, the first model from the newly formed Meta Superintelligence Labs. Muse Spark runs in three modes — Instant, Thinking, and Contemplating — that trade response speed for reasoning depth.
What makes this significant is that Muse Spark is closed-weight. Meta built its AI reputation on open-source Llama releases, which powered hundreds of third-party products and made the company the standard-bearer for open AI. Quietly swapping in a closed model for its flagship consumer device is a deliberate departure. It suggests Meta now sees frontier capability and product control as more important than the open-source goodwill that defined its earlier strategy. For where this sits in the wider field, see The AI Model Race 2026: OpenAI vs Google vs Anthropic and the The AI Directory.
Whether this is permanent or a two-track approach — open Llama for developers, closed Muse for products — remains to be seen. But it is a meaningful shift in posture from a company that spent years arguing open models were the future.
The antitrust pressure over WhatsApp
The other major thread is regulatory. In late 2025, Meta revised its WhatsApp Business terms to bar third-party AI providers from the platform when AI is their primary offering — a move that effectively locked out rivals like OpenAI and Anthropic. OpenAI’s ChatGPT, reportedly used by tens of millions of people on WhatsApp, notified users the service would be discontinued.
Regulators responded quickly. The European Commission opened a formal investigation in December 2025, Italy’s antitrust authority ordered Meta to halt the change, and other agencies across multiple regions began probing the exclusion. Separately, an Indian tribunal upheld a fine and a multi-year ban on Meta sharing WhatsApp user data with Instagram and Facebook for advertising.
For users, the practical effect is that convenient AI chatbots you may have used inside WhatsApp could disappear or change, depending on how these cases resolve. The broader significance is that Meta is now fighting to control who gets to build AI on its messaging platforms — and regulators are pushing back hard.
What it means for buyers
Three practical points. First, if you have wanted AI glasses, 2026 is the year they got cheaper and more varied; the $299 entry price makes them far more approachable. But treat first-generation, in-house hardware with normal caution — wait for independent reviews on battery life, comfort, camera quality, and how genuinely useful the AI is day to day.
Second, understand what you are buying into: Meta’s glasses now run a closed model tied to Meta’s ecosystem, so the assistant’s capabilities and privacy practices are Meta’s to define. If that matters to you, factor it in.
Third, if you rely on third-party AI chatbots inside WhatsApp, watch the antitrust cases — access could shift. For the wider consumer AI landscape and alternatives, our Best AI Chatbots 2026: ChatGPT vs Claude vs Gemini & More guide is the place to start.
The story to watch through the rest of 2026 is whether Meta actually hits its 10-million-wearable goal, whether Muse Spark proves good enough to justify the closed-model pivot, and how the WhatsApp cases reshape who can build AI on Meta’s platforms.
FAQ
How much do Meta’s new AI glasses cost?
Meta’s new in-house AI glasses start at $299, undercutting the Ray-Ban Meta models that begin around $379. A Kylie Jenner collaboration costs $399 and offers her voice as the AI assistant. Meta reportedly plans several more models before the end of 2026.
What is Muse Spark?
Muse Spark is the first AI model from Meta Superintelligence Labs and now powers most of Meta’s glasses. It runs in three modes — Instant, Thinking, and Contemplating — trading speed for reasoning depth. Notably, it is closed-weight, a departure from Meta’s open-source Llama approach.
Why did Meta move away from open-source Llama?
Meta built its AI reputation on open Llama models, so quietly using the closed-weight Muse Spark on its flagship glasses is a strategic shift. It suggests Meta now prioritizes frontier capability and product control over open-source goodwill, at least for its consumer devices. See Meta AI Review 2026: Is It a Good Brand?.
Why is Meta facing antitrust action over WhatsApp?
Meta revised its WhatsApp Business terms in late 2025 to bar third-party AI providers whose primary offering is AI, effectively locking out rivals like OpenAI and Anthropic. The EU opened a formal probe, Italy ordered Meta to halt the change, and other regulators followed.
Should I buy Meta’s AI glasses in 2026?
They are more affordable and varied than before, which makes them tempting. But as first-generation in-house hardware, wait for independent reviews on battery, comfort, camera, and real-world AI usefulness. Also weigh that the assistant runs a closed model tied to Meta’s ecosystem.
Can I still use ChatGPT on WhatsApp?
It became uncertain in 2026. Meta’s rule changes barred third-party AI providers whose main offering is AI, and OpenAI notified WhatsApp users that ChatGPT there would be discontinued. Ongoing antitrust cases could alter this, so access may change depending on how they resolve — see OpenAI Review 2026: Is It a Good Brand?.
Zen Tech Hub may earn a commission from links on this page, at no extra cost to you.
