Three days at AWE 2025 in Long Beach. Sixteenth edition of the conference. Six thousand attendees, two hundred and fifty exhibitors, four hundred and fifty speakers. The official theme was the AI+XR imperative. The press releases declared XR’s mainstream moment. They’ve been declaring it every year since 2010.

What I actually saw, sorted by what matters for the kind of work I get hired to do — flagship immersive projects for cultural institutions, brand activations, conference producers, and entertainment clients.

What was real

Display quality crossed a threshold. Samsung Display’s 1.4-inch OLEDoS panel hit 5,000 pixels per inch on a screen the size of a watch face. Sony’s XYN headset prototype and the Play for Dream MR — Android-powered, 8K micro-OLED, $1,999 versus Apple Vision Pro’s $3,499 — both showed that the visual fidelity gap between consumer-grade XR and theatrical projection is closing fast. For studios designing for headset-based experiences, this matters within the next eighteen months. For studios designing live conferences and built environments, less.

Snap is back in the conversation. Evan Spiegel’s keynote — eleven years and three billion dollars into AR glasses development — set up the fifth-generation Spectacles for a 2026 release. They’re integrating OpenAI and Gemini directly into Snap OS. Snap is the first company in years to articulate a vision of AR glasses that isn’t a phone-replacement story; it’s a “look up instead of down” thesis. Worth tracking.

Spatial intelligence became a real category. The Metaverse Standards Forum’s session on the convergence of spatial AI — 3D reconstruction, scene understanding, SLAM — and immersive platforms wasn’t vapor. AI agents that can perceive and reason about 3D space will, in a few years, change how interactive environments get authored. Today, mostly research. Worth watching closely.

What was theater

The headline narrative — that AI and XR are converging into a single mainstream product category — is true at the platform level and mostly false at the live-events level I work in.

The conference floor was full of enterprise-pilot demos: Siemens and Sony’s immersive engineering tool for designing race bikes in 3D CAD. FundamentalVR’s surgical training platform. Virnect’s industrial AR safety goggles with embedded LLMs. These are real products doing real work. They are also, to a one, B2B SaaS workflows. They have nothing to do with the question a transformational-event producer is actually asking, which is: how do I make eight hundred people in a hotel ballroom feel something they’ll talk about for a year?

Most of what gets demoed at AWE is a step on the road to a future where every person walks around in a pair of AI-powered AR glasses. That future will arrive. It is not arriving in time for the Q1 2026 leadership offsite my clients are planning. For that event, the answer is still a well-orchestrated combination of projection mapping, sound design, lighting cues, ritual structure, and a story.

What I’m taking back

When a client asks about putting Apple Vision Pros in their event, the answer is still no, but here’s what would work. Headset-based experiences at live events haven’t solved the throughput problem. Six minutes to fit and onboard each guest. Fifteen guests through per hour. Against a thousand-person audience. The math hasn’t changed since 2018.

Real-time AI image generation is becoming useful for content development and rapid prototyping, not for live deployment. I’m using it now to mock up environments faster than I could before. I am not yet using it in front of guests, because the failure modes — hallucinated typography, off-brand color drift, uncanny faces — are still publication-stopping.

The interesting story at AWE was not on the floor. It was in the side conversations. Conference producers, brand strategists, museum directors — the people who actually commission immersive work — are converging on a different language than the platform companies. Language about transformation, ritual, threshold. That conversation is the one that matters for the next decade of this work.

What I’m watching now

Snap Spectacles Gen 5 in 2026 — the first AR glasses with a credible developer ecosystem story. The Sphere/Cosm playbook scaling down — what happens when 360-degree LED domes drop below the $5M install threshold. Real-time AI for live show control — lighting, sound, and video cues that adapt to audience input.

What I’m not watching: the metaverse, in any of its 2022-vintage definitions. That conversation is over.

Alt Ethos works with a small number of producers each year on flagship immersive projects. To start a conversation: [CTA link]