What 8.5 Million Developers Need to Understand About Product Video

Field Notes · Google I/O 2026
Google I/O 2026 · Mountain View, CA · May 19–20

What 8.5 Million Developers Need to Understand About Product Video

We were at Google I/O 2026. The agentic era is here. And the video problem most app developers ignore just got significantly harder to ignore.

By Advids
May 26, 2026
3,100 words · 12 min read
Participant · Attendee
8.5M
Developers building with Google models monthly
3.2Q
Tokens processed per month (7× growth since I/O 2025)
85+
Sessions & codelabs on demand at io.google
100
Announcements across two days
Google CEO Sundar Pichai opened Google I/O 2026 with a line that will be quoted for years: "We are firmly in our agentic Gemini era." It wasn't a product announcement. It was a repositioning statement. And for the 8.5 million developers now building with Google's models every month, it described a shift in what their apps do — and simultaneously created the most consequential video problem most of them have never thought about.

We were at Shoreline Amphitheatre in Mountain View on May 19th. Two days, 100 announcements, a new model family, a 24/7 AI agent, and smart glasses with Gentle Monster branding. The energy at Google I/O 2026 was different from prior years — less "here are features" and more "here is the new shape of computing." Sundar framed it clearly: ten years since Google pivoted to AI-first, and the company is now repositioning from an operating system business to an intelligence system business.

For Advids — a video production company that works almost entirely with app developers, SaaS companies, and technology platforms — two days at Google I/O felt like a clinical diagnosis. The ecosystem is producing apps and platforms of increasing complexity and capability. The gap between what those products actually do and what their marketing video communicates is widening at exactly the same pace.

This is what we observed, what it means, and what to do about it.

Section 01
The Event — What Actually Happened at Google I/O 2026
Google I/O 2026 ran May 19–20 at Shoreline Amphitheatre, Mountain View, California — the same venue Google has used as its developer conference home for over a decade. The opening keynote began at 10 a.m. PT on May 19, streamed live on YouTube and at io.google, followed by a developer keynote and two days of technical sessions, codelabs, and Q&A stations. Over 85 sessions, codelabs, and presentations are now available on demand at io.google for anyone who missed the live event.

From Sundar Pichai's Keynote — Confirmed Data
Two years ago, Google was processing 9.7 trillion tokens per month. At I/O 2025, that had grown to roughly 480 trillion. At I/O 2026, Pichai announced the number jumped 7x to over 3.2 quadrillion tokens per month. The implication: developers and enterprises are building at a scale that makes previous cycles look like prototypes.

The structural theme of the event — what Google called the "agentic Gemini era" — played out across every major announcement. This was not a refinement year. Gemini 3.5 Flash launched as the headline model, described by Google as delivering frontier intelligence at 4x the speed of competing models on output tokens per second. Gemini Omni arrived as a new model series combining reasoning with creation — able to accept any input and generate any output, starting with video. Gemini Spark became the consumer-facing 24/7 personal AI agent, built on Gemini 3.5 and the Antigravity platform, designed to work autonomously in the background even when a user's device is off.

Google Antigravity 2.0 was the developer story of the conference: an agent-first development platform with a new desktop app, CLI, and SDK — giving developers the infrastructure to build, orchestrate, and deploy multi-agent systems at production scale. Android 17 previewed its Material 3 Expressive design language. Samsung's Intelligent Eyewear — Android XR smart glasses produced in partnership with Gentle Monster and Warby Parker — were announced for a fall 2026 launch. WebMCP introduced a new open web standard for browser-based AI agent interaction. And Veo 3, Google's video generation model, became available in the Gemini API at prices that the developer community immediately noted were usable at startup scale.

It was, by any measure, a significant event for the developer ecosystem.

Section 02
Three Things We Observed at Google I/O 2026 That Nobody's Writing About
01
The Agentic Gap: Apps That Do Things Are Harder to Show Than Apps That Display Things
The central thesis of Google I/O 2026 — that AI is transitioning from assistive to agentic — creates a new product marketing problem that almost nobody at the conference was talking about. When Sundar Pichai demonstrated Gemini Spark, the 24/7 personal AI agent that works in the background even when your phone is off, the demo was essentially: "here is a thing happening that you can't see." That's a beautiful product experience. It's a nightmare to put on video.

Traditional product demo videos are screen-recording-plus-voiceover: user clicks here, thing happens there. The agentic paradigm inverts this entirely. Gemini Spark schedules your email while you sleep. Antigravity 2.0 orchestrates multiple subagents to compress multi-day engineering tasks into hours. These aren't UI interactions — they're outcomes that materialize without visible user input. Showing what an agentic product does requires animation, metaphor, and narrative sophistication that a screen recording cannot provide. The developers building on these platforms are going to need video production partners who understand this shift — and almost none of them have thought about it yet.

02
Veo 3 in the API Does Not Solve the Product Marketing Video Problem — It Creates a New One
Veo 3 becoming available in the Gemini API was one of the most discussed technical announcements at I/O 2026. At $0.05 per second for the Lite tier and $0.40 per second for Standard (with native synchronized audio generation), it crossed a pricing threshold that makes AI video generation viable for developer-scale content production. The developer community reaction was immediate and accurate: product explainer videos from feature specs, app onboarding animations without a design team, educational content at scale.

Here is the problem we observed in the conversations around this announcement: developers were conflating content generation with product marketing. Veo 3 is excellent for producing supporting content at volume — social posts, quick demos, internal documentation clips. It is not a replacement for the strategic product launch video that drives app store conversion, investor confidence, or enterprise sales cycles. Those videos require brand consistency, accurate UI representation, tested narrative structure, and the creative judgment to decide what story is actually worth telling. The companies that outsource their flagship product video to text-to-video generators in 2026 will look like the companies that outsourced their website copy to the first wave of GPT tools in 2023 — discoverable, and not in a good way.

03
The Android 17 Design Language Change Means Every Existing App Marketing Video Is Aging Out
Material 3 Expressive — Android 17's new design language, previewed at I/O 2026 — is not a minor refinement. It's a visual identity reset. The fluid animations, vibrant colors, and new typography introduced in the Gemini app's "Neural Expressive" redesign are the same language Google is rolling into the OS itself. For app developers, this creates a compounding problem: existing product demo videos were shot against a visual language that is being deprecated.

We've seen this cycle before. When Google moved from Material Design to Material You in Android 12, app marketing videos that featured the old design language began looking dated within six months of the OS rollout. Material 3 Expressive is a more significant visual shift — the expressive color, motion, and typography changes are visible at a glance in a way that subtler updates aren't. Developers who haven't refreshed their product video since 2024 or early 2025 are, right now, producing acquisition funnels with visually outdated assets. The window to update before the gap becomes glaring is roughly the next two quarters.

Section 03
What Advids Was Doing at Google I/O 2026
We attended Google I/O 2026 as participants and observers — embedded in the developer community that makes up a large portion of Advids' client base. Our team attended sessions across the Android, Gemini API, Antigravity, and web development tracks. We were in the developer area for the codelabs, in the main amphitheatre for both keynotes, and in the conversations around the edges of the formal program where the most useful signal actually lives.

What we were paying attention to was not the model benchmarks or the API pricing. We were listening for the moments when a developer or product team described what their app does and the person they were talking to didn't understand it. Those moments happened constantly, and in every one of them, the gap was the same: the developer had a precise technical understanding of their product's capability, and zero vocabulary to communicate what that meant for a non-technical user or buyer.

The most common question our team fielded during the event — from developers who asked what Advids does — was some version of: "So you make explainer videos? Like, for things that are hard to explain?" Yes. Exactly that. The apps being built on Gemini 3.5, Antigravity, and Android XR are not easy to explain. They are invisible by design. They are agentic, ambient, and asynchronous. And they need video that can communicate what they actually do — not what they look like while doing it.

Observation from the Floor
During a session on Antigravity 2.0 subagent orchestration, a developer in the audience asked the Google engineer on stage: "How would you explain this to a non-technical customer?" The engineer paused, started three sentences, abandoned all of them, and said: "Honestly, I'd show them a demo." The audience laughed — because every developer in the room recognized the problem. The demo wasn't going to help either. Showing a terminal running subagents doesn't communicate value to a buyer who needs to understand outcomes.

Section 04
The Product Video Problem for App Developers: A Deep Dive
The core challenge for app developers building on Google's AI ecosystem is not unique to 2026 — but the agentic shift makes it significantly more acute. The problem is this: software products create value through interactions that are either invisible, rapid, or deeply contextual. Video is the best medium for communicating how products work. But most app developers produce videos that show what the product looks like — not what it does, who it helps, or why it matters.

Frequently Asked Questions — App Developer Video Production
How do app developers create compelling product demo videos without a production team?
App developers need three things: a screen-capture-plus-animation layer that makes UI interactions readable at scale, a script that leads with user outcome rather than feature description, and professional voiceover that bridges the technical-to-consumer language gap. Advids produces dedicated app walkthrough videos for mobile and web developers that combine enhanced screen recordings, UI animation overlays, and outcome-first scripting — typically delivered in two to three weeks from brief to final file.
What kind of video works best for Android and iOS app launches in 2026?
For app store listings, a 30–60 second product demo video showing the core user flow — no narration, clear captions, single compelling outcome demonstrated — consistently outperforms longer, feature-heavy videos. For developer marketing and B2B app launches, a 90–120 second explainer that leads with the problem, shows the solution in action, and closes with a specific call to action is the format Advids recommends. For agentic apps built on platforms like Gemini Spark or Antigravity, outcome animation (showing the result rather than the process) is the emerging format of choice.
What does the shift to agentic AI at Google I/O 2026 mean for app marketing video?
Agentic AI — as demonstrated by Gemini Spark, Antigravity 2.0, and the broader shift Sundar Pichai described in his keynote — means apps increasingly do things on behalf of users rather than waiting to be used. This creates a new video challenge: showing what an app does autonomously is harder than showing what a user clicks. Product videos for agentic apps need to demonstrate outcomes achieved without visible interaction — a format requiring more sophisticated animation, scripting, and visual storytelling than traditional UI walkthroughs.
Does Veo 3 / AI video generation replace professional video production for app developers?
Not for product marketing video. AI-generated video tools like Veo 3 are useful for rapid content generation — social fillers, quick demos, test content. App marketing video that drives conversion requires brand consistency, accurate UI representation, precise scripting, and narrative structure that current AI video tools cannot reliably produce. Advids uses AI tools as part of our production workflow while maintaining the human creative direction and quality control that high-stakes app launches require.
The Three Video Formats That Actually Work for App Developers
Based on the products we observed being built and launched at Google I/O 2026 — and drawing on Advids' production work for apps across Android, iOS, and web — these are the three video formats that reliably drive adoption for developer-built products:

Format Best For Length Key Requirement
Outcome Animation Agentic apps, AI-native features, invisible automation 60–90 sec Show the result state, not the process. No UI recording needed.
UI Walkthrough Video Traditional apps, SaaS tools, onboarding flows 90–120 sec Enhanced screen capture + animation overlay. Outcome-first script.
App Store Demo Consumer apps, Google Play feature listing, social ads 15–30 sec No narration. Single-core-value-proposition. Material 3 visual language.
Section 05
5 Video Insights from Google I/O 2026 — Field Notes
01
Agentic products need outcome-first video, not UI-first video. Every agentic capability shown at I/O 2026 — Spark, Antigravity subagents, Managed Agents API — does its most valuable work invisibly. The video format that communicates this effectively is outcome animation: show the world before the agent ran, then show the world after. The process is irrelevant to the non-technical buyer. The result is everything.
02
The developer-to-consumer translation gap has never been wider. Sundar Pichai announced 8.5 million developers building with Google models monthly. That number means the supply of AI-capable apps is growing faster than the marketing vocabulary to explain them. The developers who invest in a well-crafted product video in 2026 will have a structural advantage in a sea of technically superior but communicatively invisible products.
03
Material 3 Expressive is a video refresh trigger. Android 17's new design language — fluid, vibrant, emotionally expressive — makes existing product videos look dated faster than a typical OS update. If your app's marketing video was produced before Google I/O 2026, run a visual consistency check now. The gap between your video and the OS aesthetic will widen every month until you update.
04
Android XR video formats don't exist yet — and the market belongs to whoever creates them. Samsung's Intelligent Eyewear launching this fall with Android XR and Gemini Intelligence embedded means there will be, by end of 2026, a new device category that needs product explainer videos. Nobody has established what the right format looks like for spatial computing apps. The first production companies and developer marketers to define this format will own it.
05
The VC pitch video is becoming standard practice for AI apps. Across conversations at I/O 2026, multiple early-stage teams mentioned using video assets in investor outreach — 60–90 second "visual pitch" videos sent before meetings to give investors context. This is the same pattern we've seen at SaaStr and in the broader SaaS ecosystem. AI app founders are adopting it now. A well-produced investor pitch video that demonstrates what your Gemini-powered app actually does — without requiring a live demo — is becoming a table-stakes fundraising asset.
Section 06
What Advids Produces for App Developers and Google Ecosystem Builders
Advids works with app developers across Android, iOS, and web — from early-stage teams launching their first product to established platforms refreshing for a new market. The video work we do for developers in the Google ecosystem falls into four categories:

Service What It Delivers
App Launch Videos App store-optimized demo videos (15–60 sec) for Google Play and web. Outcome-first scripting, Material 3-compatible visual language, no narration or with professional voiceover options. Produced from screen recordings plus custom UI animation.
Product Explainer Videos 90–180 second explainer videos for developer tools, SaaS applications, and AI-native platforms. Specialised scripting for products that are difficult to explain — particularly agentic, ambient, and background-operating systems. Animation-led visual storytelling for invisible processes.
Developer Marketing Video Developer-facing content for teams building on Gemini API, Antigravity, Firebase, or Android Studio. Explains technical capabilities in the language of developer outcomes — what you can build, how fast, and what it enables. Useful for developer relations, conference talks, and technical blog accompaniment.
Investor & Fundraising Video 60–90 second investor pitch videos for AI app founders. Designed to communicate platform capability, market opportunity, and technical differentiation to non-technical VC audiences — without requiring a live demo. Increasingly used as a pre-meeting asset in seed and Series A fundraising sequences.
On Veo 3 and Our Production Approach
Yes, we've been building with Veo 3 and other AI video generation tools since they became available. We use them for rapid ideation, B-roll generation, and accelerating specific production stages. We do not use them as the primary output for client-facing product videos, because AI video generation currently cannot produce brand-consistent, UI-accurate, narratively precise content at the quality level that app launch and investor video requires. Our position: AI tools lower production cost and increase speed for supporting content. Human creative direction remains the quality driver for strategic video.

Section 07
The Forward Look: What Comes Next After Google I/O 2026
Sundar Pichai said at I/O 2026 that "we're now in the part of the AI cycle where people want to see the value in the products they use every day." That is a statement about marketing as much as product. Capability is no longer the differentiator — communication is. The companies that will win in the Gemini era are not necessarily those with the most sophisticated models. They are those who can make what their product does legible, desirable, and trustworthy to a buyer who has seen a hundred AI pitches this year.

The next major event for this developer community is Automate 2026 in Chicago (June 22–25), where AI-native workflow automation — built on exactly the kind of agentic infrastructure Google announced at I/O — will be the dominant theme for industrial and enterprise buyers. The pattern repeats: technically extraordinary products, underdeveloped communication assets, and a widening gap between what's been built and what buyers understand.

The 8.5 million developers now building with Google's models monthly represent a generation of products that will need to explain themselves. Some will do it with text. Most will try to do it with live demos. The ones that break through will do it with video that was purpose-built to communicate invisible value — clearly, quickly, and memorably.

If you're building on Gemini, Android, Antigravity, or any of the platforms announced at Google I/O 2026, the video problem is real. We solve it.

Building an App on the Google Ecosystem?
Talk to the Advids team about app launch video, product explainer production, or developer marketing content. We've worked with app developers from pre-launch through Series B.