The supplied reporting does not support a defensible count of startup categories absorbed during the last five years of Google I/O and WWDC. It does support the pattern worth testing: Google and Apple keep moving capabilities into devices, operating systems and default apps, reducing the value of products whose main advantage is convenient access to a single feature.
Google’s latest hardware announcement fits that pattern. The company introduced four Pixel 11 phones, the Pixel Watch 5 and Pixel Tag, while presenting the Tensor G6 around faster local AI processing. The important signal for founders is the distribution layer. A capability that runs locally, ships on new phones and connects to first-party hardware starts with advantages an independent app cannot buy.
How to count platform absorption without inflating the result
A useful five-year audit should begin with every Google I/O and WWDC keynote, then record each newly announced capability in a structured table. The columns should include year, platform, feature, previous user workaround, startup category affected, price before inclusion, geographic availability and whether the feature shipped.
That last field matters. Keynotes mix released products, previews and broad technical direction. Counting every demo as absorption would exaggerate the total. A category belongs in the final count only when users can perform the core job through an operating system, default app or first-party device without buying the independent product.
The standard should also exclude routine improvements. A better camera does not automatically absorb photo-editing startups. Native document scanning, transcription or object removal could qualify if the platform now completes the task that previously required a separate download or subscription.
Categories need consistent boundaries. “Productivity” is too broad to count. Password managers, transcription tools, package trackers, sleep trackers and Bluetooth item finders are narrow enough to compare across years. Otherwise, one analyst may count “AI assistants” once while another splits writing, summarization, image generation and search into four categories.
Without that coding sheet and the underlying keynote sources, publishing a precise total would create false confidence.
The categories carrying the most platform risk
The highest-risk products sit close to hardware sensors, private user data and frequent operating-system actions. Platform owners control the camera, microphone, location services, notifications, authentication system and background processing rules. They can connect those components with fewer permissions and fewer setup steps.
Local AI raises the pressure. Faster on-device processing can make transcription, image editing, message classification and personal assistance more useful without sending every interaction to a remote server. It can also reduce latency and support functions when connectivity is limited. The supplied announcement establishes only that Tensor G6 emphasizes faster local AI processing, so any specific Pixel 11 capability beyond that remains unconfirmed here.
The Pixel Tag also places item tracking squarely inside Google’s first-party hardware portfolio. That does not establish the commercial effect on independent trackers, but it gives buyers another option connected to the platform they already use.
The category most exposed next will probably combine three traits:
- The task happens several times a week.
- The platform already controls the required data or sensor.
- The current standalone product charges mainly for access to one narrow capability.
That profile deserves more attention than keynote vocabulary. “AI-powered” tells an investor little. Control of distribution, proprietary inputs and a workflow that survives native competition tell them much more.
Free inclusion changes the buying calculation
Platform absorption rarely erases an entire market overnight. It changes the default.
Once a usable version arrives with the phone or operating system, a startup must explain why someone should install another app, grant permissions, create an account and pay. The comparison shifts from “nothing versus this product” to “the included option versus this product.”
That creates room for specialists, but the basis of competition changes. A dedicated product can still win through deeper workflows, regulated use cases, cross-platform support, collaboration, audit trails or unusually strong accuracy. Convenience alone becomes fragile.
Founders should apply the same scrutiny to AI products. The Wrapper Audit offers a useful companion test: if the underlying model or operating system adds the feature, identify what remains defensible.
What to watch before the next keynote
Build a platform-risk register now, before keynote week turns into a stream of demos. List every dependency on Apple or Google, including permissions, sensors, default-app access, app-store rules and first-party models. Then mark which customer outcome each dependency enables.
Next, run a removal test. Assume the next operating-system release includes the product’s most visible feature at no additional charge. Write down what customers would still pay for. If the answer is vague, the roadmap needs work.
The strongest response may be a narrower market, deeper operational integration or ownership of data the platform cannot reproduce. It may also require support across iOS, Android, desktop systems and existing business tools.
When the next keynote ends, the useful question will have little to do with how polished the demo looked. Open the register, add the shipped capability, and check whether customers still have a reason to choose the separate product.
Sources
No source links were supplied with the event context.
Comments
No comments yet.