Google Pixel phones ship with AI capabilities that consume significant storage space—regardless of whether users actually employ them. A journalist documented how background AI services, including one app consuming nearly 7GB, left only 96GB usable from a stated 128GB. The situation highlights growing tension between on-device AI ambitions and storage constraints.
At a Glance
- AICore app reserves 6.73GB for on-device AI models
- Device Intelligence app takes 1.06GB
- Pixel Screenshots uses 1.14GB
- Pixel Studio occupies 193MB
- Total usable storage on 128GB Pixel 10 Pro fell to 96GB after setup
The Storage Reality
On a Pixel 10 Pro with 128GB advertised storage, system files, background apps, and preinstalled AI features left the user with approximately 96GB of usable space within two weeks of purchase. AICore, the primary on-device AI service, consumed 6.73GB alone—more than most users' largest applications.
Unlike traditional bloatware, these AI services cannot be disabled or uninstalled. They run continuously, updating AI models on-device and providing functionality to other apps. The result is a significant storage penalty that affects every user, regardless of whether they engage with AI features.
Cache Clearing Doesn't Help
Standard remedies for storage issues prove ineffective here. AICore stores no cache data, and clearing the app's storage causes immediate reconsumption. The app rebuilds itself, often faster when users employ AI-powered applications. This creates a frustrating loop where using AI features to manage storage actually accelerates storage loss.
The journalist noted that clearing AICore storage caused rapid reconsumption, with the app quickly returning to its 6.73GB footprint. Unlike conventional apps that can be disabled or deleted, AI infrastructure remains baked into the system.
Google's Current Workaround
Users can delete downloaded local AI model files through AICore's storage settings to reclaim some space. However, this compromises on-device AI functionality—requiring internet access for tools like Gemini that previously worked offline. The tradeoff means choosing between storage and capability.
Google's position appears to be that on-device AI processing justifies the storage cost. For users who rely on offline AI features, the trade-off may make sense. For those who rarely use these capabilities, the storage consumption feels mandatory regardless of benefit.
Why It Matters
This issue reveals a fundamental shift in how smartphone manufacturers approach feature delivery. Previously, unused features occupied minimal space. On-device AI changes the calculus entirely—these models consume gigabytes whether you invoke them once a day or never.
The practical implication: storage math that worked last year no longer applies. A 128GB phone that felt adequate in 2024 may leave users scrambling in 2026. The journalist found himself approaching storage limits less than three weeks after purchasing a new device, despite not installing intensive games, saving large media files, or running demanding AI tasks.
Who wins:** Google benefits most—on-device AI differentiates Pixel hardware while requiring no ongoing cloud infrastructure costs. Users who actively use AI features get faster, offline-capable assistance.
Who loses:** Budget-conscious buyers stuck with 128GB models. Users who prefer cloud-based AI pay storage costs for capabilities they don't use. Those who can't afford the $100 premium for 256GB face immediate storage anxiety.
Trade-offs:** Deleting local AI models via AICore's settings recovers space but forces cloud dependency. Connectivity becomes mandatory rather than optional.
What to Watch
Monitor whether Google adds granular toggle controls for AI features—Apple permits disabling some AI functions on iPhones. User pressure may force similar flexibility. Additionally, watch if industry-wide storage minimums shift upward as more manufacturers bundle local AI models.
Storage Recovery Options
Users with limited storage can:
- Delete local AI model files through AICore settings (sacrifices offline AI)
- Purchase higher-capacity models (256GB Pixel 10 costs $100 more)
- Use cloud storage for photos and music
- Uninstall non-essential apps
Practical Considerations
Power users requiring maximum storage should factor AI overhead into purchasing decisions. The 32GB consumed by AI infrastructure represents roughly one-quarter of a base model's capacity before installing any personal apps or media.
Key takeaways
- AI infrastructure on Pixel phones consumes significant permanent storage
- Standard storage management techniques don't work against AI app growth
- 128GB models may leave insufficient usable space for power users
- Cloud-dependent AI offers an alternative at the cost of offline functionality
Frequently asked questions
Can I delete AICore on my Pixel?
AICore cannot be disabled or uninstalled. You can clear downloaded AI models through its settings, but this disables on-device AI features.
How much storage does AICore use?
On recent Pixel phones, AICore consumes approximately 6.73GB for storing on-device AI models.
Does clearing AICore's storage free up space permanently?
No. The app rebuilds its storage immediately, often faster when AI features are used.
What's the solution for storage-constrained users?
Options include purchasing higher storage models, using cloud-based AI services instead of on-device processing, or choosing devices with more flexible storage management.
Do other Android manufacturers have similar issues?
On-device AI features are becoming common across manufacturers, though storage consumption varies based on implementation and which AI services are preinstalled.