Run mobile apps in the cloud
Connect business workflows
with AI
Run mobile applications in managed cloud Android environments. Use the PaaS platform to organize instances and connect business systems; integrate customer or partner AI models when a workflow calls for them.
Task understanding · screen recognition · action planning
Cloud phones provide the runtime; AI systems connect as needed
One cloud foundation, six AI business scenarios
Connect AI systems to managed Android environments for mobile workflows. The model, app adapter, operating rules and delivery scope are assessed for each use case.
Separate cloud workspaces for global content teams
Cross-border commerce, brand and social teams can organize app environments by store or account. An AI system may assist with multilingual drafts and planning; people review content and publish under each platform’s rules.
- AI system
- Drafts, information organization and publishing plans
- Cloud phone
- Separate app workspaces and mobile preview
- Workflow
- Human review before permitted publishing
Prepare multilingual content for an overseas store and create a review-ready publishing plan
- 01Collect source material and target languages
- 02Draft content and organize a schedule
- 03Preview in the app, review and publish
Coordinate multilingual replies with your support workflow
Connect authorized messaging workflows with an existing AI service to classify enquiries, retrieve approved knowledge and prepare reply drafts. Define when a support agent should review or take over.
- AI system
- Language understanding and reply suggestions
- Cloud phone
- Messaging apps and support workspace
- Workflow
- Review, reply or transfer to a person
Summarize an overseas customer enquiry and prepare a reply draft
- 01Receive an authorized customer message
- 02Retrieve knowledge and draft a reply
- 03Review the answer or route to an agent
Explore natural-language testing in repeatable environments
App development and QA teams can evaluate AI-assisted test planning and issue summaries across cloud Android environments. Review screenshots, logs and pass criteria to verify and reproduce each finding.
- AI system
- Test planning and issue analysis
- Cloud phone
- Versioned environments and test runs
- Workflow
- Check evidence against explicit criteria
Verify sign-in and form submission in a new app release
- 01Prepare the test image and scenario
- 02Run the target app workflow
- 03Review findings and reproduction notes
Reduce repetitive work across mobile business apps
For authorized internal workflows, assess information extraction, form preparation and cross-app handling. Keep human confirmation for submissions and other consequential steps where needed.
- AI system
- Extract fields and prepare structured drafts
- Cloud phone
- Mobile business apps and interaction channel
- Workflow
- Review before submission or decision
Collect order details and prepare a mobile form for review
- 01Read authorized source information
- 02Match fields and prepare a draft
- 03Review and confirm the submission
Run games in the cloud and explore AI-assisted experiences
Game developers and cloud service teams can explore screen understanding, contextual tips and authorized repetitive test tasks in cloud-hosted environments. Validate the app and interaction path for each title.
- AI system
- Screen analysis and contextual assistance
- Cloud phone
- Android runtime and media channel
- Workflow
- Evaluate in an authorized app scenario
Analyze an in-game screen and provide contextual guidance
- 01Launch the selected game environment
- 02Capture a screen for analysis
- 03Review assistance or test observations
Give mobile agents a resettable Android environment
Agent developers, research teams and educators can use managed instances to evaluate app interaction, inspect task traces and connect their own orchestration or training stack.
- AI system
- Customer or partner model and agent
- Cloud phone
- Android instances, lifecycle and available APIs
- Workflow
- Evaluate tasks and inspect outcomes
Run a defined app task and review the agent’s interaction trace
- 01Prepare an Android environment
- 02Connect the agent and run a task
- 03Inspect the result and reset for another run
What we actually build
Our capability follows one chain: compute, runtime, platform, delivery. We do not sell a single product line; we provide combinable infrastructure capability.
Compute engineering
Hardware selection, board adaptation, power and thermal design, and resource pooling around ARM SoC arrays.
Android virtualization
Custom AOSP images and cloud runtime supporting instance creation, image delivery, lifecycle and monitoring.
PaaS engineering
Tenancy, authorisation, metering, billing, device management, operations console and an OpenAPI gateway.
Resource orchestration
Scheduling and elastic scaling based on compute, storage, network and instance load.
Real-time media
Capture, encoding, transport and weak-network optimisation for remote control, cloud gaming and interaction.
Project delivery
Solution design, environment adaptation, deployment, integration, acceptance and operations handover.
Three PaaS product lines
One platform kernel, three delivery forms: domestic public cloud, global public cloud and private deployment.
Domestic PaaS Public Cloud
A cloud Android and compute platform for China-based customers, with on-demand instances, multi-tenant management, image operations, billing and OpenAPI access.
- On-demand provisioning
- Tenant isolation and quotas
- OpenAPI integration
Global PaaS Public Cloud
Multi-region service for cross-border and overseas operations, designed around network latency, data boundaries and local operations.
- Regional node planning
- Cross-border scenarios
- Unified console and APIs
PaaS Private Deployment
Deploy the platform in your own data center or dedicated cloud. Resources stay under your control and data does not leave your boundary.
- Dedicated resource pool
- Data stays on premises
- Localized stack support
Four-layer technology architecture
A complete layered system from ARM silicon to application services. Layers are decoupled through stable interfaces and support public, private and hybrid deployment.
This diagram explains the technical structure and component boundaries. Actual implementation and delivery scope are defined in the agreed solution.
A high-density compute base built on ARM
For workloads that live inside mobile operating environments, ARM avoids the overhead of cross-architecture translation. We organise compute, storage and network resources around SoC arrays to provide a stable base for cloud Android instances.
- Matched to the workload
Instruction-set alignment reduces translation overhead for mobile application workloads.
- Density and scalability
Scale out by array unit and invest in batches as business grows.
- One resource pool
Compute, storage and network are managed centrally and allocated by tenant and instance.
Regional node planning for cross-border business
With both domestic and global product forms, a business can choose the nearest access and the appropriate compliance boundary. Regional build-out follows actual demand, in phases.
- Regional deployment
Select nodes close to your users to reduce cross-region latency.
- Clear compliance boundaries
Domestic and global resources are organised separately.
- One platform experience
Shared console and APIs reduce multi-region operational cost.
How solutions are structured
The cases below outline typical solution structures and delivery approaches for each scenario; the implementation scope is defined during project assessment.
Unified runtime platform for public-sector mobile apps
A private deployment of cloud Android runtime providing centralised hosting, operations and access control for government mobile applications.
Group-level multi-account operations platform
Centralised management of cloud device clusters, accounts, images and operational data across regions for a group organisation.
Cloud-delivered lab environments
Mobile lab environments hosted in the cloud and allocated by class and course, reducing endpoint maintenance and improving consistency.
No cases in this category.
News and insights
Company updates, technical notes and industry observations.
Website redesign: leading with architecture, not product lists
The new site is organised around ARM compute, Android virtualization and the PaaS platform instead of a product-catalogue structure.
Read moreImage layering and startup optimisation for cloud Android instances
How image layering, runtime initialisation and resource warm-up shorten instance delivery time while keeping environments consistent.
Read moreThe changing role of ARM in cloud infrastructure
As mobile and edge workloads grow, ARM compute is moving from a supplement to a structured infrastructure choice.
Read moreStart with a technical conversation
Whether you are evaluating a cloud phone PaaS, planning private deployment, or looking for an infrastructure partner, we can start with a pragmatic discussion.