Private Cloud Compute
Private Cloud Compute is a privacy-first cloud computing infrastructure developed by Apple for Apple Intelligence, designed to securely extend complex AI tasks to cloud processing. It runs on dedicated Apple Silicon servers and employs end-to-end request encryption with a no-data-retention policy after processing. A publicly auditable transparency mechanism ensures that server software cannot be used by Apple or third parties to access user data, extending the on-device privacy model into cloud computing environments.
Core Facts
Apple's self-developed Private Cloud Compute architecture uses Apple Silicon chips on the server side, runs within a secure enclave, and destroys data immediately after processing.
Apple Intelligence底层依托苹果自研的端侧小模型与Private Cloud Compute混合架构
Apple Intelligence relies on Apple's self-developed Neural Engine to run a small language model of approximately 3 billion parameters on-device for local tasks, while accessing larger-scale models via a private cloud computing architecture in the cloud.
Timeline (last 90 days)
Apple's Private Cloud Compute scheme claims that the raw audio is deleted after processing and only the summary is retained.
Siri AI is a rebuilt version of Siri within the Apple Intelligence framework that Apple announced at WWDC 2024, relying on on-device large models and Private Cloud Compute.
Private Cloud Compute places computation on dedicated secure servers when larger models are needed, but promises no storage and no traces.
For more complex requests, the system offloads tasks to Apple's cloud servers via Private Cloud Compute, and request data is not persisted or used for training.
Apple has pledged that Private Cloud Compute does not store user data.
Apple's Private Cloud Compute (PCC) architecture is used to handle requests that exceed local capabilities; the servers do not retain user data and use temporary encryption keys for each request.
Apple's AI strategy adopts a "device-first" technical approach, running a small language model of approximately 3 billion parameters on-device, and only processing via the cloud through Private Cloud Compute when task complexity exceeds local computing capacity.
苹果在iOS 18中推出的Apple Intelligence架构由设备端小型模型处理日常请求，超出本地能力的任务通过私密云计算转发至服务器且苹果承诺不留存用户数据
苹果的私密云计算（Private Cloud Compute）机制采用安全飞地与可验证透明度技术，防止OpenAI服务器留存用户数据
All Facts (13)
Apple's self-developed Private Cloud Compute architecture uses Apple Silicon chips on the server side, runs within a secure enclave, and destroys data immediately after processing.
90%VerifiedApple Intelligence底层依托苹果自研的端侧小模型与Private Cloud Compute混合架构
80%VerifiedApple Intelligence relies on Apple's self-developed Neural Engine to run a small language model of approximately 3 billion parameters on-device for local tasks, while accessing larger-scale models via a private cloud computing architecture in the cloud.
75%VerifiedApple's AI strategy adopts a "device-first" technical approach, running a small language model of approximately 3 billion parameters on-device, and only processing via the cloud through Private Cloud Compute when task complexity exceeds local computing capacity.
65%Unverified苹果在WWDC2024首次提出端云协同AI架构愿景
80%UnverifiedApple has pledged that Private Cloud Compute does not store user data.
60%UnverifiedApple's Private Cloud Compute scheme claims that the raw audio is deleted after processing and only the summary is retained.
50%UnverifiedSiri AI is a rebuilt version of Siri within the Apple Intelligence framework that Apple announced at WWDC 2024, relying on on-device large models and Private Cloud Compute.
50%UnverifiedPrivate Cloud Compute places computation on dedicated secure servers when larger models are needed, but promises no storage and no traces.
50%UnverifiedFor more complex requests, the system offloads tasks to Apple's cloud servers via Private Cloud Compute, and request data is not persisted or used for training.
50%UnverifiedApple's Private Cloud Compute (PCC) architecture is used to handle requests that exceed local capabilities; the servers do not retain user data and use temporary encryption keys for each request.
50%Unverified苹果在iOS 18中推出的Apple Intelligence架构由设备端小型模型处理日常请求，超出本地能力的任务通过私密云计算转发至服务器且苹果承诺不留存用户数据
50%Unverified苹果的私密云计算（Private Cloud Compute）机制采用安全飞地与可验证透明度技术，防止OpenAI服务器留存用户数据
50%