October 6, 2026, (Inside AI) — A 19-year-old has raised $11 million to build a $3,499 personal AI computer that runs AI agents locally with full privacy. The startup, Ghost, announced the Core on October 5, a screenless device designed to keep all data on-device. Preorders opened the same day, with shipping set for the last week of October.
The Core packs an NVIDIA RTX Pro 4000 SFF Blackwell GPU into a compact box manufactured at Ghost's San Francisco facility. It ships with three preinstalled AI models: Qwen-3.8-Next, Qwen-3.8-27B, and Gemma-4-31B. Users can also load additional models from Hugging Face. Since the device has no screen, users interact with it through a phone app or voice commands.
Zain Javaid, co-founder and CEO, described the Core as a "brain in a box" that continuously runs AI agents as personal assistants. These agents learn a user's context and proactively offer help. For example, the Core can analyze heart rate data alongside seasonal flu patterns and alert a user about potential illness before symptoms appear.
"The user owns all of it," Javaid said of the device, memory, and data. "Once you buy the hardware, everything is stored, run, and processed on the device... All of the source code, logic, and model weights live on the device, not the services. We build this device to not be dependent on having an outside entity maintaining the service."
Read: Meta Unveils Muse Charm AI Device at Connect 2026
Privacy sits at the center of the product's pitch. All processing happens on-device with encryption. A built-in firewall monitors every outbound request to prevent unauthorized data access. Users also control how much autonomy their AI agents have. If Ghost ever shuts down, the device continues to function independently.
That local-first approach sets it apart from cloud-dependent AI assistants like Alexa or Google Home. It also differs from a Mac Mini, which serves as a general-purpose computer. The Core combines hardware and software built specifically for running AI agents out of the box.
Read: Meta launches open-source Muse Gadgets for DIY AI hardware
At $3,499, the Core targets early adopters and privacy-conscious users willing to pay a premium for on-device AI. Whether it can compete against cheaper cloud subscriptions remains the central question as the product heads to market.