A 19-Year-Old Founder’s Big Bet on Personal AI

Ghost, a startup building a dedicated computer for personal AI agents, has emerged from stealth with $11 million in seed funding. The round was led by Andreessen Horowitz, with participation from Abstract, Audacious Ventures, SV Angel, and Nova. The company is led by 19-year-old founder Zain Javaid, who also serves as CEO. Its founding team includes Nicholas Chua, 19; Yifei Chen, 24; and Gautam Sharda, 23. After a year of building, Ghost is introducing Core, a $3,499 desktop computer designed to run AI agents locally and continuously.

The announcement places Ghost at the center of a growing race to define what personal AI hardware should look like. Rather than treating AI as a feature inside a phone, laptop, or cloud subscription, Ghost is betting that people will want a dedicated machine that remembers their context, acts on their behalf, and keeps sensitive data under their control. Core is not a general-purpose PC in the traditional sense. It is a screenless “brain in a box” that users access through a phone or app, with voice mode available through the companion app rather than the box itself.

From Quant Aspirations to AI Hardware

Javaid’s path to founding Ghost began with an interest in quantitative finance. For much of his life, he wanted to become a quant who uses mathematics and data to guide investment decisions. He spent time in that world, but eventually found it boring. Then, in 2022, ChatGPT arrived. The experience, he said, made him feel fundamentally excited in a way he never had before. That excitement pulled him toward personal AI and the question of how AI could become a persistent, useful part of everyday life rather than a chatbot people visit occasionally.

That question became Ghost. The company spent a year in stealth developing both hardware and software before revealing its first product. The seed round gives it capital to manufacture, ship, and support Core while continuing to develop the software layer that makes the device useful. The investor list suggests that some prominent venture firms see potential in a local-first approach to AI agents, even as cloud-based AI services dominate headlines.

What Core Actually Is

Core is a personal computer designed specifically for AI agents that can take actions on a person’s behalf. According to Javaid, the idea is to centralize a person’s data—from desktop files and applications to smart home devices—into one piece of hardware. That hardware is built to run AI agents continuously, allowing them to act as an assistant, perform tasks, and understand a person’s life over time. “You can run models on existing hardware, but it’s a terrible experience on many levels,” Javaid said, explaining why Ghost chose to build a dedicated machine rather than rely on off-the-shelf computers.

The device includes an Nvidia RTX Pro 4000 SFF Blackwell GPU, a type of chip commonly used to run AI models. The GPU is bundled into the device’s price, so customers do not need to buy one separately. Core also has its own software layer, browser, and filing system. It can continuously process information and tasks without constant prompting. The price is $3,499. Preorders for the first batch open Monday, and shipping is scheduled for the last week of October.

Core is described as a “brain in a box” without a screen. Users access it through a phone or app, and voice mode is available through the Core app. The device comes with three models preinstalled: Qwen-3.8-Next, Qwen-3.8-27B, and Gemma-4-31B. Users can also install other models from Hugging Face, where developers share AI models, or load their own trained models. That flexibility matters for enthusiasts, developers, and privacy-focused users who want to choose which model powers their agent.

A Purpose-Built Machine vs. a General-Purpose PC

Asked how Core differs from a Mac Mini—Apple’s small desktop computer that some people set up to run AI tools—Javaid said Core’s hardware and software were built specifically for running AI agents, with models already installed. The Mac Mini, by contrast, is a general-purpose computer that can be configured for AI. The distinction is central to Ghost’s pitch. A general-purpose computer asks users to assemble the right software stack, manage dependencies, and configure models. Core aims to arrive as a fully functioning computer that does not require complex setup or configuration.

Ghost also plans to push over-the-air updates, similar to how Tesla updates its vehicles. As new models or improvements to existing models come out, the company can send updates to the device. If a new generation of models arrives, Ghost asks users whether they want to switch. Users can also turn on auto-updates so they are always running the latest generation of a given model. That approach blends the convenience of a managed appliance with the flexibility of local model choice.

Local Processing, Encryption, and User Ownership

Privacy and ownership are central to Ghost’s message. Javaid said the models and personal memory run locally, and all personal information is encrypted so that no cloud provider or third party—including Ghost—can meaningfully access it. Users can control how much autonomy they give their Core agents. The device is also programmed to keep working if Ghost shuts down. “All of the source code, logic, and model weights live on the device, not the services,” Javaid said. “We build this device to not be dependent on having an outside entity maintaining the service.”

That independence is unusual in consumer AI. Many AI products rely on cloud servers to process requests, store memory, and run large models. Local processing can improve privacy and reduce latency, but it also requires capable hardware, efficient models, and careful software design. Core’s inclusion of a dedicated GPU and preinstalled models is meant to address the hardware and setup hurdles. The promise is that the device can act as a long-running agent without sending personal data to a remote server.

The company is also building a firewall to limit how the agent can use its credentials, such as logins and passwords. According to Javaid, the firewall monitors every outgoing network request on the computer. If it sees the agent performing an unsanctioned action or exfiltrating data it should not, it rejects the request and flags it for user review. “The user owns all of it,” he said of the device, memory, and data. “Once you buy the hardware, everything is stored, run, and processed on the device.”

Competitive Landscape and Incumbent Confidence

Javaid said he is not too worried about competition, including from established tech companies. “People sometimes give incumbents credit for having already built the product because they have the ingredients,” he said. He credited Apple with good hardware and disruption, but said much of what Core is aiming to tackle is software. The model must reason over a person’s life, remember context, take action, and know when to reach out. That is a software and systems challenge as much as a hardware one.

The broader AI hardware category has seen a wave of experiments, from wearable pins and pocket companions to dedicated AI computers. Many products have struggled to justify their existence when smartphones already run AI apps and cloud services keep improving. Ghost’s bet is different: it is not trying to replace a phone or become a wearable. It is trying to become the always-on, local brain that coordinates a person’s digital life. At $3,499, Core is priced like a premium workstation rather than an impulse gadget. That price reflects the Nvidia GPU, local storage, and the cost of building a specialized device in small volumes.

An Example of Proactive Assistance

Javaid offered an example of how Core is meant to work. Core might proactively send a notification that a user could be getting sick because it sees their resting heart rate is elevated, sees from computer history that they were up late working on a project, and has seasonal context that it is flu season in San Francisco. That scenario combines biometric data, desktop activity, and environmental context. It also shows why local processing and persistent memory matter. A cloud assistant could theoretically do some of this, but a local device can keep sensitive health and activity data on the machine while still generating useful inferences.

The company currently builds the units at its San Francisco facility. Javaid said the goal is to sell a fully functioning computer that does not require complex setup or configuration. The team’s main focus is building a product people love. “If we can’t do that, Apple isn’t the reason we failed,” he said. With preorders opening and shipping scheduled for the last week of October, Ghost is moving from stealth to the market. The coming months will test whether a $3,499 screenless computer for personal AI can find an audience among users who want agents to run locally, remember deeply, and act autonomously while keeping their data under their own control.

Source:TechCrunch News