Silicon Valley has a new contender in the race for on-device AI. Underdog, launched by Thiel Fellow Sigil Wen, aims to be the most private AI assistant yet—and it comes with a business model that does not rely on harvesting your data.
Who Is Sigil Wen?
Sigil Wen’s path to launching a private AI assistant began when he was just 17. A self-taught coder, he moved to Silicon Valley and lived in an AI hacker house alongside famed AI researcher Andrej Karpathy.
During that time, he coded with people who would become some of the biggest names in AI, including Perplexity founder Aravind Srinivas and OpenAI researcher Noam Brown. He also tested early versions of tools that later became well known: a chatbot shared by Anthropic co-founder Ben Mann that would become Claude, an image generator from David Holz that would become Midjourney, and what would become OpenAI’s GPT-3 and the image generator Stable Diffusion.
Prominent investor and entrepreneur Naval Ravikant hired Wen for Airchat, Ravikant’s now-defunct rival to Clubhouse. For fun, Wen even figured out how to get GPT-2 running on his Apple Watch.
“It was a magical time,” Wen said.
Underdog Launches as a Private AI Assistant
Now a Thiel Fellow—the program from investor Peter Thiel that invites young founders to pursue projects instead of college—Wen launched an invite-only beta of Underdog on Monday.
Underdog is one of the most private AI assistants Silicon Valley has yet to offer. The model runs wholly on-device, meaning user data remains on devices they already own. It currently supports Macs and Windows PCs, with Linux, iPhone, and Android versions coming soon.
How Underdog Keeps Your Data Private
Underdog is powered by Husky, the inference engine Wen built to run AI models quickly on a user’s own hardware. According to Wen, Husky moves less data between a computer’s main chip and its graphics chip than rival on-device engines.
The private AI assistant includes other security features as well, such as encrypting the keys to the email and other accounts that users authorize Underdog to access.
The Trade-Off: Smaller On-Device Models
Underdog uses much smaller models than today’s state-of-the-art models hosted in data centers. It currently runs a 27-billion parameter reasoning model fine-tuned from Qwen3.8-27B.
Wen argues that this model compares favorably with Claude Opus 4.6 in some benchmarks—or what was considered top performance six months ago. He says that means it can handle everyday tasks people want from an AI assistant, like shopping research or answering math homework questions.
“You don’t need to sacrifice your privacy for the capability because they’re just as capable,” Wen says. He adds that small on-device models will continue to grow more capable over time.
Underdog’s Unusual Business Model
Perhaps the most interesting aspect of Underdog is its early business model. The app will be free at first and never ad-supported.
Because the AI runs on users’ machines, Underdog does not carry the giant overhead of paying a provider for inference. “I don’t have to charge you a subscription to run this because my costs are so super low,” Wen said.
Instead, with Stripe co-founder Patrick Collison as one of his angel investors, Wen is borrowing a play from the fintech era. Underdog will take a tiny percentage of payment transactions that the AI assistant makes using Stripe’s secure payment rails—something like an interchange fee.
In this way, the private AI assistant never has to mine your data. As Wen describes it, Underdog is as aligned with you as your bank or credit card provider.
Why Data Collection Matters for AI Assistants
This approach runs contrary to the business motivations of many other AI assistant players. Their privacy policies allow them to collect user data that they may sell to advertisers or other third parties, or use to train other models.
That kind of data collection could be a particularly treacherous trade-off for AI assistant users. An assistant may need access to the most intimate details about you to be useful—from medical conditions to financial data to information about your kids.
As Wen wrote in what he calls his AI manifesto: “Why should using AI require surrendering your private information?”
He added: “I honestly want to build Underdog for myself. I’m building a product that I would be proud for my future children to use.”
Who Is Behind Underdog?
The startup behind Underdog is named Conway Research, and Collison is not its only big-name investor.
Conway is backed by Andreessen Horowitz via partner Chris Dixon, as well as Khosla Ventures, Hummingbird, SV Angel, and the Anthology Fund—the partnership fund between Menlo Ventures and Anthropic.
A prominent list of angel investors also includes Vercel founder Guillermo Rauch, Noam Brown, and Deedy Das, among others.
What Underdog Means for Private AI
Underdog enters a market where privacy concerns around AI assistants are growing. By running entirely on-device and avoiding ad-based monetization, it offers a different path than many competitors.
| Feature | Underdog’s Approach |
|---|---|
| Data location | Stays on your device |
| Model size | 27-billion parameter (Qwen3.8-27B fine-tune) |
| Pricing | Free at first, never ad-supported |
| Revenue | Small percentage of Stripe payment transactions |
| Platforms | Mac, Windows (Linux, iPhone, Android coming) |
Whether Underdog can compete with larger, cloud-based models remains to be seen. But for users who prioritize privacy, it represents a notable alternative—one built on the belief that you should not have to trade your data for a capable AI assistant.
