Fabraix (YC S26) — The world's frontier hacker for AI agents
Fabraix builds state-of-the-art AI red-teaming agents that continuously detect security vulnerabilities in customer-facing AI. Our product, Nyx, has already found vulnerabilities in agents at dozens of Fortune 500 companies.
Sell to companies like these? Paste your website and see 10 with a filed reason to buy this month. Free.
Opens in a new tab. No signup.
What it does The world's frontier hacker for AI agents (source)
Example from a real report (our sample on Target): “Target eliminated approximately 1,800 corporate positions, with around 1,000 employees receiving layoff notices and the remainder accounted for by eliminating open roles.” Source: NBC News
Fabraix builds state-of-the-art AI red-teaming agents that continuously detect security vulnerabilities in customer-facing AI. Our product, Nyx, has already found vulnerabilities in agents at dozens of Fortune 500 companies. On AgentHarm, the leading benchmark for offensive AI security, Nyx achieved a 78% attack success rate, compared with 67% for GPT-5.6 Sol.
AI agents can change more often than teams can test them. A new model, prompt, tool, permission, or data source can change what an agent does, even when the application code stays the same. And AI is also increasing how much software companies produce and how often it changes. We built Nyx to automate the work required to red-team AI agents. It is multi modal by design and interacts with the target indirectly through controlled replicas of SaaS products and websites that we maintain, placing malicious payloads in webpages, documents, files, messages, and tool outputs to test how an agent behaves under different conditions.
Nyx often finds its first vulnerability within minutes or hours rather than days or weeks. Because the work is fully automated, companies can repeat the test with every change and at a much lower cost than what a 7-figure red-teaming (and point-in-time) engagement would require.
We're also the team behind ACE (Adversarial Cost to Exploit), a benchmark that measures AI security in terms of how much it costs attackers to break an AI system; providing a game-theoretic framework to understand how motivated a rational attacker would be in exploiting the system.
What the site is built with
FastlyhostingGoogle Fontsui
Read from the page's own HTML and response headers on 10 September 2026, at https://fabraix.com/. Each badge matched a string the site serves about itself — hover one to see which. Nothing here is a traffic, ranking or revenue figure: those cannot be read off a page and are not claimed.
Checked away from the directory entry. Neither line below comes from the company's Y Combinator page.
Fabraix builds red-teaming AI agents that continuously find security vulnerabilities in customer-facing AI.
The title and description the site itself served, answering 200. Not a cached copy and not the company's directory entry — what was live when we asked.
Counted only where a thread links to this company's own site, announces it by name on that forum, or names it beside its own Y Combinator batch. A thread whose title merely carries the words is not counted — a name is not an identifier. Hacker News is not us, and nothing above is our opinion of the company.
Ahmed Aly
Founder/CEO
Co-founder @ Fabraix, building a frontier hacker agent that finds failure modes and security exploits in AI systems before real users do. Previously the first data scientist at Sequoia-backed Two, where he built AI systems for adversarial environments from scratch and scaled them to process $1B+ in annual B2B transactions, averting $50M in losses for multiple Fortune 150 companies. A published researcher with multiple papers in Q1 journals and a UCL PhD drop-out.
ibrahim abdu
Founder/CTO
Red-teamed AI agents in Meta's Privacy org; built an SRE agent triaging 3M+ errors/month, saving 10,000+ engineering hours a year. Built a compiler for insider-trading detection at TradingHub: $100M+ in daily trades; Oxford PPE, top 8% of cohort.
Founders, titles and bios as the company published them. Photographs are our own copies of the ones on that record, so they do not break when the original link expires.
No investor is drawn from a Form D — the form states how many invested and never who. The one investor on this graph is named because a batch has exactly one.
Want the whole picture on Fabraix?
Example from a real report (our sample on Target): “Target eliminated approximately 1,800 corporate positions, with around 1,000 employees receiving layoff notices and the remainder accounted for by eliminating open roles.” Source: NBC News
A Full Report on Fabraix: verified, sourced, every claim cited, ready in a few minutes. See pricing.
What the Full Report on Fabraix covers
The Short Answer
Which Company This Is
What They Do, And How The Money Works
Who Runs It, And How To Reach Them
The Money: Funding, Valuation, Runway
Who Pays Them, And Who They Are Up Against
Warning Bells
Sources
10 cited sources found
Plus the sections for what you are deciding, which you pick after ordering.
You get an alert in MentionFox the next time this company shows up.
Is this you?
If you're behind Fabraix: acceptance into a YC batch is one of the few weeks a young company gets looked up by people who matter — investors, other founders, press. FoxPlug turns the commits, deploys, and releases you're already making into that story automatically, so anyone who looks finds something current. GetFoxChat puts a guide on your own site so the people YC's own page just sent you don't read one line and leave.
