Ex-DOJ antitrust chiefs built an AI startup for the cases they couldn't bring
Jonathan Kanter ran DOJ Antitrust for three years and says his team could only pursue about 10% of the cases it thought were worth bringing. His answer, with former deputy Doha Mekki, is a $6.4 million AI startup called FluxCap.
Ten percent. That’s the share of cases Jonathan Kanter says his team at the Justice Department’s Antitrust Division believed were worth bringing but couldn’t, simply because there weren’t enough lawyers. He told ABC News it’s what kept him up at night while running the division from 2021 to 2024: not the cases the government won, but the ones it never got to file.
His fix isn’t a bigger budget request. It’s a company.
Kanter and Doha Mekki, who served as acting head of the same division, unveiled FluxCap on October 28, an AI platform for legal research and litigation strategy that started, by design, with antitrust. The launch came with $6.4 million in seed funding from Y Combinator, Bloomberg Beta, Deep Work Capital, Coalition Operators, Burford Capital and Omidyar Network, according to the company’s own announcement on PR Newswire and corroborated by Law.com, Dealroom and Global Legal Post. The engineering team is led by CTO Marshall Seid and includes several former Palantir staffers, alongside former DOJ and White House officials.
Here’s the part that’s easy to misread from the headlines: this isn’t the government deploying AI against Big Tech. Kanter left DOJ in 2024, before FluxCap existed. What’s actually happening is two of the most prominent antitrust enforcers of the Biden years cashing in that experience to sell a product, first to plaintiffs’ firms and eventually, Kanter told The Verge, to “both public and private enforcers.” Antitrust boutique Kressin Powers is an early customer.
What FluxCap actually does
Per the company’s own description, the platform runs three connected systems: Almanac, which organizes legal rules and precedent; the FluxFacts Engine, which ties factual claims to the evidence that backs them; and Vision Manager, which helps lawyers test and refine litigation strategy. Kanter told The Verge the hardest part was getting the system to catch context, not just keywords. His example: in the Live Nation-Ticketmaster trial, a recorded call had the Live Nation CEO reminding a venue customer that a new arena in town could mean fewer concerts at his venue if he switched ticketing platforms. The customer read it as a threat. Spotting that kind of implied pressure in a pile of discovery documents, Kanter says, is what generic tools like ChatGPT or Harvey “do okay” at, but miss the nuance on.
That claim is Kanter’s own comparison, not an independent benchmark. FluxCap hasn’t published test results against those tools, and I haven’t seen any third party run that comparison either.
How the seed round stacks up
FluxCap’s $6.4 million looks small once you set it next to the rest of legal AI in 2026, and that gap is the clearest sign of how FluxCap is positioning itself: as a narrow specialist, not a general competitor to the big legal-AI platforms.
| Company | Round (2026) | Reported valuation | Focus |
|---|---|---|---|
| FluxCap | $6.4M seed (Oct 28) | Not disclosed | Antitrust litigation, expanding later |
| Harvey | $550M (Sept) | $15.5B | General legal AI across practice areas |
| Legora | $550M (Oct) | $8.8B | General legal AI across practice areas |
Figures for Harvey and Legora are per Dealroom’s reporting; FluxCap has not disclosed a valuation for its round.
What’s confirmed and what isn’t
Note on sources: the launch itself, the $6.4 million seed round, the investor list and the founders’ DOJ titles are confirmed by multiple independent outlets: PR Newswire’s own release, Law.com, Dealroom, Global Legal Post and ABC News all match on those facts. I read the Verge piece in full, plus the PR Newswire release, Law.com’s report and ABC News’s coverage.
Two numbers rest on a single source. The $45.5 million Tech Modernization Fund grant to DOJ Antitrust and the division’s $245 million FY2026 appropriation both come from The Verge’s reporting (the appropriation also traces to a DOJ budget filing The Verge linked); I haven’t found a second outlet that separately confirmed the $45.5 million figure. Treat it as Verge-sourced until someone else checks it. Kanter’s claims about FluxCap’s accuracy versus ChatGPT or Harvey, and his “20-plus people” manpower estimate, are his own, unverified by any published benchmark.
My read
The 10% figure is the most honest thing in this story, and it’s doing work the funding number can’t. FluxCap isn’t trying to be a general legal-AI platform competing with Harvey or Legora on raw scale. It’s a narrow bet that a tool built by people who’ve actually tried to win antitrust cases with 400 government lawyers against companies with unlimited outside counsel will out-perform a horizontal product on that one job.
Whether it does is untested in public. No customer has said FluxCap won them a case yet, and “20-plus people working around the clock” is Kanter’s own estimate of the manpower it replaces, not a measured output. The idea that better software evens a resource gap this lopsided is appealing and also exactly the pitch every legal-tech vendor makes. Kanter’s own framing acknowledges the limit: he told The Verge that even with better tools, “government is always going to be outmatched,” and private defendants will keep outspending whoever is suing them.
Worth separating from FluxCap itself: Kanter’s team at DOJ secured that $45.5 million Tech Modernization Fund grant while he was still in office, meant to fund better case-management tools inside the division, not a startup. He says the division never got to spend it before he left.
If you’re wondering whether this product changes who wins antitrust cases against the companies this site covers, the honest answer right now is: nobody outside FluxCap and Kressin Powers knows yet. That’s not a knock. It’s just where a six-month-old company with one named customer actually is.
This is also the second time this year a federal legal fight has run into the limits of what counts as a reliable AI tool in court; courts are still working out which uses of AI on legal material even clear the bar, as the Third Circuit’s ruling that training on Westlaw headnotes wasn’t fair use showed in September.
Sources
- Jonathan Kanter is using AI to take on Big Tech
- Former Justice Department Antitrust Leaders and Leading Engineers Launch FluxCap
- Former DOJ attorneys launch AI company to help lawyers help more people
- Antitrust Leaders Dive Into Legal Tech With Agentic AI Litigation Platform FluxCap
- Ex-DOJ antitrust chiefs raise $6.4M seed for litigation startup FluxCap
- Former US antitrust head Jonathan Kanter launches AI legaltech start-up