From $2M to $50M in Two Years: Rogo Pulls Away From Hebbia as Claude Looms Over Wall Street AI

The two most closely watched AI startups fighting to become the default operating system for Wall Street just showed their hands — and the gap between them is widening fast.

On September 3, 2026, The Information reported that Rogo, the AI platform for investment banks and asset managers, has passed $50 million in annual recurring revenue — more than tripling from roughly $15 million at the end of 2025. Hebbia, its closest and most storied rival, has also grown, but the top-line divergence between the two companies has become impossible to ignore. And looming over both is a threat neither can build away from: Anthropic’s Claude for Financial Services.

The numbers: a 25x run in two years

Rogo’s trajectory is one of the steepest revenue ramps in enterprise AI. The company did roughly $2 million in ARR in 2024, crossed $15 million in 2025, and has now vaulted past $50 million in 2026 — a 25x increase in roughly two years.

The adoption figures behind that revenue are equally striking. Rogo now counts more than 25,000 daily users across 150 to 250 financial institutions, according to reporting on the company’s figures. Its core product is an AI agent named Felix that automates the work junior investment banking analysts have traditionally performed: building financial models from deal documents, drafting memos and reports, and assembling pitch decks from raw source material.

The company’s April 2026 Series D provides the valuation frame: $160 million raised at approximately $2 billion, led by Kleiner Perkins. Against $50 million in ARR, that prices Rogo at roughly 40x revenue — rich by traditional enterprise software standards, but almost modest in a year when horizontal AI companies have commanded far frothier multiples on far less certain revenue.

Two bets on the same market

Rogo and Hebbia are often lumped together as “finance AI,” but they entered the market from opposite ends, and that difference now defines the race.

Rogo, founded by former Lazard banker Gabriel Stengel, is laser-focused on the transactional side of investment banking. Its tools generate the deliverables banks actually bill against: LBO models, comps tables, CIM summaries, pitch materials. The strategy is vertical-native — work the deal task list end to end, inside the bank’s compliance perimeter. It is deployed at names including Lazard and Nomura, and has raised over $300 million in total funding.

Hebbia, founded around 2020 by Stanford PhD George Sivulka, took the research end. Its Matrix platform became famous for a deceptively simple innovation: instead of answering one question about one document, it let users ask many questions across thousands of documents and arrange the answers in a grid — one row per document, one column per question, with granular citations. For a time, Hebbia appeared to be on a tear, pushing into investment banks, law firms, and asset managers, hiring a sales force, and plastering New York City with billboards.

Hebbia reported $13 million in ARR as of mid-2024 and has since claimed profitability — a meaningful achievement that Rogo has not matched. But Hebbia’s last known valuation sits around $700 million to $1.3 billion on roughly $161 million raised, meaning Rogo’s $2 billion mark is now nearly triple its rival’s, and the revenue gap has widened considerably despite Hebbia’s healthier margins.

The copying problem

Hebbia’s deeper challenge isn’t Rogo — it’s that its signature innovation became table stakes. As Hebbia president and CTO Aabhas Sharma put it in a recent Business Insider interview: “The whole market started copying it.”

Rogo has a table interface. Legal AI giant Harvey built Vault to review documents in bulk. Legora ships Tabular Review. What was once Hebbia’s moat is now a standard feature category across the vertical AI landscape.

Hebbia’s answer is a sweeping relaunch of Matrix, announced in late August 2026: the new version takes a user’s request, autonomously builds the table needed to do the job, and then converts the results into a finished memo or slide deck. In other words, Hebbia is moving from “structured document search” toward Rogo’s home turf of workflow automation and deliverable generation — while Rogo grows fastest on exactly that motion. “This is going to be our mark in the sand,” Sharma said of the relaunch.

The Claude problem

The uncomfortable backdrop to both companies’ growth is that their market’s ceiling is now set by companies far larger than themselves.

Anthropic has been pushing Claude into financial services since launching Claude for Financial Services in mid-2025, and in May 2026 it shipped ten ready-to-run finance agent templates on the Claude platform — pitch deck generation, earnings review, diligence workflows and more — underpinned by Claude Opus 4.7, which Anthropic says leads Vals AI’s Finance Agent benchmark at 64.4%. Claude now connects to dozens of market data and research platforms natively, and The Information’s reporting frames this explicitly: the Rogo–Hebbia revenue race is happening “as [a] Claude Finance threat looms.”

The strategic logic for the startups is that banks will not entrust regulated deal workflows to a general-purpose chatbot — they need finance-native compliance, audit trails, data permissions, and workflow integrations that horizontal providers are slow to replicate. The counter-logic is that frontier labs are vertically integrating fast, benchmark scores keep climbing, and a general model that is 95% as good at 20% of the effective cost eats the market from below.

Both startups are also, awkwardly, Anthropic customers and partners: Rogo has appeared in Anthropic’s own partner showcases, and both companies build on the very models whose maker is now courting their customers directly. That dual relationship — supplier, distribution channel, and existential competitor in one — is the defining tension of the vertical AI moment.

What it means

Three takeaways worth watching:

1. Revenue is the new benchmark. In a market saturated with capability claims, $50M of real ARR at 150+ institutions is the clearest signal yet that AI agents have crossed from demo to deployment in one of the world’s most demanding, compliance-heavy industries.

2. The bundling war has begun. Rogo’s bet is that deal-native workflow depth wins; Hebbia’s bet is that analytical breadth plus autonomous deliverables can retake the lead. But both are now competing not just with each other — they are competing with Harvey, Legora, AlphaSense, and the frontier labs themselves, all of whom are bundling the same capabilities.

3. Valuations will be tested. A 40x revenue multiple assumes Rogo keeps tripling. If Claude’s finance agents commoditize the analyst workstream faster than Rogo deepens its workflow moat, that multiple compresses violently. If vertical-native depth holds, Rogo becomes the template for how vertical AI survives the frontier labs.

For now, the scoreboard reads: Rogo $2B valuation and $50M+ ARR, growing fastest; Hebbia profitable and pivoting; Claude waiting at the gates. The next four quarters will decide whether vertical AI platforms become the durable operating layer of finance — or a transitional generation of software that the frontier models simply absorb.