EN
Back to the archive

The archive · Money & Fintech · Product decision · 2025–2026

Astor bets US retail investors pay for an AI analyst: $5M seed, $200M+ linked

Astor (ex-Gaus): Brazilian founders sell US retail a $15/mo AI analyst — YC S25's only Brazilian pick, $5M Monashees seed, thousands of users.

Astor (Gaus)

The betRetail investors would pay ~$15/month for an always-on AI analyst replacing $5,000-a-year human advisors; if they won't, the subscription dies.Live

What the business is

Astor (launched as Gaus) is a San Francisco-based, SEC-registered AI investment advisor founded by two Brazilians: its iOS app connects to a user's brokerage via Plaid, scores the portfolio on risk and diversification, and delivers portfolio-specific research and alerts by text or voice for a $15–40/month subscription.

Starting capitalUS$500K YC S25 check; US$5M seed led by Monashees (April 2026) with YC, Goodwater Capital, Gilgamesh Ventures, 468 Capital, Valutia and executives from Stripe and OpenAI.

How it started

Bruno Koba (data scientist at Nubank, later investor at Monashees) and Daniel Tulha (engineer at Stripe, Robinhood and Amazon) got the idea in February 2025, joined YC's Summer 2025 batch as the only Brazilian startup among 180 companies chosen from 30,000 applicants, and entered the program with a 1,000-person waitlist, a $19/month plan and coverage of ~30,000 US stocks.

What happened

The company (Gaus, Inc.) kept the Gaus name through YC, then launched its iOS app around February 2026 as Astor. In late April 2026 it raised a $5M seed led by Monashees as an SEC-registered investment advisor — Goodwater, Gilgamesh, 468 Capital and Stripe/OpenAI executives joined — and within two months of launch reported thousands of users, $200M+ in connected brokerage assets and a 4.8/5.0 rating, priced at $15/month (core) and $40/month (Pro).

How it ended up

Still live and early: as of its August 11, 2026 Launch YC post, Astor was pitching the same thesis — an always-on AI analyst that closes the retail-investor research gap — with no later revenue or funding disclosed beyond the April seed.

Background

Astor began as Gaus, a February 2025 idea from two Brazilian fintech builders — Bruno Koba (data scientist at Nubank, then investor at Monashees) and Daniel Tulha (engineer at Stripe, Robinhood and Amazon) — to give retail investors the analyst coverage institutions take for granted. The bet: individual investors, who now drive 25% of US trading volume yet mostly say they lack confidence, would pay a monthly subscription for an always-on AI analyst that watches their actual portfolio, not a generic robo-advisor allocation.

In YC's Summer 2025 batch — the only Brazilian company among 180 selected from 30,000 applicants — Gaus entered with a 1,000-person waitlist, a planned $19/month plan and monitoring coverage of roughly 30,000 US stocks, ETFs and crypto. The platform connects to the user's brokerage through Plaid, scores the portfolio on performance, risk and diversification, and answers by text or voice, from Roth IRA vs 401(k) comparisons to company-specific news on holdings.

The product launched as Astor around February 2026, and in late April the company — an SEC-registered investment advisor operated by Gaus, Inc. — closed a $5M seed led by Monashees with YC, Goodwater, Gilgamesh, 468 Capital and executives from Stripe and OpenAI. Within two months of launch it reported thousands of users, $200M+ in connected brokerage assets (some reports cite $300M+) and a 4.8/5.0 rating, priced at $15/month core and $40/month Pro.

As of August 11, 2026, Astor's Launch YC post repeated the original thesis — an AI analyst that never sleeps, tailored to each user's holdings — with no further revenue or valuation disclosed. The case is a live experiment in whether fiduciary-grade AI advice can be sold as a subscription to the mass market that human advisors' $5,000 minimums exclude.

What has to be true

  • Koba and Tulha saw both sides — institutional data advantage and retail's lack of confidence — and bet retail investors (25% of US volume) would pay for research, not allocation advice.
  • YC S25's only Brazilian company: the 0.6% acceptance rate (180 of 30,000) plus a 1,000-person pre-launch waitlist showed the idea had pull before the product shipped.
  • Astor made personalized concrete: Plaid-connected real holdings, portfolio scoring and per-position research justify a subscription where generic robo-advisors could only charge a fraction.
  • The $5M Monashees-led seed with Stripe/OpenAI executives and a 4.8/5.0 rating are early but consistent signals of product-market fit, still unproven at scale.

What can be applied

Make 'personalized' literal: Astor ties insights to the user's real holdings, so the subscription replaces a concrete $5,000-a-year advisor — portfolio-specific value is what the price defends.

Aftermath

As of August 11, 2026, Astor is live and expanding: an SEC-registered RIA in San Francisco with a team of five, thousands of users, more than $200M in connected brokerage assets and a 4.8/5.0 rating. The April 2026 $5M Monashees-led seed funds product, engineering and growth, with deeper integrations planned. Open questions: whether subscribers stay at $15–40/month, how advice holds up under regulatory scrutiny, and whether AI analyst is a durable category or a feature brokerages absorb.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases