Oleve Revenue: Inside the $6M ARR Studio Behind Quizard AI

Oleve, the New York consumer-AI studio behind the study apps Quizard AI and Unstuck AI, was generating $6 million in annual recurring revenue — roughly $500,000 a month — with a team of six when co-founder Sid Bendre sat down with Nathan Latka in mid-2025. The episode, “How This 25 Year Old College Kid Built a $500K/Month App Empire,” reached the podcast feed on July 15, 2025. In it, Bendre put profit margins in the 10–30% range and total downloads across the portfolio at about 4 million. Latka’s company database still lists Oleve at $6M in revenue, now with nine employees, as of its August 10, 2026 update.

That answers the revenue question. The more interesting story is how a homework app built in a University of Rochester dorm room got there in about two and a half years — and which of the numbers floating around it actually hold up.

Oleve revenue: the dated record

All revenue figures are self-reported by the company; Oleve is private and nothing here is audited.

The arithmetic closes: $500,000 a month annualizes to $6 million, and spread across the six people on staff at the time of the interview, that is $1 million in revenue per employee per year. The earlier version of this article advertised a “$1M revenue per founder formula” and then computed $1.5 million per founder two lines later — the cleaner, honest number is the per-employee one.

From a dorm room to 10,000 users in about 30 hours

Quizard AI was built by three co-founders — Michael Giardino (CEO), Sid Bendre, and Achraf Golli — while they were students at the University of Rochester, with Giardino’s childhood friend Ben Or-Chen joining early as founding marketing lead. Launching in January 2023 mattered: consumer AI was weeks old as a category, and a homework solver that pointed a phone camera at a problem was still novel.

Distribution came from one TikTok. The hook — ChatGPT and Photomath had a baby — did about 2.3 million views overnight and brought in roughly 10,000 users in under 30 hours, a story Bendre retold on the Latka episode and the company repeats in its own materials. Growth compounded from there: Giardino’s ASU GSV speaker profile says Quizard has helped over 4 million students with more than 75 million academic questions, and Bendre’s Forbes Technology Council bio (current as of 2025) claims 4M+ users, $6M+ ARR, 100M+ social views, and backing from the accelerator-fund Neo.

The pricing history: free queue-skipping to a $9.99 weekly paywall

The revenue engine took three attempts, as Bendre walked through on the tape. Version one was free, with paid queue-skipping at peak times — it monetized urgency but capped revenue. Version two was a coins-and-credits system that mostly created friction. Version three is the one that stuck: weekly subscriptions around $9.99 a week, with the paywall shown immediately at signup rather than after users cool off. Bendre credited the switch to the weekly-pricing playbook Nikita Bier has long pushed publicly on X.

Around that paywall sits an unusually heavy testing habit for a six-person shop. The episode notes describe 6–7 A/B tests running simultaneously on the main product, with a few more on secondary apps; Bendre named Superwall for paywall experiments and Mixpanel for analytics, and said experimentation infrastructure is the first thing built into every new product. The same notes mention AI agents that scan for app-store arbitrage opportunities — under-served niches where a proven playbook can be redeployed.

Six people, three teams

On the tape, Bendre described the company as three small organizations. A growth arm, led by Giardino and Or-Chen, owns viral content and acquisition. A product engineering arm under Golli runs each app with its engineers acting as owners of their own metrics. And a platform arm — Bendre’s — abstracts every recurring process in the company into tooling, with a stated ambition of an operation where agents handle most of the work: find the next app to build, then make it go viral.

Whether that vision scales past edtech is unproven, but as an explanation for how six people ran a $500K-a-month portfolio profitably, the structure is the substance of the interview.

What checks out, and what doesn’t

• Founding date. Latka’s database lists Oleve as founded in 2024; Quizard AI verifiably launched in January 2023. The likely reconciliation is that Oleve, the studio entity wrapped around the apps, was formalized later — but the database date and the launch date disagree, and we flag it.
• Headcount. Six at the time of the mid-2025 interview, nine in the database as of August 2026 — consistent with a company that kept hiring slowly.
• Margins. The episode notes say 10–30%; the earlier version of this article narrowed that to 20–30% without a source. Either way it is a wide, self-reported band.
• Unstuck’s early growth. The earlier article claimed 1 million users in nine weeks; Startup Spells claims 8 million in nine weeks. That is an order-of-magnitude conflict, neither number is independently verifiable, and we are not endorsing either.
• Beyond $6M. Some company materials in 2025–26 gesture at bigger numbers, but $6M ARR remains the figure Bendre stated on the record and the one Latka’s database carries as of August 10, 2026.