How Redis Labs CEO Ofer Bengal Drove Revenue Growth to a ~$72M Run Rate by 2018

In July 2018, Redis Labs co-founder and CEO Ofer Bengal told Latka his database company was growing revenue roughly 60% year over year, serving about 8,500 paying customers, and running its entire cloud database service with five DevOps engineers. He would not state revenue — "we are not sharing numbers because we are a private company" — but when host Nathan Latka worked backward from Bengal's stated goal of passing $9 million in monthly recurring revenue about a year out, he pegged the company at roughly $6 million a month, or around a $70M run rate. GetLatka's database records that estimate as $72M ARR for July 2018.

Bengal's plan on that tape was an IPO in about two and a half years. That is not what happened. The company took a $110M Series G at a $2B-plus valuation in April 2021, dropped "Labs" from its name that August, replaced Bengal with Rowan Trollope as CEO in February 2023, and — still private — announced it passed $300M in ARR in January 2026.

From Garantia Data to Redis Labs

The old version of this post claimed Redis Labs was "founded in 2011... rebranding as Redis Labs" and hung section headers on a 2011/2012 timeline the interview never supported. Here is the version the tape and the public record actually agree on.

Bengal and co-founder Yiftach Shoolman started the company in 2011 — under a different name, Garantia Data — to accelerate application performance with a caching system. Under the hood they used Redis, the open-source project Salvatore Sanfilippo had started in Sicily, "half hobby, half something that he built for Telecom Italia," as Bengal put it. "A year after, we saw that Redis is becoming so popular as an open source that we better do something about it, and we decided to become a Redis company," he told Latka. The formal rename to Redis Labs came in 2014, per the company's own press release. Sanfilippo later joined to lead open-source development, and the business split into open-source Redis plus Redis Enterprise, the commercial product sold as managed cloud databases or as software.

Redis Labs revenue growth: what the tape actually supports

Bengal gave Latka a customer breakdown in July 2018: roughly 8,500 paying customers plus about 100,000 free-tier users. Of the paying base, about 250 were enterprise software customers averaging around $60,000 a year — with Fortune 100 accounts averaging closer to $300,000 and some "paying millions of dollars per year." The other ~8,250 were long-tail database-as-a-service customers at about $1,000 per year (Bengal said "per month," then immediately corrected himself to per year; the episode's closing recap repeated the per-month figure, which is where later confusion started).

The old post's central claim — that the company was "projected to hit $9.4 million in monthly recurring revenue within a year" alongside 60% growth — was a garbled version of host arithmetic. The $9.4M figure was Latka's calculation, not Bengal's. What Bengal actually committed to on tape: the company would pass $9 million a month "approximately a year from now," and revenue was "growing approximately sixty percent year over year." When Latka reversed that into "you're probably somewhere around call it six million per month right now," Bengal declined to confirm: "I won't like to share it... we are approximately two years from an IPO readiness."

Worth being honest about: the averages Bengal quoted don't multiply out to that run rate. 250 customers at $60K plus 8,250 at $1K is roughly $23M ARR — a third of the ~$70M the host inferred. The gap sits in the top of the book, which averages conceal: $300K Fortune 100 accounts and the customers paying millions. Neither the $60K average nor the $6M-a-month estimate was a number Bengal stated as revenue, and the episode itself was titled around "$60m+ in ARR" — Latka's own, more conservative read of the same math. Treat the 2018 figure as a host estimate Bengal let stand, not a disclosure.

The actual tactics: automation instead of professional services

The durable part of the interview is the operating model, and it needed no embellishment. Redis Labs rented thousands of cloud servers across AWS, Google Cloud, and Azure and ran customer databases on them "in a fully automated manner." When Latka suggested humans must be managing things behind the code, Bengal pushed back: "We do not do professional services. We are against this model. We think it's not good." Hundreds of thousands of databases, he said, were managed by five DevOps engineers.

That automation showed up in the funding math Bengal was proudest of. The company had raised $78M by mid-2018, and he expected to raise at most $50M more before an IPO — around $130M total. "If you take a company like MongoDB, they raised approximately north of $300 million before they went public. This means that we probably are way more efficient," he argued. Latka pushed back that MongoDB raised big precisely because it had a rep model it trusted — and Bengal conceded he had no payback-period model for the field-sales team he was hiring into, from 200 employees to a planned 270 by year-end. "These are numbers... some we are not tracking at the moment because we are in a very aggressive growth period."

On sales motion, Bengal was specific that the growth engine was field reps, not inside sales: "When you talk to these giants, you need to be in constant contact with them... with solution architects, with salespeople, with finance." The retention numbers backed the enterprise focus: gross revenue churn in the single digits annually, half of each quarter's growth from expansion, and a typical customer doubling its spend within 24 months. Asked for net revenue retention, Bengal said he wasn't sure of the term — but with sub-10% gross churn and that expansion rate, well above 100% is the only arithmetic that works.

The IPO that never happened — and what did

Bengal's two-and-a-half-year IPO clock, started in mid-2018, expired quietly. The record since:

Bengal's reasoning for an IPO on the 2018 tape — "the idea is to build a long-term business and not to do a quick exit" — arguably still describes the company; it just never needed the public markets to do it. The 2021 Series G with its $200M secondary delivered exactly the investor-and-employee liquidity Latka had suggested as the IPO alternative.

Where GetLatka's own records disagree

What holds up from the 2018 playbook

Strip out the garbled numbers and the stale IPO framing, and three things Bengal described in 2018 still explain the company that passed $300M ARR: a genuinely automated cloud service (five DevOps engineers for hundreds of thousands of databases), an enterprise book where expansion — not new logos — drove half of growth, and a refusal to sell professional services at all. The part that aged worst was capital efficiency as identity: the company Bengal said would never need more than $130M raised roughly $347M net by 2021, at a $2B valuation. It got the scale; it just paid market rate for it.