Redrob.

Redrob is an AI company that builds large language models and, on top of them, a set of go-to-market tools for two of the least glamorous jobs in business: finding customers and finding people to hire. That is the whole pitch, and it is a more interesting pitch than it sounds. Because the thing about sales prospecting and recruiting is that they are both, underneath, the same problem - search, plus trust. Who is this person, is their email real, do they actually want what I'm selling, can they actually do the job. Redrob's bet is that if you put a cheap-enough language model under all of that, the economics change for a lot of companies that were previously priced out.

The company sits under the McKinley Rice umbrella and now runs offices in San Francisco, New York, New Delhi, Mumbai and Seoul with roughly 100 people. Its data layer is the part that makes the sales team's eyes light up: a pool of more than 700 million contacts, over 100 million verified accounts, and 180 million-plus email verifications that it claims land at 98% accuracy. Numbers like these are the sort of thing you either believe or you don't, but they are the raw material for the actual product - a prospecting engine that hands a small sales team the kind of contact database that used to be a Fortune 500 luxury.

That framing is not marketing gloss; it is the founder's stated reason for existing. "We are taking the resources that Fortune 500 companies have globally," Felix Kim has said, "and giving these resources to small businesses and startups." Redrob's mission language leans hard on the word affordable, and its LLM pitch - roughly 90% of flagship-model performance at around 5% of the cost, built with a Mixture-of-Experts architecture plus aggressive distillation and quantization - is the technical expression of the same idea.

The strategic wrinkle is who Redrob is building this for. Not, primarily, the enterprises that can already write a check to OpenAI. Redrob is aiming at emerging markets and cost-conscious companies - "the next three billion users," in its telling - and describes itself, memorably, as the Android to ChatGPT's iPhone. It is an audacious analogy and also a genuine strategy: be the affordable layer that shows up everywhere, rather than the premium object everyone photographs.

Which brings us to the founder, because Redrob is one of those companies where the founder's biography is not a footnote. Felix Kim grew up moving between the United States, South Korea and Peru, and by his own account was bullied at each stop for being the outsider. He went into law, practiced internationally in Seoul, and rose to become Vice President of the Korean Bar Association - a resume that reads nothing like "AI founder." In 2018 he left it, having concluded that the real friction in global business was trust: people, and companies, distrust what they do not know.

He relocated to Noida, India, and spent about five years bootstrapping an IT outsourcing venture before Redrob took its current shape. That long unglamorous runway matters, because it is why the company's later numbers arrive with revenue attached rather than just a model demo. By the 2023 seed round Redrob reported around $2.1 million in annual recurring revenue and 200,000 users. By the 2025 Series A, those had become roughly $7 million ARR and about 3 million users, many of them flowing in through 500 university partnerships in India where Redrob's recruitment tools became a distribution channel.

The money tells its own quiet story about geography. Redrob's $4 million seed came from Murex Partners and DS & Partners; its $10 million Series A, announced in November 2025, was led by Korea Investment Partners with KB Investment, Kiwoom Investment, Korea Development Bank Capital, Daekyo Investment and DS & Partners along for the ride. Seoul money, Indian users, San Francisco sales motion. It is a genuinely tri-continental company, which is either a coordination nightmare or exactly the point, depending on how much you believe the trust-across-borders thesis.

The stated destination is deliberately immodest: Redrob wants to be the world's third-largest LLM platform by monthly active users by 2028, behind only ChatGPT and Gemini. That is the kind of goal that is easy to raise an eyebrow at, and the company knows it. But the more grounded read is that Redrob does not need to beat the frontier labs at their own game. It needs to be the default, affordable AI layer for sales and hiring in markets the frontier labs aren't optimizing for - and to keep the revenue growing while it tries. On that narrower scoreboard, the $2.1M-to-$7M jump is the number worth watching.