The Miss Is the Signal, Not the Score: Why We Backed Hive
Most literacy platforms treat a wrong answer as the end of the transaction. A student misspells a word, the app marks it wrong, and everyone who got it wrong gets the same drill next. Hive starts from a different premise: the wrong answer isn't noise; it's the most useful data point in the entire system.
Three Kids, One Word, Three Different Problems
Take the word "occurred." One student writes Ocurred, a prefix assimilation error. Another writes Occurd, a grammar and e-dropping error. A third writes Ockurred, a Latin-derived spelling pattern error. On a normal spelling test, all three get an identical red X and an identical follow-up drill. Hive treats them as three different students with three different gaps, because that's what they actually are.
What Hive Built
Hive is an AI-powered adaptive learning platform that captures every miss with letter-level context, maps it to the specific literacy concept behind it, updates an evolving map of what the student has and hasn't mastered, and assigns the next lesson, explanation, or drill based on that map, not a fixed curriculum. The company is building this in public in three stages: a live V1 focused on spelling bee preparation (curriculum, competitions, analytics, already live in the US, UAE, and Oman), a current V2 layer of adaptive practice that recommends next steps from a student's most recent missed words, and a next-stage concept-level AI engine that builds full miss-to-concept maps from exact error patterns. Spelling is the starting point because it produces clean, objective, letter-level data. The same diagnostic engine is built to extend into vocabulary, reading, and writing next.
A Team Built for Exactly This Problem
Ekansh Rastogi placed 5th at the 2022 Scripps National Spelling Bee, has nine years of spelling coaching and competition experience, and serves as a national spelling subject matter expert for an international nonprofit working with 15,000+ students across 75+ chapters. Akshat Rastogi is a software engineer with experience at Honeywell Aerospace and Siemens, and previously founded Gathr, a campus engagement app that reached 50,000+ users and 3,000 student organizations. One brother understands exactly where literacy instruction breaks down. The other knows how to build and ship software at scale.
Proof at the Hardest Level First
Hive didn't start with a beta test in a general classroom. It started at the elite level, where the diagnostic engine had to work under real pressure. Hive users are 2x more likely to place in the world's top 100 spellers, and 14 Hive users currently rank among that top 100. The company is the official learning platform for both the UAE and Oman National Spelling Bees, and all three UAE podium finishers this year used it. That's 1,400 registered users across four countries, with 119 paying users even in the off-season, a strong signal for a product whose core use case is seasonal.
The Path From Spelling Bees to Every Classroom
The go-to-market is deliberately sequenced. Spelling bee partnerships bring in students at scale, every missed word from that usage becomes training signal for the adaptive engine, a stronger engine makes better miss-to-concept decisions, and that improved product is what expands Hive into direct family subscriptions and, eventually, school and district licensing. Today, families can go from free to Plus ($10/month) to Elite ($25/month) directly on the web. Competition partners generate revenue two ways: $3.40 per speller registered for a competition, plus $2.05/month for paying users converted through that freemium funnel, a model already live through Hive Learn in the UAE and Oman. School and district licensing, charged per seat, is currently in development.
The Size of the Opportunity
Hive is targeting a beachhead inside a much larger market: an $11.8B global children's English learning TAM, a $6.2B global digital children's English learning SAM, and a 3-year attainable revenue target (SOM) of $10.7M starting from spelling.
The Bigger Bet
Edtech has spent a decade promising personalization and mostly delivering adjustable difficulty sliders. Hive is betting that real personalization starts one level deeper, at the level of why a student got something wrong, not just whether they did. They've already proven that diagnostic engine works under the most competitive conditions in the sport. Now they're building the on-ramp from spelling bees to every classroom. That's the bet we're making with them.
