Loora raises $22M Series B to scale its AI English tutor

Loora, the AI English tutor, announced a $22 million Series B led by Union Tech Ventures, with vgames plus existing investors Emerge, Hearst Ventures, and QP Ventures. The company says the raise brings total funding to $43.25 million and will fund AI research and engineering, go-to-market across the US, Europe, Asia, and South America, and learning-data infrastructure, with headcount expected to double from 30 to 60 within six months.

FINANCE desk — same-day primary that an AI English tutor the company says has reached 15 million learners just closed a disclosed Series B to double the team and expand globally, a concrete money signal in AI for skills and language.

The round was led by Union Tech Ventures, with participation from vgames and existing investors Emerge, Hearst Ventures, and QP Ventures. The company says the raise brings total funding to $43.25 million. File the $22 million figure, the Series B label, the Union Tech Ventures lead, those named participants, and the $43.25 million cumulative line as Loora’s. This desk is not inventing a valuation, an ownership split, or a check size. None was printed.

Use of proceeds, as the same wire has it: three priorities. Scale AI research and engineering. Accelerate go-to-market — sales and launch work — across the US, Europe, Asia, and South America. Expand the data and research infrastructure that underpins Loora’s learning outcomes. File that three-part spend plan as Loora’s. This desk did not sit in a board meeting.

Hiring, still company: headcount is expected to double from 30 to 60 within six months. The majority of new hires, the company says, will focus on AI research and engineering, plus continued investment in learning science. File the 30-to-60 / six-month / research-and-engineering picture as Loora’s. This desk did not count employees.

Product, as the same wire has it: Loora is a conversational AI English tutor — speaking-practice software that talks with a learner, not a generic chatbot wrapper on someone else’s model. The company says building fluency takes more than connecting users to a frontier model. It says it needs a data flywheel: a system that, from hundreds of millions of real interactions, learns which paths actually produce retention and fluency. Retention here means learners coming back and keeping what they practiced. A large language model, or LLM, is the kind of AI that writes and answers in sentences; the company says general-purpose LLMs are built for breadth — coding, productivity, search — not the personal, iterative work of language learning. File that conversational-tutor / data-flywheel / not-just-a-wrapper picture as Loora’s. This desk did not run a lesson.

Learner and revenue claims, company-attributed — do not independently certify: 15 million learners worldwide; more than 200 million learner interactions; revenue more than doubled in the last twelve months; annual recurring revenue, or ARR, “in the tens of millions.” ARR is the company’s yearly-run-rate sales figure if current subscriptions kept going. File those 15 million / 200 million / revenue-more-than-doubled / tens-of-millions-ARR lines as Loora’s. This desk did not count learners, sit on the books, or invent an exact ARR dollar number. The wire does not print one.

B2B and who uses it, still company: Loora says it built its business-to-business, or B2B, side from scratch over the past year and signed more than 100 corporations, governments, and universities globally. The core user, the company says, is a fluency-driven adult aged 25 to 55, with 72% learning English to advance their careers. Retention, which the company calls one of the clearest indicators of learning outcomes, has more than doubled since launch. File the more-than-100 / 25-to-55 / 72% / retention-doubled lines as Loora’s. This desk did not audit a customer list or a retention cohort. The Sep 16 wire does not name those 100 organizations. This desk is not inventing logos.

Named voices on the release: Roy Mor, CEO; and Roxanne Horesh, managing director at Union Tech Ventures. The wire also names Yonti with Roy as the founding team. File the names and titles as theirs, via Loora. Their other quotes are color only and stay in Sources. This is not investment advice.

Opportunity backdrop, still company-attributed: Loora says English fluency is one of the most reliable predictors of advancement, and that 95% of the global population are not native speakers, yet must reach fluency to access the 98.5% of employers across 38 developed economies that use a formal English competency assessment during hiring. The average online human tutor, the company says, ranges from $15 to $50 an hour. File those 95% / 98.5% / 38-economies / $15-to-$50 lines as Loora’s figures — not a desk census. This desk did not rerun a labor survey.

Plain English for the rest of the card: Series B = a growth venture round after Series A. ARR = annual recurring revenue, a yearly-run-rate sales figure. conversational AI English tutor = speaking-practice software that talks with a learner, not a generic chatbot wrapper. data flywheel = a loop where more practice sessions make the tutor better at picking the next useful path. LLM = large language model, the kind of AI that writes and answers in sentences. GTM / go-to-market = sales and launch work in new countries. B2B = selling to companies, governments, and universities, not only to individual app users. fluency = being able to speak and understand English well enough for work or school. headcount = how many people the company employs.

PRIMARY here: Loora’s 16 Sep 2026 Business Wire company release — Tier A PRIMARY company source, the original record, wire id 20260916632941. The Wedbush FinancialContent page is a full-text redistributor of that same originating wire, not a second newsroom. The loora.com company home is product-home context, not a third originating announce. The $22 million Series B, the Union Tech Ventures lead, vgames / Emerge / Hearst Ventures / QP Ventures, the $43.25 million total-funding line, the AI-research / GTM / learning-data spend plan, the 30-to-60 headcount plan, the conversational-tutor / data-flywheel product, the 15 million learners / 200 million interactions / revenue-more-than-doubled / tens-of-millions-ARR lines, the more-than-100 B2B / 25-to-55 / 72% / retention-doubled picture, the Mor / Horesh / Yonti titles, and the 95% / 98.5% / $15-to-$50 backdrop are company-attributed. Learner counts, ARR framing, B2B counts, and hiring plans stay company-attributed — not independently audited here. NOT claimed: a valuation, an exact ARR dollar number, churn, independently verified learner or retention counts, that Loora is only a generic chatbot wrapper, that this desk tested the app or saw a term sheet, a stock tip, or investment advice. Distinct from the already-filed footprint-25m-series-b, backops-42m-series-b, arcee-series-b-1b, abre-vera-analytics-agent, phrase-atlas-ga, gates-foundation-1b-ai-access, xai-grok-build-memory, and apex-intelligence-50m-angel.