Leviathan

What do they actually do

Leviathan is an early YC S24 startup building an AI encyclopedia — effectively an AI-native Wikipedia that aims to deliver citable, encyclopedic answers. Public listings name founder Sam Crombie and link to a landing site and company profiles YC listing, LinkedIn, Founder X.

Today there is no clearly accessible public product: no open beta, pricing page, product docs, or demos are visible, and the footprint looks founder-led and early-stage rather than a launched consumer app YC listing, LinkedIn, Founder X.

Based on the positioning, a plausible current focus is building source ingestion, citation and provenance, answer generation, and a search/UI layer, with internal tests or invite-only trials. This is an inference from the AI encyclopedia concept and founder posts, not confirmed product documentation Founder X.

Who are their target customer(s)

• Researchers and students: They need quick, readable summaries with reliable citations. They lose time stitching together facts across papers and web pages and worry AI answers won’t point to trustworthy sources.
• Journalists and analysts: They need fast, citable background briefings. Current tools return scattered or paywalled results, and many AI answers hallucinate or omit sources.
• Developers and startups: They want a drop-in knowledge API for apps and assistants. Building their own ingestion, search, and attribution stack is costly and error-prone.
• Fact-checkers and subject-matter editors: They need an auditable, editable reference set. Manual verification across many sites and tracking provenance is slow and prone to mistakes.
• Internal knowledge teams at enterprises or universities: They need a single, searchable, canonical knowledge base. Documentation is fragmented across wikis, chat, and drives, making accuracy, discovery, and citation difficult.

How would they acquire their first 10, 50, and 100 customers

• First 10: Founder-led outreach to trusted domain experts (researchers, senior journalists, knowledge-team leads) via YC and personal networks; run free, time-boxed pilots that ingest their sources, co-author a small set of entries, and let them directly review citations and accuracy YC listing.
• First 50: Convert successful pilots into a closed beta cohort; add similar teams via referrals, targeted outreach to university departments and newsroom innovation desks, and a few focused demos/webinars. Standardize the pilot playbook (ingestion checklist, review workflow, success metrics) to lower onboarding cost.
• First 100: Offer a limited self-serve or institutional pilot with onboarding docs, API keys, and packaged ingestion tools; begin charging pilot fees or small subscriptions. Drive growth with integrations (citation managers, Slack/wiki connectors), early case studies, and targeted conferences.

What is the rough total addressable market

Top-down context:

Leviathan targets spend across knowledge management software (~$20B in 2024), enterprise search (~$4.9B in 2023), higher-education technology (tens of billions), and academic publishing and research platforms (tens of billions), acknowledging heavy overlap among these segments GVR KM, GVR Enterprise Search, GVR EdTech, DataIntelo Academic Publishing.

Bottom-up calculation:

Start with the knowledge-management market as the core target (~$20B), add 30–50% of enterprise search spend (~$4.9B) for AI encyclopedia use cases, and include a modest slice of higher-ed and academic publishing budgets to reach a realistic near-term TAM of roughly $10–30B, rather than summing full market sizes GVR KM, GVR Enterprise Search, GVR EdTech, DataIntelo Academic Publishing.

Assumptions:

• Significant overlap exists between KM, enterprise search, and higher-ed/publishing buyers, so double-counting is avoided by taking partial shares rather than full sums.
• Initial focus is on citable reference and API use cases within enterprises, universities, and research platforms rather than the entire consumer or general EdTech markets.
• Winning enterprise and institutional budgets requires solving provenance, auditability, and licensing, which gates how much of each market is realistically addressable in the next 3–5 years.

Who are some of their notable competitors

• Perplexity: AI answer engine that searches the web in real time and returns concise answers with numbered citations; offers an API and enterprise features for embedding citation-forward answers Perplexity help.
• Consensus: AI search focused on peer‑reviewed literature and evidence-backed summaries, targeting academic and research workflows Consensus homepage.
• Elicit: AI research assistant for literature reviews and data extraction from scientific papers, competing for researcher synthesis and sourcing tasks Elicit product.
• ChatGPT / OpenAI: General-purpose conversational AI with web search, browsing, deep research, source links, and enterprise/API options; competes on breadth and developer integrations OpenAI ChatGPT Search, Capabilities overview.
• Wikipedia: Human‑curated, non‑commercial encyclopedia and default reference for many users; competes on trust, breadth, and editorial provenance Wikipedia.