$6 billion in legal-tech funding last year. Harvey at $11B. EvenUp at $2B. Bench IQ at $5.3M for federal bankruptcy. Trellis ships state-court data pipes to BigLaw. None of it lands where Sidebar lands: Florida civil, at 3–25-attorney firms, with a dossier on both sides of the bench.
Four dimensions where every legal-AI raise is making a bet. No incumbent is executing in all four at once — the gaps are where Sidebar wins.
| Dimension | Enterprise / BigLaw lane | Sidebar's lane |
|---|---|---|
| Court system | Federal (Bench IQ, Pre/Dicta, Lex Machina) | Florida state trial courts |
| Side of bench | Judges or attorneys | Judge × Attorney graph — both, bundled |
| Output shape | Probability scores / research dataset | Editorial dossier — facts, not ratings |
| Price / ICP | $3K–$14K / seat / yr (AmLaw 100) | $149 / seat / mo — 3–25 attorney FL firms |
$5.3M seed · Aug 2025 · Battery + Inovia · total $7.4M. Federal bankruptcy only. AmLaw 100 buyers. Toronto HQ. Ships judges only.
Their money helps us: $5.3M commits them to federal + BigLaw for 18 months. They cannot pivot into state-civil + attorney graph + sub-$200 seats without rebuilding the company. We are asking 28% of their seed for a market 10× the size.
Claims 85% accuracy predicting federal MTD rulings. Inputs include judge demographics (age, gender, political affiliation) — which is also their regulatory risk. We ship facts, not predictions built on demographics. Rule 4-7.13 / ABA 2.3(B) safe; Pre/Dicta's not.
$4.2M seed May 2025 + $2.2M pre-seed Nov 2024. Plaintiff-side intake triage — a totally different buyer from trial counsel. No judge intel. Different product shape entirely.
Established. Expanded to FL / TX / NY. 21 states of state-trial-court data. API-first.
Why they're dangerous: they already cover Florida civil dockets. That's the one part of our thesis that isn't unique.
Why we still win: Trellis sells data pipes — API, CSV export, search UI — to litigation-support teams and BigLaw analytics groups. Sidebar sells finished editorial dossiers to trial lawyers at 3–25-attorney firms. Their output is a dashboard; ours is a PDF a litigator reads walking into the hallway before argument. Different buyer, different price point (their ~$5–10K/seat vs. our $149/mo), different deliverable.
Mitigation: attorney graph is the moat. Trellis has zero attorney dossier inventory. Ship the two-sided graph in 90 days and the mimicry window closes.
UniCourt = records aggregator, generalist. Lex Machina (LexisNexis) = federal-first, $10K+ seats, BigLaw. vLex / Fastcase = 850M-record research tool — research UX, not dossier shape. Docket Alarm = alerts, not intelligence. Each useful; none is shaped like Sidebar.
The rails we price against:
| Company | Round / Valuation | Lane | Price point |
|---|---|---|---|
| Harvey | $11B valuation · Mar 2026 | Enterprise legal LLM | $1K–1.2K/seat/mo |
| Legora | $5.5B valuation | Enterprise workflow | $3K/seat/yr · 10-seat min |
| Paxton | $22M Series A · early 2025 | Legal research assistant | $159/mo Pro |
| Eve Legal | $103M · $1B valuation | Plaintiff-bar AI | Enterprise |
| EvenUp | $150M Series E · Oct 2025 · $2B | Personal injury settlement AI | Enterprise (20% of top 100 PI) |
| Supio | $60M Series B · Apr 2025 | PI medical-record AI | Enterprise |
| Centari | $14M Series A | Contract review | Enterprise |
| DeepIP | $25M | IP drafting | Enterprise |
| DeepJudge | Undisclosed | Opposing-counsel negotiation intel | Enterprise |
| Antes | $4.6M seed · Apr 2026 | Legal↔engineering bridge | SMB |
| Vesence | $9M seed | AI legal workflow | SMB / mid |
| Felix | $1.7M pre-seed · Apr 2026 | AI workflow (ex-Parrot) | SMB |
| Sidebar | $1.5M seed · Apr 2026 | FL civil judicial + attorney dossiers | $149/seat/mo |
Market context: Legal tech raised $6B in 2025. 2026 tracking to $4.3B across 356 deals YTD (70% AI-powered). Industry framing: "2026 is the year legal AI stops being about neat tools and starts being about operating infrastructure."
Enterprise tier ($3K–$14K/seat/yr) is saturated and venture-rich. SMB tier (~$100–200/mo) has one serious legal-AI player — Paxton — and it's a horizontal research assistant, not a judicial dossier product. Sidebar lands at Paxton-adjacent price with a vertical product a generalist tool can't replicate.
Legal AI raised $6B in 2025. None of it was allocated to the 4,300 small Florida civil firms doing 95% of the actual hearings. Bench IQ went federal at $5.3M. Harvey went enterprise at $11B. Trellis sells API pipes to BigLaw. We ship the one thing every FL civil litigator needs in the 40 minutes before a hearing — a dossier on the judge and the lawyer across the table — for the price of a bar-prep subscription.
| Risk | Severity | Mitigation |
|---|---|---|
| Trellis launches FL small-firm tier | HIGH | Ship attorney graph in 90 days; API DNA is wrong shape for dossier output |
| Paxton adds judicial analytics module | MEDIUM | Paxton = horizontal research; judicial intel is a specialty vertical |
| EvenUp/Supio PI war chest expands into general civil | MEDIUM | PI AI = settlement math, not judge intel; different core model |
| Bench IQ pivots to state & lowers price | LOW | Their $5.3M commits federal + BigLaw for 18+ months |
| Thomson Reuters v. ROSS-style ingestion lawsuits | MEDIUM | Corpus = public court data + FL Bar. No editorial-source ingestion |
| LexisNexis bundles state analytics into Lex Machina | MEDIUM | Their model is federal-native; FL state expansion 2–3 yrs out |