Ema's $77M Raise and the "SaaS Becomes a Database" Replacement Thesis
The displacement thesis got its best-funded articulation yet this cycle: Ema, a two-year-old enterprise agent platform founded by ex-Google/Coinbase executive Surojit Chatterjee and ex-Okta executive Souvik Sen, raised a $77M Series B led by Creaegis (with Accel, Section 32, and Prosus increasing stakes) on September 23, 2026, bringing total funding to $140M at a valuation TechCrunch reports "more than quadruples" its 2024 round (TechCrunch, Sept 23, 2026).
The replacement claim, verbatim
Chatterjee's pitch is not "copilot for your SaaS" but replacement of it: "Many of our customers are already on the way to replace [large SaaS applications] completely, removing dependency on them, because they are mostly becoming like a database." The mechanism is "wrap, then replace": Ema first wraps an enterprise's existing applications, then customers reduce dependence on some and replace them altogether. This is the operational version of the buyer-side displacement Qualitate quantified the day before (see Buyer-Side Evidence: Qualitate's Full AI Disruption Report — HubSpot Most-Displaced Vendor, ServiceNow/Workday/Adobe in the "Flat-Budget Waiting Room," Anthropic the Top Winner) — and of the SaaStr argument that legacy products are now inferior to AI-native alternatives (Legacy B2B Software Products Now Actively Worse Than AI-Native Alternatives).
Traction metrics (company-reported)
- 50+ active enterprise deals; 1M+ active enterprise users; 5M+ actions/queries handled
- Customers include NTT DATA, Hitachi, ADP, PwC, Google, KPMG, Wipro, and Microsoft
- Revenue up 50-fold in two years; bookings >$150M (multiyear TCV including 2–3 year deals, not ARR — Chatterjee declined to disclose run-rate)
- ~180% net dollar retention; >90% of customers expanded beyond their initial use case
- ~80% gross margins with ~200 employees — Chatterjee says margins improve as deployments need less human support
Two structural signals beyond the raise
- Pricing is outcome-based, not seats or tokens: Ema charges "tied to the completion of tasks and business outcomes" — the same pivot incumbents are being forced into (The Outcome-Based Pricing War: HubSpot's Breeze AI Shift, GitLab's Flex Consumption, and Progress/Domo's Seat-Churn Admission), but natively, without a seat-revenue base to cannibalize.
- It eats services too: "A lot of the services companies are working with us… They are also dramatically changing or disrupting their own business models because they understand the human-forward model may not be the best model going forward." This extends the AI-labs-vs-consultancies dynamic (OpenAI and Anthropic Launch Enterprise Deployment Firms, Directly Challenging Consulting Giants) down to the startup layer — and Chatterjee notes Ema draws on 150+ models, so frontier-lab progress is an input, not a threat.
What it means
An outcome-priced, model-agnostic agent layer with $150M+ in bookings and 180% NDR is now commercially validated venture scale. For buyers evaluating vendors, the "wrap-then-replace" pattern means incumbent contracts can be hollowed out (incumbent demoted to database-of-record) before any headline churn shows up — consistent with Qualitate's finding that spend decreases (27% AI-driven) run ahead of churn (18%).