Delphina funding, valuation and investors

Product & use cases

Illustrative buyer tour for a finance operations leader asking why net revenue retention fell: (1) The data owner connects a read-only Snowflake, Databricks, BigQuery or other supported warehouse using role-scoped credentials; (2) imports dbt definitions, metric dashboards and written policies via documented connectors/MCP; (3) Delphina's Context Factory agents propose definitions and surface ambiguities; (4) relevant subject matter experts approve consequential changes in a versioned shared knowledge base; (5) an executive asks a question in Delphina, Slack or a supported MCP client and receives an answer tied to approved metrics, tables and context; (6) changes in data trigger reviewed updates instead of leaving a stale semantic layer. These are documented supplier capabilities, not a hands-on VCT test. Free tier is one user and public datasets/files only; Pro $250/month + metered consumption includes $300 monthly credits and one connected warehouse; Enterprise is quotation-based with additional controls, single-tenant/BYOC add-ons. Public prices are not audited contract economics. Product does not guarantee that source data or human-approved definitions are correct.

• CFO/finance leader investigates retention, margins, customer-acquisition economics using reviewed metric definitions instead of manually requesting a dashboard.
• Data team reviews and versions the business glossary, trusted tables, access policies and AI-proposed changes; it remains accountable for accuracy.
• Operations, growth and sales leaders request reproducible answers from structured company systems through Delphina, Slack or MCP-connected agents.

Key facts

• November 12, 2026: $7.5M newly announced financing, co-led by Radical Ventures and Costanoa, bringing total disclosed financing to $15M; formal round stage not disclosed.
• December 13, 2023: separate $7.5M seed co-led by Costanoa and Radical Ventures, according to Delphina's original issuer announcement.
• Company reports 73% of platform questions now come from business teams and CEOs at more than half of customers are active users; customer-usage metrics, not audited revenue.
• Public Nov 2026 pricing: free Community for one user; Pro $250/month plus consumption, 10 users and $300 monthly usage credits; Enterprise custom.
• Issuer names Substack, LATAM Airlines, Coursera, Upstart and others as customers; testimonials are vendor-selected evidence, not controlled proof of ROI.

Competitive landscape

Edge: Governed semantic/business-context upkeep performed by agents, with human review and shared, version-controlled metric definitions. The buyer value is reduced analyst-request queue and more trustworthy self-serve answers; data quality, auth and maintenance remain material constraints. Vendor-reported adoption is not proof of revenue or independent accuracy improvements.

Two $7.5M tranches separated by almost three years total $15M; the Nov 2026 financing must not be described as $15M new capital. Unknown: 2026 round formal stage, valuation, absolute paying customer count, ARR, gross margin and retention. Delphina claims adoption and selected customer benefits; 73% of question volume from business users is not share of paying seats. Buying the $250 Pro plan does not include unlimited model and compute usage. Validate permission scope, prompt-injection/data leakage risk, metric definitions, approvals, credit metering, exportability and data residency.

• ThoughtSpot direct

  Search and natural-language analytics; compare governance, semantic modeling, time-to-answer, permissions and enterprise deployment.

• Microsoft Power BI Copilot incumbent

  Natural-language analytics within existing Power BI tenant; often easier for Microsoft-centric customers but separate model/domain work is still needed.

• Looker / LookML adjacent

  Strong managed semantic definitions and BI delivery; more human-defined modeling than Delphina's claimed agent-maintained context.

• dbt Semantic Layer adjacent

  Governed metrics-as-code foundation rather than a complete AI question-answering product.

• Seek AI alternative

  Conversational access to business data; compare reliability, approval workflows and cost under identical data questions.

Market / IPO context