Fivetran

Growth decel + connector commoditization + merger friction compress to ~5-8x ARR (~$3-5B) - a flat-to-down mark versus 2021.

~10-13x combined ARR (~$6-8B)

roughly reaffirming the 2021 $5.6B mark to modestly above, reflecting growth offset by commoditization risk and multiple reset.

Combined ~$600M ARR growing ~30%+, sole independent open full-stack data-infra leader in the AI-agent wave -> ~15-20x ARR (~$9-12B+), IPO-premium case.

• Category leader in managed automated ELT with 600+ pre-built, maintained connectors (company-stated) - deep breadth and reliability moat
• Large blue-chip and AI-native customer base (OpenAI, HubSpot, Zendesk, Coupa, DocuSign) with usage-expansion economics
• dbt merger unifies ingestion + transformation + semantic/governance into one open stack - rare full-pipeline coverage; ~80-90% of Fivetran customers already used dbt
• Company-stated ~50% YoY ARR growth to ~$300M standalone, sustained new-logo and expansion motion
• Backed by top-tier investors (a16z, General Catalyst, Matrix, ICONIQ, D1, YC Continuity)

• Consumption/MAR pricing is widely criticized as expensive and hard to forecast; Airbyte claims materially lower cost on typical workloads - price is a leading churn/objection vector
• No public financial transparency — gross margin, NRR, operating margin, GAAP P&L all undisclosed; opacity limits diligence
• COGS exposed to cloud compute/egress inflation on high-volume CDC pipelines
• Integration risk from a large all-stock merger with dbt Labs — dual leadership (Fraser CEO / Handy President), product + culture merge execution
• Long-dated $5.6B (2021) mark predates the SaaS-multiple reset — internal marks/liquidity for employees uncertain

• Position as the trusted, governed 'data layer for AI agents' — agentic workloads need continuous, fresh, governed data (secular tailwind and the explicit strategic framing of the merger)
• Cross-sell/upsell dbt transformation + semantic layer into Fivetran's ingestion base and vice versa
• Displace legacy ETL freed up by the Salesforce/Informatica acquisition (customer uncertainty during integration)
• Open-standards positioning (avoid warehouse lock-in) as a wedge against Snowflake/Databricks native ingestion
• Eventual IPO of a ~$600M+ ARR combined data-infra company if public markets reopen

• Warehouse-native ingestion from Fivetran's own destination partners is the most direct structural threat: Databricks Lakeflow Connect now advertises 100+ built-in managed connectors with Salesforce, Workday and SQL Server generally available, and Snowflake built Openflow on its Datavolo (Apache NiFi) acquisition — meaning the two platforms Fivetran loads into can bundle ingestion into their own contracts.
• Hyperscaler and platform bundling compounds this, since Snowflake and Databricks can absorb ingestion cost into an existing consumption commitment while Fivetran must be justified as a separate line item; a single competitor-vendor blog account (Nexla, undated, anecdotally sourced) reports Databricks held analyst briefings within two weeks of the merger and that Snowflake did the same, but the structural bundling argument does not depend on that timing claim.
• Open-source ELT caps Fivetran's pricing power at the low and mid market: Airbyte's self-hosted edition is free and, per third-party 2026 comparisons, Airbyte Cloud undercuts Fivetran up to roughly 300M MAR/month, so Fivetran's consumption pricing is defensible mainly at high volume and on hard-to-build connectors.
• Fivetran's MAR (monthly active rows) pricing counts distinct source primary keys synced per calendar month, so high-churn tables inflate spend regardless of net data growth; and since the March 2025 pricing change, volume scaling is applied per connection rather than aggregated at the account level, removing account-wide volume discounts for multi-connector estates and giving warehouse-native and open-source alternatives a ready displacement argument at renewal.
• Post-merger integration risk is unusually high given three acquisitions announced within five months — Census (reverse ETL, announced May 2025), Tobiko Data/SQLMesh (September 2025) and dbt Labs (announced October 13 2025, closed June 1 2026) — leaving overlapping transformation engines (SQLMesh and dbt Fusion) and go-to-market motions to rationalise while competitors sell against the uncertainty.
• dbt Core's perceived vendor neutrality — the merger's most valuable asset, with 100,000+ combined data teams — is at risk now that the de facto open transformation standard is owned by an ingestion vendor; third-party accounts of Coalesce (October 2025) report community concern over pricing power, vendor lock-in and dbt Core's long-term future as engineering investment shifts toward dbt Cloud and Fusion, and a credible community fork or migration would erode the strategic rationale for the deal.