Databricks Public-Source Startup Diligence Report
The core thesis is that Databricks is expanding from a lakehouse pioneer into a broader data-and-agent operating layer, using open formats, multi-cloud distribution, and acquisition-led product expansion to widen its moat. The core counter-thesis is that this expansion now depends on ambiguous late-stage financing, hyperscaler coopetition, and legal/IP exposure that could constrain margins, strategic freedom, or valuation durability.
Databricks Public-Source Startup Diligence Report
Databricks looks like one of the strongest late-stage private data-and-AI platforms visible from public sources: real product breadth, credible enterprise scale, strong open-source lineage, and a demonstrated ability to raise capital and ship new categories. The counterweight is that core financial quality, partnership economics, and several material IP risks remain unresolved without a data room.
The Makkai case links DBRX and Mosaic model development to allegedly pirated books, and public reporting says the court denied Databricks’ motion to dismiss while damages theories remain very large.
Diligence request: Request outside-counsel assessment, reserve analysis, insurance coverage, and Mosaic acquisition diligence on training-data provenance.
Azure Databricks is a first-party Microsoft service while Microsoft Fabric is a competitor; Databricks also depends on AWS, GCP, and model providers for distribution and infrastructure.
Diligence request: Review all major cloud and model-partner contracts for revenue sharing, MFN terms, exclusivity, and termination or change-of-control rights.
Snowflake, Microsoft Fabric, AWS, BigQuery, and Palantir all market overlapping data, AI, and governance capabilities, which can compress Databricks pricing or force bundling.
Diligence request: Request segment-level win-loss data, pricing concessions, and churn drivers versus named competitors.
The official Series J announcement uses unusual non-dilutive language while also funding employee liquidity and taxes, leaving open questions on primary cash, secondary flow, structured financing, and preference overhang.
Diligence request: Obtain the Series J term sheet, cap table, liquidation preferences, and primary-versus-secondary allocation.
Public revenue and margin signals are management disclosures without audited statements, backlog, deferred revenue, or AR aging detail.
Diligence request: Request audited historical financials, revenue-recognition policies, deferred revenue, ARR definitions, and gross-margin bridges.
Public materials show impressive logo breadth and $1M-plus customer counts but do not disclose top-account share, concentration, GRR, or NRR.
Diligence request: Request top-10 customer concentration, gross and net revenue retention, and renewal trends by cohort and geography.
Patent suits by R2 Solutions and Byteweavr, plus the Databricks v. Weisfield matter, suggest a recurring IP-defense burden and former-employee dispute risk.
Diligence request: Request the IP litigation schedule, trade-secret inventory, and any settlement or indemnity obligations tied to active disputes.
Databricks is integrating MosaicML, Arcion, Tabular, Neon, Anthropic models, Lakebase, and Agent Bricks at once, increasing roadmap complexity and dependency on external model and infrastructure ecosystems.
Diligence request: Request attach rates, integration milestones, sunset plans, and contingency architecture for major partner-model dependencies.
Databricks’ public financial story is unusually rich for a private company on financing and growth markers, but still thin on the documents that matter most for diligence. The public record supports a rapid scaling narrative and a striking capital-raising history, while leaving cap-table mechanics, balance-sheet quality, and earnings quality unresolved.
not publicly verifiable confidence: low
Public sources give Databricks unusually good scale markers for a private company — including ARR, revenue, margin, and customer-count disclosures — but not audited statements, backlog schedules, AR aging, or debt detail. The financial picture is therefore directionally informative yet still insufficient for true earnings-quality diligence.
| period | public metric | source | verification status | diligence caveat |
|---|---|---|---|---|
| 2018 | Exceeded $100M ARR | Series E press release | verified | No audited 2018 statements or revenue-recognition detail were public. |
| 2021-02 | $425M+ ARR; 75%+ year-over-year growth | Forbes Series G coverage | verified | Third-party reporting still depends on management access; no audited financial package was available. |
| 2021-08 | $600M+ ARR; 75% annual growth | Forbes Series H coverage | verified | No public cohort, retention, or gross-margin bridge accompanies the ARR number. |
| 2023-07 | Crossed $1.5B revenue run rate; 85% non-GAAP subscription gross margin; >300 $1M+ ARR customers | Series I press release | partially_verified | Metrics are management-stated and non-GAAP; no audited income statement or working-capital view is available. |
| FY2024 ended 2024-01-31 | Reached over $1.6B revenue with 50%+ growth | NVIDIA June 2024 partnership press release | partially_verified | The statement confirms scale but still does not supply audited statements, segment margin, or tax detail. |
| Q4 FY2025 guidance | Expected to cross $3B revenue run rate and become free-cash-flow positive for the first time | Series J press release | partially_verified | This is forward-looking guidance and not a confirmed actual result in the reviewed public set. |
Public scale visibility is unusually good for a private company, but revenue quality remains a data-room issue.
partially verified confidence: medium
Databricks does not publicly disclose multi-year projections, but management does provide short-horizon guidance around run-rate revenue, growth, and free cash flow. Those signals support momentum, not forecast reliability.
partially verified confidence: medium
The public record is strong enough to map Databricks’ board and investor mix, but not its actual equity capitalization, option pool, or liquidation preferences. Series J makes the structure more rather than less important.
| stakeholder | public position | public evidence | diligence caveat |
|---|---|---|---|
| Founders and senior management | Seven founders remain publicly listed; Ali Ghodsi is CEO | Founder and leadership pages identify founders and current operating roles. | No public founder share counts, vesting, or dilution history were verified. |
| Board | Founder-led board with a16z and NEA investor representation plus independent directors | Board page lists Ion Stoica, Ali Ghodsi, Matei Zaharia, Ben Horowitz, Pete Sonsini, Elena Donio, Jonathan Chadwick, and Scott Shenker. | No public committee structure, observer rights, or reserved-matter schedule was verified. |
| Strategic investors | Microsoft, NVIDIA, AWS-linked, and Alphabet-linked participation is public | Series E, Series I, and Series G/H materials name Microsoft, NVIDIA, AWS, and CapitalG as investors or strategic backers. | Investment size, side letters, information rights, and commercial tie-ins remain private. |
| Employees and former employees | Series J explicitly funds liquidity and related taxes | Series J press release says the capital is expected to provide liquidity for current and former employees and pay related taxes. | No public tender mechanics, strike prices, employee pool size, or preference stack was verified. |
| Debt and other instruments | not_publicly_verifiable | Reviewed public sources did not confirm any current debt schedule or off-balance-sheet liability inventory. | Request debt instruments, venture debt, revolvers, guarantees, and any off-balance-sheet commitments. |
The governance picture is public enough for an initial map, but the economic rights remain largely opaque.
partially verified confidence: medium
Public financing history is unusually visible, and Databricks has clearly used capital to fund product expansion and employee liquidity. The public record is much weaker on taxes, accounting policy, debt, and the exact economics of later rounds and acquisitions.
| date | round or event | public amount or term | valuation | lead or counterparties | verification status | diligence caveat |
|---|---|---|---|---|---|---|
| 2019-02-05 | Series E | $250M equity financing | $2.75B | Andreessen Horowitz; Coatue; Microsoft; NEA | verified | Need full stock terms and any side letters tied to Microsoft commercial rights. |
| 2021-02 | Series G | $1B equity financing | $28B post-money | Franklin Templeton; AWS; CapitalG; Salesforce Ventures; Microsoft | verified | Historical round terms come from archived independent coverage rather than Databricks PRs retrieved in-session. |
| 2021-08 | Series H | $1.6B equity financing | $38B | Morgan Stanley; Baillie Gifford; ClearBridge; UC Investment Office | verified | Need capitalization, preference, and employee-liquidity mechanics for the round. |
| 2023-09-14 | Series I | Over $500M; $73.50 per share | $43B | T. Rowe Price; NVIDIA; Ontario Teachers; Capital One Ventures | verified | Need exact round size, class rights, and strategic-investor restrictions. |
| 2024-12-17 | Series J | $10B expected financing; $8.6B completed at announcement | $62B | Thrive; a16z; DST Global; GIC; Insight; WCM | partially_verified | Need primary-versus-secondary split, any debt component, and the legal meaning of non-dilutive. |
| 2025-2026 | Later private-market pricing references | Paywalled or indirectly cited later-round references only | not_publicly_verifiable | Undisclosed or paywalled later syndicates | inconclusive | Confirm whether any post-Series J round has actually closed and at what price. |
Later valuation references beyond Series J were not independently validated in-session and are intentionally presented as a diligence gap rather than a settled fact.
| event | date | disclosed value | strategic rationale | financing implication | verification status |
|---|---|---|---|---|---|
| MosaicML acquisition | 2023-06-26 | Approx. $1.3B inclusive of retention packages | Bring generative AI model training and ownership onto the Databricks platform | Largest disclosed acquisition in the reviewed set; likely accelerated later capital needs and now anchors current copyright risk. | verified |
| Arcion acquisition | 2023-10-23 | Over $100M inclusive of incentives | Add real-time replication and CDC into the Lakehouse Platform | Supports ingestion expansion and shows continued tuck-in M&A spending beyond flagship AI deals. | verified |
| Tabular acquisition | 2024-06-04 | Price not disclosed in Databricks PR | Bring Delta Lake and Apache Iceberg interoperability closer together | Strategically important format-unification move; actual cash outlay remains a diligence gap. | partially_verified |
| Neon acquisition intent | 2025-05-14 | Price not disclosed in Databricks PR | Add serverless Postgres built for AI-agent workloads | Expands Databricks into operational databases and agent-native application infrastructure. | partially_verified |
| Series J stated capital uses | 2024-12-17 | New AI products, acquisitions, international GTM expansion, and employee liquidity | Support accelerated AI demand and company scaling | Signals that not all headline financing converts into balance-sheet cash available for operations. | verified |
This is a selected disclosed set rather than a complete acquisition register; private side letters, earnouts, and integration budgets remain unreviewed.
A sparse valuation chart is still useful because it frames what later pricing remains unverified.
Databricks clearly operates a real multi-product platform rather than a single analytic tool. The challenge for diligence is less about whether the products exist and more about which of them produce durable economics and defensible differentiation.
verified confidence: high
Databricks now publicly presents a broad stack that spans engineering, warehousing, AI agents, governance, and operational databases on top of an open lakehouse foundation. Product existence and current scope are well supported; product economics, attach rates, and relative maturity are not.
| product | audience | key features | public evidence | verification status |
|---|---|---|---|---|
| Lakeflow | Data engineering and platform teams | Jobs orchestration, declarative pipelines, ingestion connectors | Pricing and product pages show Lakeflow Jobs and related engineering SKUs. | verified |
| Databricks SQL | Analytics engineers, BI users, and warehouse buyers | Classic, Pro, and Serverless SQL warehouses | Pricing pages and product descriptions confirm SQL analytics as a major warehouse SKU. | verified |
| Mosaic AI / Agent Bricks | ML, GenAI, and application teams | Agent development, model serving, training, evaluation, vector search, governance | AI page shows Agent Bricks, model serving, training, evaluation, and customer metrics. | verified |
| Lakebase | Application developers and AI-agent builders | Serverless Postgres with autoscaling, branching, and Unity Catalog integration | Lakebase product and pricing pages describe operational database positioning for AI apps. | verified |
| Unity Catalog and open formats | Security, governance, and platform teams | Universal governance, lineage, and interoperable open lakehouse formats | Open-source and platform pages position Unity Catalog, Delta Lake, and Apache Iceberg as part of the platform foundation. | verified |
Public product breadth is clear; what remains unresolved is product-level monetization and usage mix.
| workload | databricks public price | public competitor anchor | comparability note |
|---|---|---|---|
| Batch and pipeline engineering | $0.15 / DBU classic; $0.35 / DBU serverless for Lakeflow Jobs | AWS markets a purpose-built analytics suite rather than a single unified engineering SKU | Databricks publishes per-DBU list prices; AWS packages multiple services with separate billing. |
| SQL warehousing | $0.22 / DBU classic; $0.55 / DBU pro; $0.70 / DBU serverless | Snowflake describes a serverless managed platform with governance and consumption billing | Snowflake public pricing and warehouse sizing were not normalized in-session, so this is a positioning comparison not a TCO model. |
| AI model serving and agents | $0.50 /M input tokens, $1.50 /M output tokens, plus hourly serving rates on reviewed public pricing pages | BigQuery and Microsoft Fabric position AI and agent workflows as integrated into the core analytics platform | Cross-vendor AI pricing is highly workload-specific and public list prices were not normalized across token, compute, and platform charges. |
| Operational database | $0.092 / capacity unit hour and $0.345 / GB-month for Lakebase | Palantir Foundry markets an AI-powered operating system rather than a Postgres-native operational database | Lakebase is directly database-oriented; Palantir is a broader operational platform, so overlap is strategic rather than apples-to-apples pricing. |
| SAP Business Data Cloud shared processing | Normal 1.25x DBU consumption uplift; temporary 1.0x promo through 2026-08-31 | not_publicly_verifiable | The reviewed public evidence confirms SAP-specific Databricks pricing terms, but no directly comparable partner pricing from peers was validated in-session. |
All Databricks rates are public list pricing and likely differ materially from committed-use contract economics.
Databricks has strong public customer and partner breadth, but not the concentration and monetization detail that most investors or acquirers would want before underwriting a late-stage platform company. The customer story is credible; the revenue-quality story is still private.
partially verified confidence: medium
Public customer evidence is strong on breadth and named case studies across industries, but still weak on monetization quality and repeatability. The most credible public examples come from named customers with attributed executives and specific metrics.
| customer | industry | use case | public outcome | verification status |
|---|---|---|---|---|
| AT&T | Telecommunications | Fraud detection and operational decision support | Up to 80% fraud reduction with 100+ ML models in production | partially_verified |
| Virgin Australia | Travel and aviation | Real-time analytics, ML deployment, baggage operations | 75% increase in near real-time data availability and 44% reduction in mishandled bags | partially_verified |
| Block | Financial technology | AI-agent productivity system for seller operations | $10M in productivity gains claimed on Databricks AI page | partially_verified |
| Comcast | Media and communications | Personalized viewer experience and voice commands | 96% response accuracy and 10x cost reduction claimed on Databricks AI page | partially_verified |
| UK and ANZ regional logos | Cross-industry | Regional proof of adoption in regulated and enterprise sectors | Public releases name Unilever, Rolls Royce, Nationwide, Virgin Atlantic, Airwallex, Atlassian, NAB, Telstra, and Queensland Health | partially_verified |
These cases prove buyer diversity and use-case breadth, but not concentration or renewals.
partially verified confidence: medium
Strategic relationships are central to Databricks’ distribution and product strategy. Microsoft, Anthropic, SAP, and the broader partner network are real strengths, but they also concentrate channel and model dependence.
| partner | relationship type | public evidence | revenue or distribution signal | gap |
|---|---|---|---|---|
| Microsoft / Azure Databricks | First-party cloud distribution and strategic investment | Azure Databricks is a first-party Azure service and Microsoft joined Series E. | Microsoft-controlled billing makes Azure a major channel rather than just infrastructure. | Exact revenue sharing, pricing control, and Fabric-related conflict terms were not public. |
| Anthropic | Five-year model partnership | Claude is offered natively on the Databricks platform across AWS, Azure, and GCP. | Adds frontier model breadth and potentially deepens AI-agent adoption. | Public economics and any minimum commitments are not disclosed. |
| SAP Business Data Cloud | Data-sharing and processing integration | Databricks publishes explicit SAP-related DBU pricing and promotional uplift terms. | Suggests SAP-originated data and AI workflows could become a distribution vector. | Revenue sharing and exclusivity terms were not public. |
| Global partner ecosystem | Cloud, ISV, and consulting ecosystem | Databricks says it has 1,200+ global cloud, ISV, and consulting partners. | Large partner ecosystem can multiply implementation and co-sell reach. | Public materials do not disclose partner-sourced revenue or the top partner mix. |
| BrickBuilder / partner-led delivery | Partner program and go-to-market enablement | Databricks public materials describe partner programs and ecosystem routes alongside direct sales. | Supports enterprise implementations and global reach beyond direct field sales. | Public partner tiers and economics remain limited. |
The public record shows a real partner moat, but commercial dependencies may be more important than the headline ecosystem size suggests.
The purpose of this chart is to expose the disclosure gap, not to imply a precise concentration model.
not publicly verifiable confidence: low
Public sources support broad customer adoption but do not support a reliable concentration analysis. This section is therefore an evidence gap rather than a clean concentration bill of health.
not publicly verifiable confidence: low
No material severed customer, partner, or supplier relationships were directly verified in the reviewed public set. That absence should be treated as a gap, not as proof of clean relationship history.
partially verified confidence: medium
Databricks’ supplier picture is best understood as multi-cloud and model-platform co-dependency rather than single-supplier concentration. Microsoft, AWS, GCP, NVIDIA, and partner models all matter in different ways.
| supplier | role | public evidence | dependency risk | verification status |
|---|---|---|---|---|
| Microsoft Azure | Cloud platform plus first-party Azure Databricks channel | Azure Databricks is a first-party service and Microsoft joined Series E. | Microsoft controls billing and also competes via Fabric. | verified |
| AWS | Cloud infrastructure, marketplace route, and analytics/AI competitor | AWS invested in Series G and markets a broad analytics plus SageMaker stack. | Databricks competes with AWS services while also relying on AWS infrastructure and marketplace reach. | verified |
| Google Cloud | Cloud infrastructure and adjacent data-plus-AI competitor | CapitalG invested in Series G and BigQuery markets an autonomous data-to-AI platform with Iceberg support. | Open-format interoperability may lower switching friction while keeping Google competitive. | verified |
| NVIDIA | GPU, model-serving, and accelerated compute partner | Databricks publicizes DGX Cloud training, Photon CUDA acceleration, and NVIDIA strategic investment. | Supplier pricing power can influence Databricks AI margins. | verified |
| Anthropic and other model providers | Third-party model access through Databricks AI workflows | Databricks markets Anthropic Claude and partner-model availability on its platform. | Partner-model concentration can weaken Databricks control over roadmap and pricing. | partially_verified |
The most material public dependency is not a single supplier but the channel and infrastructure leverage of hyperscalers and model partners.
Databricks no longer competes only with warehouse vendors; it competes across data engineering, governance, AI platforms, agent frameworks, and operational AI systems. That enlarges the opportunity set and intensifies pricing and bundling pressure.
partially verified confidence: medium
Databricks competes across several overlapping markets at once: warehouse, lakehouse, AI platform, governance, and operational data apps. Public product pages make the overlap obvious even without neutral market-share data.
| competitor | segment | product overlap | differentiator | source |
|---|---|---|---|---|
| Snowflake | AI Data Cloud and cloud data platform | Warehouse, governance, AI, app development, and multi-cloud data platform | Serverless managed platform with governance, FinOps, and observability | Snowflake platform page |
| Microsoft Fabric | End-to-end analytics SaaS over OneLake | Data engineering, data science, warehouse, real-time analytics, databases, and governance | Tight Microsoft 365 and OneLake integration plus SaaS packaging | Microsoft Fabric overview |
| AWS analytics and SageMaker | Purpose-built services plus integrated analytics and AI experience | ETL, SQL analytics, governance, ML, and lakehouse-style data access | Breadth of native AWS services and deep cloud distribution | Analytics on AWS |
| Google BigQuery | Autonomous data-to-AI platform | Warehouse, AI, agents, vector search, and Iceberg interoperability | Agentic analytics experience with Google-native data and AI integrations | BigQuery page |
| Palantir Foundry | Enterprise operating system and ontology platform | Operational decisioning, data governance, and AI-driven applications | Ontology-driven operational workflows rather than warehouse-first analytics | Palantir Foundry page |
This matrix intentionally focuses on publicly visible product overlap rather than an unverified market-share ranking.
| axis | databricks position | competitor pressure | public evidence |
|---|---|---|---|
| Open formats and interoperability | Strong via Spark, Delta Lake, MLflow, Iceberg, and Unity Catalog | Google BigQuery and AWS now also emphasize Iceberg and open lakehouse interoperability | Databricks open-source and Tabular pages versus AWS and BigQuery positioning |
| Unified end-to-end platform breadth | Strong with data engineering, SQL, AI, governance, apps, and operational DB | Microsoft Fabric is closest on explicit end-to-end SaaS scope | Databricks platform pages and Microsoft Fabric overview |
| Channel reach | Strong but partner-dependent via Azure, AWS, GCP, and 1,200+ partners | Microsoft and AWS control major native cloud channels | About page, Series E, and TechCrunch Azure article |
| AI and agent depth | Strong public product momentum through Mosaic AI, Anthropic, Lakebase, and Agent Bricks | Snowflake, Fabric, BigQuery, and Palantir all market integrated AI and agent workflows | Databricks AI page, Anthropic PR, and competitor pages |
| Pricing transparency | Moderate because list prices exist, but enterprise economics are opaque | All major peers still require workload-specific normalization for true comparison | Databricks pricing pages and competitor positioning pages |
Scoring is qualitative and should be validated with win-loss and pricing data from actual deals.
Entity placement is an analyst judgement grounded in public product scope rather than a formal market-share ranking.
Databricks’ public GTM posture is coherent: one-platform messaging, cloud and partner leverage, and visible regional investment. What the public record cannot answer is how efficient those routes are or how much pricing pressure sits underneath the strong growth narrative.
partially verified confidence: medium
Databricks’ public GTM strategy centers on one-platform positioning, open and private data control, partner leverage, and geographic expansion. The story is coherent, but its efficiency is not publicly measurable.
| channel | region or buyer | public evidence | gap |
|---|---|---|---|
| Direct enterprise sales and committed-use contracts | Large enterprise accounts | Pricing pages invite customers to contact Databricks for committed-use discounts and custom requirements. | No public quota, win rate, or sales-cycle metrics were found. |
| Azure first-party route | Azure customers and public-sector workloads | Azure Databricks is a first-party Azure service with Microsoft billing and support. | No public revenue-share schedule or Azure-versus-direct mix was found. |
| Cloud marketplace and hyperscaler alignment | AWS, Azure, and GCP-aligned buyers | Databricks markets itself as multi-cloud and highlights cloud-provider compatibility. | Marketplace share and CAC are not public. |
| Partner ecosystem and consulting delivery | Large enterprise transformation programs | Databricks says it has 1,200+ cloud, ISV, and consulting partners. | No public split of partner-sourced revenue, top SIs, or partner dependency. |
| Self-serve learning, free trial, and platform adoption routes | Developers, learners, and future enterprise users | Pricing and careers materials reference free trial, free edition, and training pathways. | No public conversion rates from self-serve users to paying enterprise accounts were found. |
The GTM picture is visible at a surface level, but the economics and productivity of each route remain private.
partially verified confidence: medium
Major-customer evidence is persuasive on brand quality and cross-industry relevance, but the public record does not support a reliable view of account health, expansion, or pipeline conversion.
partially verified confidence: medium
The principal public new-business routes are enterprise sales, Azure/AWS/GCP procurement, partner delivery, training funnels, and increasingly AI-agent and database expansion attached to existing data estates.
This is a narrative funnel rather than a fully comparable cohort funnel because public conversion data is unavailable.
not publicly verifiable confidence: low
Databricks clearly operates a serious enterprise-sales motion, but public materials do not disclose quota, compensation, average cycle length, or new-hire productivity.
partially verified confidence: medium
Public regional investment commitments strongly suggest Databricks can fund continued marketing and GTM expansion. What remains unclear is whether those budgets are economically efficient and conversion-positive.
| signal | description | public evidence | verification status |
|---|---|---|---|
| UK expansion | More than $850M to expand UK presence over three years | New 137,000-square-foot EMEA hub, 500+ employees today, and training target of 100,000 people by 2028. | partially_verified |
| ANZ expansion | $300M commitment over three years | New Sydney headquarters, 85% year-over-year Q1 growth claim, and 100,000 learners target over five years. | partially_verified |
| Large-event brand presence | Data + AI Summit and training ecosystem support broad market reach | Databricks public materials repeatedly route users to trials, events, training, and academy resources. | verified |
| Regional GTM priorities | Lakebase, Genie, and Agent Bricks are featured as regional growth products | UK and ANZ releases explicitly call out these products as part of the expansion program. | verified |
Budget visibility is strongest for regional commitments rather than marketing efficiency or CAC.
Databricks still presents itself as a founder-led technical company with substantial research depth, but the current pipeline also reflects a more aggressive commercial expansion into agents, databases, and interoperability. Public sources support the direction of travel, not the eventual economic quality of that pipeline.
verified confidence: high
Databricks still presents itself as a founder-led technical organization with real research depth, specialized engineering teams, and explicit academic ties. Public materials support technical credibility more strongly than R&D resource allocation.
| name | role | technical scope | source |
|---|---|---|---|
| Matei Zaharia | Co-founder and CTO | Spark lineage, MLflow, research leadership, platform strategy | Founders and research pages |
| Ion Stoica | Co-founder and Executive Chair | Distributed-systems lineage and board-level technical influence | Founders and board pages |
| Reynold Xin | Co-founder and Chief Architect | Core platform architecture and Spark lineage | Founders page |
| Patrick Wendell | Co-founder and VP of Engineering | Platform engineering leadership | Founders page |
| Vinod Marur | SVP of Engineering | Public engineering leadership within current executive team | Leadership team page |
| Research organization | 48-publication public research footprint | Distributed systems, AI/ML, applications, and specialized engineering teams | Research page |
Public leadership depth is visible, but exact team size and span of control remain private.
The public org chart is partial by design and should be supplemented by an internal span-of-control map.
partially verified confidence: medium
Databricks’ public product pipeline is active and acquisition-fed: AI agents, operational databases, interoperability, and partner-model integrations are all moving at once. The missing question is how much of this pipeline becomes high-quality recurring revenue.
| project | status | expected date or signal | source | verification |
|---|---|---|---|---|
| Agent Bricks and Mosaic AI agents | Publicly marketed and regionally highlighted | Featured on current AI page and in ANZ release | AI page and ANZ expansion release | verified |
| Lakebase | Launched and priced publicly | Promoted in UK and ANZ expansion materials | Lakebase product and pricing pages; UK and ANZ releases | verified |
| Anthropic Claude integration | Five-year partnership live on AWS, Azure, and GCP | Claude 3.7 Sonnet available via Databricks | Anthropic partnership press release | verified |
| Delta/Iceberg interoperability | Strategic integration roadmap after Tabular acquisition | Public interoperability ambition over multiple years | Tabular acquisition press release | partially_verified |
| Neon integration for AI-native operational workloads | Announced acquisition intent | Transaction still subject to customary closing conditions and regulatory clearances | Neon acquisition press release | partially_verified |
The public roadmap is rich on product launches and partner integrations but thin on product-level economics or kill rates.
Databricks’ public management picture is strong enough to confirm a mature late-stage leadership bench, but still weak on compensation, equity, and attrition. The company looks operationally serious; what remains unclear is how resilient that seriousness is beneath the visible top layer.
verified confidence: high
The public org chart is strongest at the board and top-management level. Founders, investor-linked directors, and late-stage executive hires are all visible, but sub-C-suite operating structure is not.
partially verified confidence: medium
Databricks is clearly large enough to be operating like a global public-company candidate in headcount terms, but the public record still provides only a few anchor points rather than a full workforce model.
| signal | detail | source | diligence note |
|---|---|---|---|
| Current workforce size | 10,000+ employees worldwide across 30+ offices in 20+ countries | Careers page | Strong current scale anchor, but no function-by-function breakdown is public. |
| 2021 hiring plan | Plan to exceed 3,000 employees by end of 2021 | Forbes Series H coverage | Useful historical anchor, but it is a plan rather than a confirmed year-end actual. |
| UK / Ireland team | 500+ employees today with expectation to surpass 1,000 over the next few years | UK investment release | Shows meaningful regional GTM and engineering build-out. |
| Research hiring | Research page explicitly recruits PhDs across research and specialized engineering teams | Research page | Signals continued investment in technical depth rather than only GTM scaling. |
| ANZ footprint expansion | New 22,000-square-foot Sydney HQ and continuing team growth in ANZ | ANZ investment release | Supports regional capacity growth but not net global hiring economics. |
Public headcount data is sparse, so workforce quality and attrition still require direct diligence.
This sparse chart is intentional: it makes the headcount disclosure gap visible instead of hiding it.
verified confidence: high
Publicly named leaders cover finance, legal, field, people, security, information, and engineering — enough to confirm a mature C-suite. Public biographies remain partial and still need diligence support.
| name | role | public signal | source |
|---|---|---|---|
| Ali Ghodsi | Co-founder and Chief Executive Officer | Core operating leader and public face of financing, platform, and AI strategy | Leadership and founders pages |
| Andy Kofoid | President, Global Field Operations | Field-operations scale-up and public-market readiness signal | Leadership page and Forbes Series H coverage |
| David Conte | Chief Financial Officer | Late-stage private-company finance leadership | Leadership page |
| Trâm Phi | SVP and General Counsel | Elevated legal function while IP and model litigation risks rise | Leadership page |
| Ion Stoica and Matei Zaharia | Executive Chair and CTO; founders on the board | Founders remain inside both technical and governance structures | Founders and board pages |
The public roster is credible but still not a substitute for a full internal org chart and biography pack.
partially verified confidence: low
Public materials show that Databricks promotes benefits and flexible work, but not executive compensation or broad compensation design. This section is therefore mostly a diligence gap.
not publicly verifiable confidence: low
No meaningful public stock-plan detail was verified. Given the late-stage financing profile and explicit employee-liquidity references, this is a core private-data-room issue.
inconclusive confidence: low
No significant employee-relations problems were directly verified in the reviewed public set. That should be read as inconclusive, not as evidence that no issues exist.
inconclusive confidence: low
Public turnover evidence is thin, but founder-role distinctions and former-employee litigation suggest that people-related diligence should go deeper than the positive culture narrative.
| signal | date or period | public evidence | implication | verification status |
|---|---|---|---|---|
| Ion Stoica moved from founding CEO to Executive Chair years ago | 2015 onward, public history summarized in Forbes | Forbes recounts the governance transition and Ali Ghodsi’s rise to CEO. | Shows the company has already navigated one founder-governance reset. | verified |
| Andy Konwinski remains a founder but is not on the current executive-team page | Current as of 2026-05-14 | Founders page lists Konwinski, while leadership page does not. | Suggests a founder-role transition or reduced operational involvement that warrants direct clarification. | partially_verified |
| Andy Kofoid joined to run global field operations | 2021 | Forbes Series H coverage frames Kofoid as a late-stage field-scale hire. | Supports the view that Databricks has been professionalizing GTM for public-market readiness. | verified |
| Former employee James Weisfield is now a litigation counterparty | 2024 onward | Databricks sued Weisfield and related entities in 2024. | Raises former-employee IP and retention-governance questions. | verified |
| Comprehensive attrition metrics | Past two years | not_publicly_verifiable | No public voluntary, regretted, or function-level turnover data was found. | not_publicly_verifiable |
This table should be read as a sparse public signal set, not a true turnover ledger.
Databricks’ legal and regulatory diligence picture is materially more active than a casual reading of its trust center would imply. The combination of AI copyright exposure, recurring patent disputes, and private contract economics makes legal diligence a central workstream, not a check-the-box exercise.
partially verified confidence: medium
Databricks has material public litigation exposure, led by the Makkai copyright case and supported by a broader set of patent disputes. This is not peripheral diligence — it is a central risk area.
| case | court | filed date | matter | status | source |
|---|---|---|---|---|---|
| Makkai v. Databricks, Inc., et al. | N.D. California, 3:24-cv-02653 | 2024-05-02 | Copyright and model-training-data dispute tied to Mosaic/DBRX | Complaint filed; public reporting says motion to dismiss was denied in 2026 | CourtListener and The Register |
| Byteweavr, LLC v. Databricks, Inc. | E.D. Texas, 2:24-cv-00162 | 2024-03-08 | Patent infringement | Public docket remains active with later filings visible through 2025-06 | CourtListener |
| R2 Solutions LLC v. Databricks, Inc. | E.D. Texas, 4:23-cv-01147 | 2023-12-28 | Patent infringement | Public docket shows the case remained active through 2026-05 | CourtListener |
This table reflects only the matters directly verified in-session; it is not a complete litigation schedule.
partially verified confidence: medium
The clearest publicly verified company-initiated dispute is Databricks v. Weisfield. Beyond that, the offensive or threatened-litigation picture remains largely private.
| defendant or counterparty | court | filed date | matter | status | source or gap |
|---|---|---|---|---|---|
| James Weisfield and related entities | W.D. Washington, 2:24-cv-01417 | 2024-09-06 | Databricks complaint involving former employee and affiliated IP entities | Filed and publicly docketed; multiple defendants served | CourtListener docket |
| Other offensive disputes | not_publicly_verifiable | not_publicly_verifiable | No broader offensive litigation schedule was validated in-session | not_publicly_verifiable | Need company counsel schedule of active and threatened matters |
The absence of more cases in this table should not be read as proof that Databricks has no additional offensive or confidential disputes.
partially verified confidence: medium
For a software platform like Databricks, the most relevant public safety and exposure signals are security, privacy, and regulated-workload obligations rather than classic environmental liabilities. The public record is strongest on certifications, weakest on operating evidence and regulator interaction.
partially verified confidence: medium
Databricks’ public IP story is strongest on open-source lineage and current litigation, and weakest on the proprietary patent and trademark portfolio itself. That asymmetry is common for private software companies but still important in diligence.
| asset or right | jurisdiction or owner | public status | source | diligence note |
|---|---|---|---|---|
| Apache Spark lineage and trademark context | Apache Software Foundation ecosystem | Databricks states its engineers are the original creators of Apache Spark. | Open-source page | Databricks influence is substantial but not equivalent to sole control of community governance. |
| Delta Lake and Apache Iceberg interoperability | Linux Foundation Delta Lake and Apache Iceberg communities | Databricks publicly positions itself as bringing the two leading open lakehouse formats closer together. | Open-source page and Tabular acquisition release | Need the actual interoperability roadmap, governance rights, and any contributor concentration analysis. |
| MLflow and Unity Catalog | Open-source / Databricks-led ecosystem | Public pages identify MLflow and Unity Catalog as part of Databricks’ open-source-led governance stack. | Open-source page | Need license, trademark, and contribution agreements for all materially used open projects. |
| Mosaic/DBRX model-training rights and copyright exposure | Disputed in U.S. litigation | Makkai alleges Mosaic-related training data use that now affects DBRX-related IP risk. | CourtListener and The Register | Need training-data provenance records, outside-counsel analysis, and any indemnities from the Mosaic transaction. |
| Databricks patent and trademark portfolio | Databricks proprietary rights | not_publicly_verifiable | No patent or TTAB docket sweep was completed beyond the public lawsuit set | Request a full patent, trademark, and license schedule from counsel before any transaction decision. |
The strongest public IP evidence is around open-source positioning and current litigation, not the full Databricks proprietary portfolio.
not publicly verifiable confidence: low
No public insurance coverage schedule was verified. This matters because Databricks’ current legal and compliance posture could create material D&O, E&O, cyber, and IP exposure.
partially verified confidence: medium
Publicly visible contracts and transactions show Databricks is strategically intertwined with Microsoft, Anthropic, SAP, and recent acquisition targets. These relationships are central to the diligence case, but their economics remain private.
| counterparty | contract or transaction | public terms | strategic importance | diligence gap |
|---|---|---|---|---|
| Microsoft | Azure Databricks first-party service and strategic investment | Azure billing and support flow through Microsoft; Microsoft joined Series E. | Major distribution and infrastructure dependency with direct competitive overlap from Fabric. | Need revenue sharing, termination rights, and change-of-control protections. |
| Anthropic | Five-year strategic partnership | Claude models offered natively through Databricks on AWS, Azure, and GCP. | Expands AI-agent offering and deepens third-party model dependence. | No public committed spend, minimums, or exclusivity terms were found. |
| SAP | SAP BDC delta-sharing pricing relationship | Normal 1.25x DBU uplift; temporary 1.0x promotional rate through 2026-08-31. | Signals a real commercial integration path into large SAP data estates. | Need the full OEM, resale, or referral economics and renewal structure. |
| MosaicML | Acquisition | Approx. $1.3B inclusive of retention packages | Defines Databricks’ GenAI positioning and current litigation exposure. | Need reps and warranties, indemnities, and post-close integration economics. |
| Tabular / Neon / Arcion | Acquisitions and acquisition intent | Databricks disclosed strategic rationale but not all prices or post-close economics. | Extends interoperability, ingestion, and operational database scope. | Need actual purchase prices, earnouts, retention packages, and synergy tracking. |
This table captures only contracts and transactions visible in public sources; the operative legal documents remain private.
inconclusive confidence: low
Databricks’ public regulatory profile is dominated by certifications and compliance signals rather than a directly verified adverse-action record. That is directionally positive, but not the same as a clean regulatory bill of health.
| regime or action | public status | source | diligence note |
|---|---|---|---|
| FedRAMP and DoD IL5 | Visible in Databricks Trust & Compliance as supported programs | Trust & Compliance page | Request actual authorization packages, scope boundaries, and current ATO status by deployment. |
| ISO, SOC, HIPAA, HITRUST, PCI-DSS, GDPR, CCPA, IRAP, ISMAP, K-FSI, and related programs | Visible in Databricks Trust & Compliance | Trust & Compliance page | Request latest reports, exceptions, and remediation trackers rather than relying on logo-level certification references. |
| Due-diligence package and SOC 2 availability | Databricks says customers can download a due-diligence package and request SOC 2 Type II reports | Trust & Compliance page | Need the actual package contents and report dates. |
| Public adverse agency actions or sanctions | inconclusive | Reviewed public evidence emphasized certifications rather than agency actions | Absence of a found public action is not proof of absence; request regulator correspondence and enforcement logs. |
This table should not be read as clearance from regulators; it only reflects what was visible in the reviewed source set.
| ID | Claim | Status | Sources |
|---|---|---|---|
| EC-001 | Databricks says more than 15,000 organizations worldwide, including 70% of the Fortune 500, rely on the platform, and that it has 1,200+ global partners. | partially verified medium | SRC-001 |
| EC-002 | Databricks’ customer page says more than 20,000 customers across the globe and more than 60% of the Fortune 500 use Databricks. | partially verified medium | SRC-002 |
| EC-003 | Databricks describes its Data Intelligence Platform as a private, open, unified foundation for data and AI built on a lakehouse architecture. | verified high | SRC-003 |
| EC-004 | Databricks publicly ties itself to Apache Spark, Delta Lake, Apache Iceberg, Unity Catalog, MLflow, Delta Sharing, and Redash as core open-source technologies. | verified high | SRC-004 |
| EC-005 | Databricks markets Mosaic AI and Agent Bricks as a governed AI-agent system that can connect enterprise data with any AI model. | verified high | SRC-005 |
| EC-006 | AT&T’s Databricks case study says the company reduced fraud by up to 80% and now runs over 100 fraud-detection ML models in production. | partially verified medium | SRC-006 |
| EC-007 | Virgin Australia’s case study says Databricks drove a 75% increase in near real-time data availability and a 44% reduction in mishandled bags. | partially verified medium | SRC-007 |
| EC-008 | Databricks’ careers page says the company has 10,000+ employees, 30+ offices in 20+ countries, and is still growing aggressively. | verified high | SRC-008 |
| EC-009 | Databricks’ research page shows 48 public publications and explicitly ties the company to specialized teams across distributed systems, AI/ML, and applications. | verified high | SRC-009 |
| EC-010 | The current executive roster publicly includes Ali Ghodsi, Andy Kofoid, David Conte, Amy Reichanadter, Trâm Phi, Ron Gabrisko, Rick Schultz, Hatim Shafique, Fermín Serna, Naveen Zutshi, Vinod Marur, David Meyer, and Adam Conway. | verified high | SRC-010 |
| EC-011 | Databricks’ founders page still lists seven founders: Ali Ghodsi, Ion Stoica, Matei Zaharia, Patrick Wendell, Reynold Xin, Andy Konwinski, and Arsalan Tavakoli-Shiraji. | verified high | SRC-011 |
| EC-012 | Databricks’ board page shows founder control plus investor representation from a16z and NEA alongside independent directors. | verified high | SRC-012 |
| EC-013 | Databricks raised $250M in Series E at a $2.75B valuation. | verified high | SRC-013 |
| EC-014 | Series E also confirmed Databricks exceeded $100M ARR in 2018 and that Microsoft joined the round after Azure Databricks traction. | verified high | SRC-013 |
| EC-015 | Databricks raised over $500M in Series I at a $43B valuation and added NVIDIA as a strategic investor. | verified high | SRC-014 |
| EC-016 | Series I also said Databricks crossed a $1.5B revenue run rate, had over 10,000 global customers, more than 300 $1M-plus customers, and achieved 85% non-GAAP subscription gross margins. | partially verified medium | SRC-014 |
| EC-017 | Databricks’ Series J announcement says the company is raising $10B of expected non-dilutive financing at a $62B valuation and had completed $8.6B to date. | partially verified high | SRC-015 |
| EC-018 | Series J also says Databricks grew over 60% year-over-year in Q3 FY2025, expects to cross a $3B revenue run rate, expects positive free cash flow in Q4 FY2025, has 500+ customers above $1M annual run rate, and sees Databricks SQL at a $600M run rate. | partially verified medium | SRC-015 |
| EC-019 | Forbes reported Databricks raised $1B in Series G at a $28B valuation, had $425M+ ARR growing 75%+, and added AWS, CapitalG, Salesforce Ventures, and Microsoft to the investor base. | verified high | SRC-016 |
| EC-020 | Forbes reported Databricks raised $1.6B in Series H at a $38B valuation, exceeded $600M ARR, aimed for 3,000+ employees by year-end 2021, and hired Andy Kofoid from Salesforce. | verified high | SRC-017 |
| EC-021 | Databricks’ June 2024 NVIDIA partnership release says Databricks reached over $1.6B in revenue for its fiscal year ending January 31, 2024, representing over 50% year-over-year growth. | partially verified medium | SRC-018 |
| EC-022 | Databricks and Anthropic publicly announced a five-year partnership that brings Claude models natively to the Databricks platform across AWS, Azure, and GCP. | verified high | SRC-019 |
| EC-023 | TechCrunch reported Azure Databricks is a first-party Azure service, with Microsoft handling billing and support. | verified high | SRC-020 |
| EC-024 | Databricks said the MosaicML acquisition was valued at approximately $1.3B, inclusive of retention packages. | verified high | SRC-021 |
| EC-025 | Databricks said the Arcion acquisition was valued at over $100M, inclusive of incentives. | verified high | SRC-022 |
| EC-026 | Databricks’ Tabular acquisition release says the company wants to bring Delta Lake and Apache Iceberg closer together over time to reduce lakehouse format fragmentation. | verified medium | SRC-023 |
| EC-027 | Databricks’ Neon acquisition release says over 80 percent of databases provisioned on Neon were created automatically by AI agents. | partially verified medium | SRC-024 |
| EC-028 | Databricks announced more than $850M of UK investment, 500+ UK employees, a new 137,000-square-foot London hub, and over 50% of the FTSE 100 as regional customers. | partially verified medium | SRC-025 |
| EC-029 | Databricks announced a $300M ANZ investment, said Q1 regional growth exceeded 85% year over year, and highlighted Lakebase, Genie, and Agent Bricks as regionally available priorities. | partially verified medium | SRC-026 |
| EC-030 | Databricks publishes Lakeflow Jobs pricing at $0.15 / DBU for classic jobs and $0.35 / DBU for serverless jobs. | verified high | SRC-027 |
| EC-031 | Databricks SQL pricing is publicly listed at $0.22 / DBU for SQL Classic, $0.55 / DBU for SQL Pro, and $0.70 / DBU for SQL Serverless. | verified high | SRC-028 |
| EC-032 | Databricks publicly prices Lakebase at $0.092 per capacity unit hour and $0.345 per GB-month of storage. | verified high | SRC-029 |
| EC-033 | Databricks’ SAP Business Data Cloud pricing page says normal DBU processing runs at a 1.25x uplift but is temporarily promoted at 1.0x through 2026-08-31. | verified high | SRC-030 |
| EC-034 | Databricks’ trust center publicly offers a due-diligence package and lists a broad certification set including FedRAMP, DoD IL5, SOC, ISO, HIPAA, HITRUST, PCI-DSS, GDPR, and CCPA. | verified medium | SRC-031 |
| EC-035 | The Makkai complaint shows authors sued Databricks and Mosaic ML in May 2024 over alleged copyright infringement tied to model training. | verified high | SRC-032 |
| EC-036 | Independent public reporting says Databricks’ motion to dismiss in the Makkai case was denied and frames the case as carrying extraordinary damages risk tied to roughly 196,000 titles and statutory damages up to $150,000 per work. | partially verified medium | SRC-033 |
| EC-037 | Databricks filed suit against former employee James Weisfield and related entities in September 2024. | verified high | SRC-034 |
| EC-038 | Byteweavr sued Databricks for patent infringement in March 2024 in the Eastern District of Texas. | verified high | SRC-035 |
| EC-039 | R2 Solutions sued Databricks for patent infringement in December 2023 and the public docket remained active through May 2026. | verified high | SRC-036 |
| EC-040 | Snowflake publicly describes its platform as serverless, managed, multi-cloud, and governance-heavy, confirming direct overlap with Databricks’ warehouse-plus-AI positioning. | verified medium | SRC-037 |
| EC-041 | Microsoft Fabric publicly describes itself as an end-to-end analytics SaaS platform spanning data engineering, data science, warehousing, databases, and governance over OneLake. | verified high | SRC-038 |
| EC-042 | AWS publicly markets a comprehensive analytics stack plus an integrated SageMaker experience with Iceberg-compatible lakehouse access, confirming meaningful functional overlap with Databricks. | verified medium | SRC-039 |
| EC-043 | Google BigQuery publicly positions itself as an autonomous data-to-AI platform with built-in AI, agents, and Apache Iceberg interoperability. | verified medium | SRC-040 |
| EC-044 | Palantir Foundry publicly markets itself as an Ontology/AI-powered operating system for the modern enterprise. | verified medium | SRC-041 |
| EC-045 | Reviewed public customer materials disclose broad logo breadth and high-value customer counts, but do not disclose top-customer revenue share, top-10 concentration, churn, or retention. | not publicly verifiable medium | SRC-002SRC-015 |
| EC-046 | Reviewed public financial materials do not provide audited financial statements, AR aging, backlog schedules, or a public debt inventory. | not publicly verifiable medium | SRC-013SRC-014SRC-015 |
| EC-047 | Reviewed public GTM materials do not disclose sales compensation, average quota, sales-cycle duration, or marketing-budget productivity. | not publicly verifiable medium | SRC-027SRC-028SRC-025SRC-026 |
| EC-048 | Reviewed public people materials do not disclose executive compensation schedules, stock-plan mechanics, or two-year attrition data. | not publicly verifiable medium | SRC-008SRC-010SRC-011 |
| EC-049 | The public founders and leadership pages show Andy Konwinski remains listed as a founder but is absent from the current executive-team page. | partially verified medium | SRC-010SRC-011 |
| EC-050 | The regulatory evidence reviewed in-session primarily shows certifications, standards, and due-diligence materials rather than a public record of adverse agency actions or sanctions against Databricks. | inconclusive low | SRC-031 |
This report is a public-evidence diligence snapshot, not investment advice. Important financial, legal, technical, and contractual facts remain non-public and should be verified directly with management and primary documents before any investment decision.