Sigma Computing

1.1 Identity, product, and leadership

Sigma Computing, now branded as Sigma, was founded in 2014 as Bitmoon Computing by Rob Woollen and Jason Frantz and is headquartered in San Francisco. Its product is a cloud-native analytics and business-intelligence layer that queries live data in Snowflake, Databricks, BigQuery, and other warehouses through a spreadsheet-style interface. The warehouse-native architecture is the central product claim: users can explore large datasets without writing SQL while existing warehouse permissions and governance remain relevant. The company has expanded that base into Sigma Agents, which SiliconANGLE describes as no-code agents that can run interactively, autonomously, or through external API calls. Mike Palmer is identified as chief executive, with Woollen as co-founder and CTO and Frantz as co-founder and chief architect in the company profile. Leadership depth beyond those named executives and the board's composition are not publicly established by the assigned sources, so key-person and governance diligence remains open.[CO001, CO002, CO003, CO004, CO005]

1.2 Financing, valuation, and operating scale

Public reporting indicates that Sigma closed an $80 million Series E in May 2026 at a $3 billion valuation, led by Princeville Capital with strategic participation from Databricks Ventures, ServiceNow Ventures, and Workday Ventures. SiliconANGLE and The Next Web both report the same round and describe the valuation as double the prior $1.5 billion Series D level. The same reporting says annual recurring revenue reached $200 million in April 2026, up from roughly $100 million a year earlier, and customer count exceeded 2,000. These are reported operating metrics, not audited financial statements. Assigned profile material reports conflicting cumulative funding totals ($661 million-plus in its round table versus $961.3 million in its headline), while it explicitly says public metrics vary by source and methodology. The report therefore treats the latest round, valuation, ARR, and customer count as reported values with source and period qualifiers, not as verified accounting results.[CO006, CO007, CO008, CO009, CO010, CO011]

1.3 Milestones, dependencies, and open diligence

Sigma's trajectory combines a long-standing warehouse-native BI proposition with a recent shift toward agentic analytics. The assigned evidence records the 2014 founding, the 2021 Series C, the 2024 Series D, the 2026 Series E, the April 2026 ARR milestone, and the launch and rapid adoption of Sigma Agents. Strategic investors that are also warehouse or enterprise-software ecosystem participants may support distribution and integration, but they can also create dependence on partner priorities and competitive overlap. The competitor source describes Sigma's expanding architecture as including optional materialization and AI infrastructure layers, while positioning rival Astrato as simpler; this is a competitor's comparison, not an independent benchmark. UpGuard's page confirms ongoing external monitoring across 330-plus checks but does not disclose a breach or a passing score in the assigned text. Key unresolved items are board control, cap-table terms, profitability and cash efficiency, customer concentration, retention, and independent customer economics.[CO012, CO013, CO014, CO015, CO016]

1.4 Exhibits

2.1 Market boundary and status quo

The relevant market is enterprise analytics software that lets business users query, model, and act on governed operational data. It includes warehouse-native self-service BI, embedded analytics, write-back workflows, and AI-assisted analysis when those capabilities are sold as an analytics layer. It excludes the underlying cloud warehouse, general-purpose data engineering, standalone AI foundation models, and bespoke consulting. The status quo is not one product: it is a mix of legacy dashboard suites, spreadsheets, analyst-built SQL, and internal data applications. Sigma's fit is strongest where the buyer already treats Snowflake, Databricks, BigQuery, or another supported SQL warehouse as the system of record. That boundary matters because the platform's live-query architecture can reduce extract duplication, but it does not by itself create a new data budget or eliminate warehouse compute costs.[CM001, CM002, CM003, CM004]

2.2 Sizing lenses and growth signal

Public evidence supports multiple directional sizing lenses rather than a defensible Sigma-specific TAM/SAM/SOM. A company-focused editorial source reports $200 million of ARR, more than 2,000 businesses, and 1.1 million new active users, while The Next Web reports the same ARR milestone alongside an $80 million Series E at a $3 billion valuation. These are traction and investor-valuation signals, not market-size measurements. A competitor review supplies a narrower workflow lens: warehouse-native self-service for Excel-fluent analysts, with structural exclusions for NoSQL, API-first, and highly customized customer-facing embedded use cases. The market therefore has at least three separable pools: internal warehouse BI, embedded/product analytics, and AI-native data workflows. Available sources do not disclose the spend base, addressable geography, penetration, or conversion assumptions needed to convert those pools into comparable TAM, SAM, and SOM.[CM005, CM006, CM007, CM008, CM009]

2.3 Buyers, adoption path, and constraints

The likely economic buyer is a data, analytics, finance, or operations leader who already funds a cloud warehouse and wants more business-user throughput without moving data into a separate analytics store. Users are Excel-fluent analysts and operational teams; the payer is usually the enterprise data or business function budget owner, with warehouse compute remaining an adjacent cost. Adoption is triggered by demand for faster ad hoc analysis, governed AI workflows, write-back planning, or broader access to live data. Constraints are equally material: supported warehouse dependence, optional metric governance without a mature semantic layer, iframe-only embedding according to the competitor review, and uncertain compute economics. The market's growth thesis is therefore conditional on migration from incumbent BI and spreadsheets while preserving trust, governance, and measurable ROI.[CM010, CM011, CM012, CM013, CM014]

2.4 What remains unproven

The evidence base is strong enough to define a market thesis but not to underwrite market share or valuation from market size alone. No retained source provides a comparable global or regional BI spend denominator, Sigma's win rate against Power BI, Tableau, Looker, or internal builds, segment-level pricing, renewal or expansion rates, or customer payback. The available growth figures are partly company-originated or company-focused summaries, and the independent security page reviewed here describes its monitoring methodology without publishing a Sigma market or adoption benchmark. The diligence priority is to obtain bottom-up bookings by segment and geography, customer cohort retention, competitive displacement data, warehouse-compute allocation, and a reconciled definition of ARR, customers, and active users.[CM015, CM016]

2.5 Exhibits

3.1 Competitive landscape and buyer alternatives

Buyers can solve Sigma's core job through at least five routes: Sigma and newer warehouse-native tools; AI-first exploration led by ThoughtSpot; incumbent suites such as Power BI, Looker, Tableau, Qlik, and Domo; lower-cost products such as Metabase; or the status quo of spreadsheets and internal applications. Independent profiles consistently place Sigma beside these categories rather than in a narrow standalone market. Sigma's strongest position is with cloud-warehouse customers whose operators already think in spreadsheets but need governed access to live data. The positioning map is an evidence-backed ordinal synthesis, not a measured benchmark, because the assigned evidence does not provide a common performance, adoption, or total-cost dataset across vendors.[CP001, CP002, CP009, CP011, CP012, CP015]

3.2 Capability differentiation and operating tradeoffs

Sigma's clearest differentiation is the combination of a spreadsheet interaction model, live queries against customer-controlled warehouses, and operational features such as input tables and writeback. That combination broadens the addressable user beyond SQL specialists and distinguishes Sigma from read-only dashboarding. It is not unique in every dimension: ThoughtSpot also emphasizes live-warehouse exploration, Looker has a mature governed semantic model, Omni targets analytics-engineering workflows, and several vendors support embedding. The adverse evidence is directionally consistent on two tradeoffs: complex live workbooks can increase latency and warehouse consumption, while Sigma's visualization and presentation depth may lag dashboard-first products. No independent workload benchmark in the pool quantifies either effect.[CP002, CP003, CP004, CP005, CP008, CP009, CP010, CP016]

3.3 Pricing, distribution, switching, and channel power

Sigma's public pricing is sales-led, so third-party list-price comparisons are useful only as screening signals and not evidence of realized contract economics. Microsoft can make Power BI the default inside existing enterprise agreements, Google can pair Looker with BigQuery and Gemini, and both routes reduce the need for a separate buying process. Sigma counters with warehouse neutrality, spreadsheet familiarity, and strategic ties to Databricks, ServiceNow, and Workday, but those relationships are not exclusive distribution rights. Switching cost therefore comes mainly from governed workbook logic, user training, embedded applications, and operational writeback workflows rather than proprietary data storage. Multi-homing remains plausible because several tools can query the same warehouse, although no source in the assigned pool measures actual overlap or migration cost.[CP005, CP006, CP007, CP008, CP011, CP015]

3.4 Moat durability and displacement risk

Sigma has enough reported scale and financing to compete aggressively, and independent coverage describes its 2026 expansion from warehouse-native BI into agentic analytics. However, that repositioning increases overlap with cloud platforms, incumbent BI vendors, and focused AI-search products. The durable moat is more likely to be accumulated governed workflows, embedded applications, and operator adoption than the phrase "agentic analytics," which large software vendors can copy. Competitor-authored criticism that Sigma's platform identity is broad and that some AI functions are immature should be treated as adverse directional evidence, not an independent benchmark. The underwriting conclusion is a differentiated product with real switching friction but no proven structural lock-in; customer-level win/loss, migration, and compute-cost data remain required.[CP004, CP005, CP013, CP014, CP015, CP016]

3.5 Exhibits

4.1 Revenue model and monetization

Sigma's monetization appears to be enterprise SaaS built around a cloud-warehouse analytics and AI applications platform. The product is sold to business and technical teams that query live warehouse data through spreadsheet, SQL, Python, and AI workflows; a 2026 job posting describes technical presales, RFP/RFI responses, prototypes, and concurrent enterprise engagements, which is consistent with a consultative sales motion rather than self-serve software. The same posting says Sigma has more than 2,000 customers and reached $200M ARR in April 2026, but the issuer is Sigma and the figures are not independently audited in the source. Third-party procurement commentary describes role-based tiers and custom enterprise quotes, while noting that published contract terms and typical deal sizes are not available. Consequently, list-price or role-price indications should not be treated as realized revenue, discount-adjusted ASP, or gross margin evidence.[CI001, CI002, CI003, CI004, CI016]

4.2 Unit economics and cost structure

The public file is insufficient to calculate CAC payback, gross margin, contribution margin, or sales efficiency. Sigma's live-query architecture can reduce data duplication, but it also leaves customer workloads dependent on warehouse compute; an independent competitor article flags performance under load and potentially unpredictable costs from heavy or frequent queries. Those observations are risk signals, not quantified Sigma cost data. The enterprise presales requirements, including solution prototypes, security responses, and 25% travel, indicate a non-trivial implementation and sales-support burden, but no public source supplies fully loaded selling expense, deployment cost, support cost, or payback period. A buyer should therefore model economics with actual cohort bookings, services effort, warehouse-cost pass-through, and renewal data rather than extrapolating from ARR growth alone.[CI005, CI006, CI007, CI008]

4.3 Capital adequacy and financial verdict

Sigma's May 2026 Series E provides an observed financing event and a reported valuation anchor, but it does not disclose cash on hand, burn, runway, debt covenants, or use of proceeds. The Next Web reports an $80M round at a $3B valuation and describes the financing as a test of the company's agentic-analytics repositioning; its discussion of 15x ARR is an editorial calculation based on the reported $200M ARR, not an audited accounting multiple. The round size can plausibly extend operating flexibility, but cash available cannot be inferred from funding raised because expenses, debt, working capital, and transaction terms are unknown. The financial verdict is therefore constructive on reported growth and financing access, but research-more on earnings quality, capital efficiency, and downside protection.[CI009, CI010, CI011, CI012]

4.4 Underwriting gaps

The most decision-relevant missing inputs are recurring-revenue definitions and cohort behavior, realized price and discount data, gross margin by product and customer segment, cash and burn forecasts, implementation and warehouse-cost allocation, debt or project-finance obligations, and concentration by customer, partner, and warehouse. Public profiles explicitly state that profitability figures were not found and that reported metrics vary by source and methodology. These are material gaps rather than evidence of weak performance: the available data can support a reported-growth narrative, but not a reliable cash-flow or valuation underwriting case.[CI013, CI014, CI015]

4.5 Exhibits

5.1 Customer workflow and product surface

Sigma's core workflow is a business user exploring live warehouse data without waiting for a data team to publish a static dashboard. The product combines a spreadsheet-style workbook with SQL, Python, natural-language assistance, collaborative sharing, write-back input tables, embedded analytics, and parameter-driven data applications. The official announcement describes the product moving beyond passive business intelligence toward AI-powered workflows, with Sigma Agents, Sigma Assistant, Client Connectors, and API Actions. This broadens the addressable workflow from analysis to action, but it also creates a product-scope diligence question: the public record describes capabilities and customer examples more clearly than it quantifies adoption, latency, error rates, or the amount of implementation work required for each module.[CE001, CE002, CE003, CE004]

5.2 Warehouse-native architecture and dependencies

The product's central architectural choice is to query the customer's cloud warehouse rather than copy data into a separate analytics store. Independent coverage says Sigma runs spreadsheet operations, SQL, Python, and AI Apps against the warehouse compute layer; the company says agents run where the data lives and inherit the permissions already controlled there. This can reduce extract duplication and align the user experience with existing governance, but it makes product performance and availability dependent on warehouse configuration, cloud-provider regions, query workload, semantic modeling, and customer data quality. The reviewed sources identify Snowflake, Databricks, BigQuery, Redshift, and PostgreSQL connectivity, yet they do not independently establish comparative performance or the completeness of each connector's feature support.[CE005, CE006, CE007, CE008]

5.3 Maturity, deployment, and roadmap signals

Public maturity signals are mixed. Sigma's April 2026 announcement reports $200 million of ARR, more than 2,000 businesses, 1.1 million new active users in the latest fiscal year, and customer workflows that include real-time analytics and embedded AI applications. A current enterprise customer-success posting emphasizes adoption programs, training, success plans, integrations, production BI implementation, and net-dollar-retention work, which is consistent with a product that requires active deployment and expansion support rather than a purely self-serve workflow. A separate solution-engineering listing calls out discovery, demos, technical and security RFP responses, data migration, and partner-channel execution. These are useful operating signals, not independent proof of implementation success, retention, uptime, or roadmap delivery.[CE009, CE010, CE011, CE012]

5.4 Trust, reliability, and technical diligence

Sigma's stated trust proposition is inherited warehouse governance: live queries, row- and column-level controls, and agents operating inside existing permissions. The official release also describes API Actions and external-system connectors, which expand the control plane beyond read-only analytics and raise the importance of action authorization, auditability, secrets handling, and failure recovery. The reviewed UpGuard page reports continuous monitoring across 330-plus external checks but does not expose a company-specific score or control-level results in the fetched text. A review aggregation reports a 99.9% monthly target availability and notes dependence on upstream warehouses and cloud regions, while also saying historical multi-month uptime is not fully summarized. Investors should therefore treat security posture, incident history, service-level attainment, agent evaluation, connector coverage, and customer-side operating burden as open diligence items rather than infer them from architecture language.[CE013, CE014, CE015, CE016]

5.5 Exhibits

6.1 Customer base and segmentation

The public customer picture is enterprise-oriented but spans several adoption motions. Third-party reporting says Sigma served more than 2,000 customers by April-May 2026 and names AMD, Duolingo, Colgate-Palmolive, JPMorgan Chase, DoorDash, and other large organizations. The retained case-study index broadens that picture across financial services, logistics, healthcare, consumer, software, media, and manufacturing, with users ranging from data engineers and analytics leaders to product, finance, marketing, and operations teams. These references support a segmentation around enterprise analytics, analyst-led self-service, embedded analytics, and operational business workflows rather than a single narrow buyer. However, the catalog is syndicated vendor success material: it establishes named references and claimed use cases, not a representative mix of paying customers, revenue bands, geographies, contract sizes, or deployment maturity.[CU001, CU002, CU004, CU007]

6.2 Adoption trajectory and customer proof

Adoption evidence is strongest at the top of the funnel. The Next Web reports 1.1 million new active users in Sigma's latest fiscal year and more than 2,000 customers, while a review aggregator reports a 4.6 average across 508 reviews. The named proof is concrete enough to show varied use: the indexed stories claim DoorDash increased queries by 30%, Agero achieved five times more self-service analytics and 50% lower BI costs, and customers use Sigma for embedded analytics, freight visibility, financial reporting, fraud controls, marketing, and operational workflows. These are useful deployment signals, but they are mostly company-issued outcomes without disclosed baselines, cohort definitions, implementation costs, or independent replication. Sigma Agents being described as the company's fastest-adopted product adds a current expansion surface, although no installed-base penetration or paid-conversion denominator is public.[CU003, CU005, CU006, CU008, CU009, CU015]

6.3 Durability, expansion, and concentration risk

Public evidence does not close the durability question. The retained sources provide no NRR, GRR, logo churn, renewal rate, contract length, customer cohort, top-customer share, or warehouse-partner concentration. Review synthesis is directionally constructive on onboarding, warehouse connectivity, support, and collaboration, but also notes learning curves, pricing concerns, slow loads, and friction in complex workbooks. Competitor-authored accounts go further: Reflex describes one evaluated deployment with near-zero usage at renewal and estimates mid-sized annual fees of roughly $70,000-$80,000 plus add-ons, while Astrato argues that per-seat tiers and warehouse compute can make costs unpredictable. Those claims are self-interested and not independent benchmarks, yet they identify testable churn mechanisms. Microsoft Marketplace's instruction to contact Sigma for annual pricing reinforces procurement opacity. A prospective investor should therefore condition expansion confidence on customer calls, cohort retention schedules, concentration data, and renewal economics rather than logo breadth.[CU010, CU011, CU012, CU013, CU014, CU016]

6.4 Exhibits

7.1 Ranked risk view

Sigma's highest-conviction risks are architectural and executional rather than a documented enforcement event. The product runs live against customer cloud warehouses, so performance, governance quality, and customer compute economics remain partly outside Sigma's control. Public descriptions also show a move into AI agents, API actions, and workflow automation; that expands the consequence of permission, model-output, and operational-control failures even where the underlying data remains in the warehouse. The evidence supports a high residual exposure to complexity and implementation risk, but not a claim that a breach, outage, or enforcement action has occurred. Mitigation evidence is strongest where Sigma or its hiring materials describe governance, GRC, employment counsel, and human review. UpGuard describes continuous external checks, but its public page does not expose the full assessment. Evidence remains weakest on quantified incident history, uptime, burn, and concentration.[CR001, CR002, CR003, CR004, CR013, CR015]

7.2 Legal, privacy, and regulatory exposure

Sigma's public legal posture is better described as active compliance work than as a cleared legal record. A company-posted employment-counsel role covers federal, state, local, and potentially international employment law, investigations, reductions in force, outside counsel, litigation matters, and M&A diligence. The same posting says applicant data may transfer across the United States, United Kingdom, Canada, and Australia and that AI assists candidate screening while humans retain decision authority. Those statements identify concrete privacy, employment, cross-border-transfer, and AI-governance obligations, but they do not establish a violation. A buyer or investor should request the current privacy and data-processing terms, incident and litigation log, employment-law coverage by jurisdiction, and evidence that AI-assisted workflows have documented controls.[CR005, CR006, CR007, CR008, CR002]

7.3 Operational, partner, and financial dependencies

Sigma's warehouse-native design is a strategic strength and a risk concentration. Marketplace material says users query massive datasets in the cloud warehouse, while the independent review notes that poorly governed workbooks can materially increase Snowflake, BigQuery, or Databricks spend and that implementation requires connection setup, semantic modeling, row-level security, and role mapping. That creates a shared-fate dependency on customer data quality, warehouse architecture, permissions, and cloud-provider availability. The public evidence also contains unaudited metrics, requires contact for annual pricing, and provides no public profitability figure, audited burn, runway, margin, or warehouse-compute uplift; these are diligence gaps, not estimates. Strategic investors include infrastructure partners and potential competitors, adding an ecosystem-governance dimension. The risk register therefore emphasizes measurable controls: workload-cost monitoring, permission review, SLA and incident evidence, renewal and concentration data, partner portability, and a cash-efficiency package.[CR009, CR010, CR011, CR012, CR014, CR015, CR016]

7.4 Exhibits

8.1 Investment thesis and anti-thesis

The positive case is that Sigma has reached meaningful scale in a structurally attractive warehouse-native analytics market. The strongest current financing evidence is a May 2026 Series E of $80 million at a reported $3 billion valuation, while the same coverage reports $200 million of ARR and more than 2,000 customers. The product thesis is also coherent: a spreadsheet-like interface operates on live warehouse data, and Sigma is extending that architecture into agentic analytics. Strategic participation from Databricks, ServiceNow, and Workday is a useful ecosystem signal, but it is not proof of customer economics or exit value. The anti-thesis is valuation sensitivity: the reported price implies roughly 15x ARR on one private-market data source, and the public record does not disclose gross margin, net retention, burn, runway, customer concentration, or preference terms. A $3 billion mark can therefore be reasonable only if growth remains unusually strong and the agentic product expands durable monetization rather than merely changing the category label.[CV001, CV002, CV003, CV004, CV005]

8.2 Price discipline and scenarios

The entry question should be framed as a range of outcomes rather than as a single fair-value point. Multiples.vc records a $3 billion May 2026 valuation against $200 million of recorded revenue or ARR and calculates a 15.0x revenue multiple; The Next Web independently describes the same round and says the valuation doubled from $1.5 billion while revenue doubled over the same period. Those observations support a strong momentum case, but they do not establish audited revenue or recurring-revenue composition. Multiples.vc lists LTM EV-to-revenue references of 1.7x for Teradata, 4.9x for Amplitude, and 7.6x for Planisware, all below Sigma's recorded 15.0x figure, but those companies are imperfect business-model and growth matches rather than a dispositive benchmark. Latka instead reports $100 million of annual revenue in April 2025 and $581.3 million raised across seven rounds, illustrating the data-quality problem rather than resolving it. A base case should require refreshed management financials, cohort retention, margin and burn data, and cap-table terms before accepting the reported price. The bear case is a multiple reset if growth slows, AI-agent adoption proves shallow, or enterprise budgets consolidate around incumbent platforms. The bull case requires sustained growth, expansion revenue, and evidence that agents increase paid workload rather than only usage.[CV006, CV007, CV008, CV009, CV010, CV015]

8.3 Exit readiness and diligence asks

Sigma has credible ingredients for an eventual strategic or public-market outcome: private-market sources describe a broad enterprise customer base, strategic investors, and a product positioned across analytics, data apps, and AI workflows. However, exit readiness cannot be inferred from a financing headline. The latest round was privately reported rather than disclosed in a public filing, and the source set contains materially different totals for funding, revenue, ARR, and headcount. Silicon Valley Investclub estimates about 13% dilution in the Series D and about 3% in the Series E from disclosed round sizes and post-money values, while noting that the shareholder register is not public; these calculations do not reveal preference rights or fully diluted ownership. UpGuard's current vendor-risk page also demonstrates that external monitoring exists, but its public page does not provide a complete score or incident history; absence of a disclosed incident is not evidence of absence. Before investment, diligence should obtain the executed financing documents, updated cap table, liquidation preferences, option-pool treatment, debt and convertibles, ARR bridge, gross margin, net retention, logo churn, customer concentration, and AI-agent usage-to-revenue conversion. Thesis-break triggers should include a material down-round, weakening net retention, sustained margin deterioration, a security or reliability event that causes enterprise churn, or failure to convert agent adoption into contracted recurring revenue.[CV011, CV012, CV013, CV014]

8.4 Exhibits