Review of o9 Solutions, supply chain planning software vendor
Go back to Market Research
o9 Solutions (often stylized “o9”) is a Dallas-based enterprise software editor founded in 2009 that sells a cloud-hosted planning suite branded as the “Digital Brain” for integrated business planning (IBP), demand and supply planning, supply chain analytics, and revenue growth management. The platform centers on an Enterprise Knowledge Graph (EKG) and an in-memory “Graph-Cube” store that model business entities and hierarchies, fed by batch ETL and real-time APIs. Commercial distribution leans on hyperscaler marketplaces (Microsoft Azure, Google Cloud; collaborations with AWS) and a partner ecosystem for implementation. Public materials emphasize configuration artifacts (IBPL queries, model/report designers), packaged apps (IBP, Control Tower, MEIO, RGM), and recent “composite agents” built on GenAI. Technical specifics of predictive/optimization components are described at a high level (ML integrations, optional Vertex AI Forecast, optimizer partnerships) with some patents and training collateral but limited reproducible details. Financing milestones include KKR’s 2020 minority investment, a $295M round led by General Atlantic in 2022, a $116M top-up in 2023, and production go-lives such as Li Auto in 202512345.
o9 overview
What the product does (high level). o9 provides a multi-tenant SaaS platform that ingests enterprise data (ERP, POS/EPOS, syndicated data, etc.), models it in an Enterprise Knowledge Graph and Graph-Cube in-memory store, and exposes packaged planning applications: IBP/S&OP, demand sensing & forecasting, supply planning, control tower, production scheduling, multi-echelon inventory optimization, and Revenue Growth Management (pricing, promotions, assortment). Connectivity is via ETL services and secure REST APIs; deployment is cloud-native on Azure/AWS/GCP with listings on Azure & Google Cloud marketplaces and joint go-to-market with hyperscalers67891011121314. o9 positions the EKG/Graph-Cube as the differentiating data/metadata layer across these apps6789.
How it does it (high level). Customer data is staged and loaded (batch and real-time) into the platform; business entities and hierarchies are modeled; planners work in web UIs and dashboards backed by Graph-Cube. Configuration artifacts (e.g., IBPL select queries, model/report/layout designers) and Platform APIs expose Graph-Cube data; public docs and training materials reference “GraphCube Server” and Platform APIs for UI access to the model91516. Optimization/ML is described through: (1) native forecasting/planning features, (2) integrations with cloud ML (e.g., Vertex AI Forecast) and (3) a partnership ecosystem (e.g., Gurobi) for solver-backed scenarios1718. Recent releases highlight GenAI “composite agents” that orchestrate cross-functional planning actions1920.
State of the tech (high level). o9’s public footprint is consistent with a modern hyperscaler-hosted APS: cloud-marketplace distribution, API/ETL connectivity, an in-memory columnar/graph hybrid store (Graph-Cube) and a packaged-app layer. Materials substantiate the presence of an internal data/semantic modeling layer and extensive UI/config frameworks. However, reproducible disclosures about forecast model classes, uncertainty treatment, and solver formulations are sparse; o9 often references AI/ML and optimization in marketing language and partner press, with training materials and solution pages offering glimpses rather than full-stack technical transparency679211819201422.
A more detailed introduction
- Company & financing. Founded by Chakradhar (Chakri) Gottemukkala and Sanjiv Sidhu; leadership bios and the corporate “About” page confirm the founding team and positioning12324. External capital: KKR minority investment (Apr 2020)2; $295M (Jan 2022) led by General Atlantic3; $116M incremental (Jul 2023) at a $3.7B post-money valuation (company press) with a contemporaneous Form D-style trade press note425.
- Go-to-market & partners. Product purchasable via Azure Marketplace (IBP, Revenue Mgmt) and Google Cloud Marketplace; AWS collaboration and analyst coverage; Google Partner directory presence101112131422.
- Product surface. Solutions include IBP, S&OP, demand forecasting/sensing, Supply Chain Control Tower, MEIO, production scheduling, RGM; a Data Science (PaaS) offer claims an “open” platform to embed Python/R/PySpark models101121.
- Architecture signals. Platform pages emphasize Connectivity Services & APIs, ETL Services, “direct in-memory loaders,” adapters for SAP/Oracle, and Graph-Cube analytics over multi-level hierarchies; a public Guide entry confirms Platform APIs “to access the GraphCube server data”69.
- Recent claims. “Composite agents” (GenAI) to execute cross-functional planning; Vertex AI Forecast integration; Microsoft GenAI collaboration updates; AWS collaboration expansions19172014.
- Customer evidence. Press releases document production rollouts like Li Auto (Jan 22, 2025)5.
o9 vs Lokad
Scope & approach. Both vendors target quantitative supply-chain decisions, but their delivery models and technical disclosures differ. o9 markets a packaged APS suite delivered via hyperscaler marketplaces, with an EKG/Graph-Cube core, UI-heavy configuration (IBPL), and optional cloud-ML integrations; the platform’s optimization/ML internals are mostly black-box from public docs101169152118192014. Lokad, by contrast, exposes a domain-specific language (Envision) as the primary interface to build probabilistic forecasting + decision optimization pipelines; its technical docs detail the compiler/VM (Thunks) and distributed execution, and emphasize probabilistic modeling and decision-centric optimization (e.g., MEIO, prioritized action lists) as code26272829[^40]30.
Data & execution layers. o9’s Graph-Cube models hierarchies and enables fast aggregation/disaggregation for planning; access is via platform APIs and configured UIs6978. Lokad’s stack documents an event-sourced persistence and a distributed VM executing Envision bytecode over a multi-tenant cluster272829.
ML/optimization transparency. o9 references AI/ML broadly (Vertex AI Forecast, GenAI agents, solver partnerships), but algorithmic specifics (uncertainty modeling, objective functions, constraints) are not extensively documented in public sources; reliance on partner technologies suggests a pluggable ML/OR posture17181920. Lokad publishes how uncertainty is represented (random variables/quantile grids in Envision), with technical articles and case materials describing the probabilistic and stochastic optimization approach and its execution substrate26272829.
User experience & programmability. o9 emphasizes out-of-the-box applications with model/report/layout designers and IBPL queries—a configuration-first UX suited to standard processes across functions1516. Lokad is programming-first: users (often “supply chain scientists”) write Envision code; this increases transparency and flexibility for bespoke constraints at the cost of a learning curve262729.
Bottom line. o9 is a suite-style APS with a proprietary EKG/Graph-Cube model and broad functional coverage, packaged for enterprise IT procurement. Lokad is a programmable quantitative platform that surfaces the math and execution engine; its differentiation rests on probabilistic modeling and code-level control.
Corporate history & financing
-
Founding / leadership. Public bios list Chakri Gottemukkala (CEO, co-founder) and Sanjiv Sidhu (co-founder)12324.
-
Financing chronology.
-
Partners & channel. Strategic alliances with Microsoft, Google Cloud, and AWS, including marketplace availability and collaboration announcements; Google partner directory profile10111213142220.
Acquisition activity
No corporate acquisitions by or of o9 surfaced in the press archive; o9 instead pursues alliances and joint offerings, e.g., Marubeni’s Mo9 supply-chain/management solution for Japan (2024)313233. We found no third-party filings indicating M&A; such deals would typically appear in newsroom or partner disclosures31.
Product surface, modules & rollout
- Packaged solutions. IBP, S&OP, Demand Planning (incl. AI/ML Demand Forecasting & Demand Sensing), Supply Chain Control Tower & Analytics, Production Scheduling, MEIO, Supplier Collaboration, and RGM are consistently represented across solution and marketplace listings1011217.
- Deployment & integration. Public Guide and platform pages, plus training content, show batch ingestion, real-time APIs, and GraphCube as the modeled store; adapters/connectors for SAP/Oracle are mentioned across platform collateral691516.
- Training & configuration. The o9 Academy catalog and training pages mention Platform Architecture, GraphCube Server, IBPL select queries, and role-based technical configuration—indicative of a configuration-centric rollout methodology1516.
- Customer go-lives. Press releases document production rollouts like Li Auto (Jan 22, 2025)5.
Architecture, data model & platform services
- Enterprise Knowledge Graph & Graph-Cube. The Platform page and solution pages (RGM, sustainability) repeatedly cite an EKG and Graph-Cube enabling multi-granularity analytics with hierarchical dis/aggregation and “digital twin”-style modeling678. A platform article explains using customer data lakes alongside Graph-Cube, clarifying that Graph-Cube is a specialized in-memory store for supply-chain analytics while lakes/warehouses remain the broader source-of-truth34.
- APIs & UI access. The Guide portal explicitly states “public UI APIs … used to access the GraphCube server data and its model,” and lists Reference Model APIs for batch ingestion9.
- Cloud posture. o9 publishes Azure Marketplace offers and Google Cloud Marketplace availability; multiple o9/AWS collaboration posts and analyst notes reflect standard hyperscaler playbooks for scale and procurement1011121422.
Optimization, ML & “AI” claims
- Native planning + cloud ML integrations. o9 publicizes Vertex AI Forecast integration (Google) and broader GenAI collaborations (Microsoft), positioning the platform as “open” to external ML and foundation models1720.
- Composite agents (GenAI). o9’s 2024 release describes composite agents that execute complex cross-functional planning; technical specifics (agent orchestration, grounding, guardrails, evaluation) are not fully detailed in public docs19.
- Optimization stack. o9 references advanced algorithms and lists Gurobi as a partner; however, reproducible formulations (objective functions, stochastic treatment, constraints) or solver architecture for MEIO/production scheduling are not published; customers likely see recommendations rather than raw solver artifacts1867.
- Net assessment. The presence of ML/optimization is credible (press, partner ecosystems, training), but implementation depth remains a black box to outsiders. Where uncertainty modeling and decision economics (e.g., cost-aware stochastic policies) are crucial, o9’s public materials are less specific than research-oriented platforms that publish their modeling stack6211920.
Technology stack signals
- Languages/frameworks. Public docs emphasize platform-level services (APIs/ETL/Graph-Cube) rather than runtime languages. The Data Science (PaaS) page markets support for Python/R/PySpark for custom analytics embedded in o9 workflows21.
- APIs & integration. Secure REST APIs, direct in-memory loaders, adapters for SAP/Oracle; Platform Guide pages list UI APIs and Reference Model APIs for batch ingestion9.
- Cloud & security. Marketplace listings (Azure) and AWS collaboration posts imply standard hyperscaler security/compliance postures; Google Cloud partner directory and Marketplace procurement docs reflect vendor onboarding and commercial models1011141335.
Deployment & change management
- Rollout playbook. Training modules (IBP Functional → Technical), configuration artifacts (models, layouts, IBPL queries), and API integrations point to implementation projects where o9 and SI partners configure models, wire data feeds, and build dashboards; the Li Auto go-live supports this cadence15165.
- Customer data lakes + Graph-Cube. o9 explicitly explains dual-storage: the customer’s lake remains system-of-record; Graph-Cube is used tactically for supply-chain analytics/decisions (the article addresses duplication concerns and design intent)34.
Discrepancies & open questions
- Algorithmic transparency. Despite frequent AI/ML/optimization claims, technical documentation that would allow third parties to reproduce forecasting/optimization behavior (e.g., distributional forecasting over demand/lead-time, MILP/MINLP formulations, stochastic search) is not public; evidence is indirect (partner pages, training, solution overviews)621181920.
- EKG/Graph-Cube semantics. Naming is consistent across pages, but schema/typing, incremental update mechanics, and persistence guarantees are not described beyond marketing-grade pages and Guide blurbs6978.
- Automation boundary. It is unclear from public materials how far closed-loop automation goes (e.g., auto-creation of POs/transfer orders vs. decision-support hand-offs); marketplace pages suggest decision support with integrations, not transactional execution101167.
Conclusion
What o9 delivers (technical, non-promotional): a cloud-hosted APS with an EKG/Graph-Cube data layer, ETL/APIs for ingestion, packaged apps for IBP/supply/revenue planning, and configuration frameworks (IBPL and model/report/layout designers) to tailor workflows. The system surfaces recommendations and plans across hierarchies and time horizons, and can integrate external ML/solvers. Procurement via Azure/Google marketplaces and collaborations with AWS align with enterprise IT expectations.
How it achieves outcomes (mechanisms & evidence):
- Modeling of entities/hierarchies in Graph-Cube (confirmed across platform/solution pages and the Guide’s API description);
- Connectivity (ETL, in-memory loaders, REST APIs) and configurable UIs;
- Optimization/ML via a mix of native features and partner integrations (Vertex AI Forecast, Gurobi);
- Operational rollout supported by training paths and partner SIs.
State-of-the-art assessment: o9’s platform engineering and cloud posture (marketplaces, API surface, in-memory store) are contemporary and credible. However, algorithmic transparency around uncertainty modeling and decision optimization remains limited in public sources. By contrast, a programming-first competitor like Lokad publishes technical details of its probabilistic + optimization stack and execution VM. For organizations prioritizing suite breadth, hyperscaler procurement, and packaged process coverage, o9 fits well. For organizations prioritizing white-box probabilistic optimization and code-level control, a DSL-centric platform like Lokad represents a different trade-off.
Sources
-
General Atlantic invests in o9 Solutions — Jan 19, 2022 ↩︎ ↩︎ ↩︎
-
Existing investors add $116M at $3.7B valuation — Jul 19, 2023 ↩︎ ↩︎ ↩︎
-
Platform — Technology powering the Digital Brain — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Revenue Growth Management — Solution page — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Environmental Footprint — Solution page — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
o9 Guide — Platform/UI APIs (landing) — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Azure Marketplace — o9 Integrated Business Planning — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Azure Marketplace — o9 Revenue Management ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
o9 on Google Cloud Marketplace — Announcement (Aug 2020) ↩︎ ↩︎ ↩︎ ↩︎
-
Google Cloud — Partner Directory: o9 Solutions — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎
-
o9 and AWS expand collaboration — Feb 2024 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
o9 Academy — IBP Technical Training — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
o9 launches Vertex AI Forecast integration (2021) ↩︎ ↩︎ ↩︎ ↩︎
-
Gurobi — Technology Partner: o9 Solutions — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
GenAI composite agents added to Digital Brain — Jul 1, 2024 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
o9 expands Microsoft collaboration for GenAI — Apr 16, 2024 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Data Science (PaaS) — Python/R/PySpark — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Futurum Group — o9 + AWS collaboration analysis (2024) ↩︎ ↩︎ ↩︎ ↩︎
-
Intelligence360 — Notice of exempt offering ($116M) — Aug 16, 2023 ↩︎ ↩︎
-
Lokad Docs — Envision Language — retrieved Sep 2025 ↩︎ ↩︎ ↩︎
-
Architecture of the Lokad platform — retrieved Sep 2025 ↩︎ ↩︎ ↩︎ ↩︎
-
Envision VM (Part 2): Thunks & Execution Model — Nov 22, 2021. ↩︎ ↩︎ ↩︎
-
Envision VM (Part 4) : Distributed Execution — Dec 6, 2021. ↩︎ ↩︎ ↩︎ ↩︎
-
Lokad Solutions (landing) — solution scope & case studies (retrieved Sep 2025). ↩︎
-
How to use your data lake for storage (Graph-Cube rationale) — retrieved Sep 2025 ↩︎ ↩︎
-
Google Cloud Marketplace — Partner concepts & procurement — retrieved Sep 2025 ↩︎