2012 – 2026 · SNOW · CIK 1640147 · FY ends Jan 31
Snowflake Products — the AI Data Cloud
Snowflake is usually flattened to “a cloud data warehouse,” but the company sells one platform — which it brands the AI Data Cloud — organized as a stack of layers on a single engine. The engine separates storage from compute across AWS, Azure, and Google Cloud; an open-format and governance layer sits on top of it; a data-sharing network rides across deployments; a developer / app platform and a transactional engine run alongside; and the Cortex AI layer sits at the top. Every workload — analytical, transactional, developer, and AI — runs against the same substrate, and the platform is billed on consumption (credits for compute, plus storage), not per-seat licenses. Sourced from snowflake.com/product/, the FY2026 10-K, docs.snowflake.com, and Snowflake's acquisition press releases.
Jump to a layer: The engine · Open formats & governance · Data sharing · Developer & apps · Transactional · Cortex AI.
Roster row: Snowflake on /orgs/ · Related: Tech Filings.
How the layers stack on one engine
Read the diagram bottom-to-top. Data flows from the customer's existing systems into the engine — the multi-cloud substrate whose defining choice is the separation of storage from compute, so independent “virtual warehouses” scale without contending for each other. The open-format and governance layer lets the same data live in the open Apache Iceberg format under one governance model. The data-sharing network moves live data between accounts, regions, and clouds without copies. The developer / app platform and the transactional engine run alongside each other on the same substrate, and the Cortex AI layer — with Snowflake Intelligence and Cortex Agents as the conversational surface — sits at the top.
Diagram caption: the engine at the base is a fully managed, multi-cloud substrate (AWS, Azure, Google Cloud) whose defining architectural choice is the separation of storage from compute — data sits in cloud object storage while independent virtual warehouses run against it, so workloads scale without contending. On top of the engine: the open-format and governance layer (Apache Iceberg tables, the Apache Polaris / Open Catalog catalog, and the Horizon governance layer); the data-sharing network (Secure Data Sharing, the Marketplace, and Data Clean Rooms, all riding the cross-cloud Snowgrid mesh); the developer / app platform (Snowpark, Snowpark Container Services, Native Apps, Streamlit, Notebooks, Openflow) and the transactional engine (Unistore / Hybrid Tables and Snowflake Postgres) side by side; and the Cortex AI layer at the top, with Snowflake Intelligence and Cortex Agents as the natural-language and agentic surface.
The engine · Multi-cloud · GA since 2015
The core platform — the AI Data Cloud engine
What it is
The base of everything is a fully managed, cloud-native data platform that runs on the three major public clouds — AWS, Microsoft Azure, and Google Cloud. Its defining architectural choice is the separation of storage and compute: data sits in cloud object storage, while independent “virtual warehouses” (elastic compute clusters) run against it. Storage, compute, and the cloud-services layer scale independently, so multiple workloads run at once without contending for the same resources, and a customer pays for the compute and storage actually consumed. Per the FY2026 10-K, the platform “enables customers to consolidate data into a single source of truth to drive meaningful business insights, build data applications, and share data and data products, as well as applies AI for solving business problems.”
The “data warehouse → AI Data Cloud” repositioning
Snowflake began as a cloud data warehouse and is still widely described that way, but the platform has expanded to cover lakehouse, data-lake, data-engineering, application-hosting, transactional, and AI workloads — all on the same engine. The company now brands the whole thing the AI Data Cloud, and organizes its own product hub around four workload families: Data Engineering, Analytics, AI, and Applications & Collaboration. The current corporate narrative centers the Cortex AI layer (see below); the engine underneath is what makes that layer's promise — AI grounded in governed enterprise data — technically coherent. Snowflake's own framing on snowflake.com/product/ sums the platform up as “From Analytics and AI to Engineering and Apps—power it all in the AI Data Cloud.”
When it shipped
Founded 2012 by Benoit Dageville, Thierry Cruanes, and Marcin Żukowski; out of stealth in 2014 and generally available in June 2015 as the “Snowflake Elastic Data Warehouse.” Multi-cloud support (Azure, then Google Cloud) followed later in the 2010s, and the company went public on the NYSE in September 2020 (ticker SNOW, CIK 1640147).
How it relates to the others
Every other layer on this page runs on this engine. The open-format and governance layer, the data-sharing network, the developer / app platform, the transactional engine, and the Cortex AI layer are all capabilities of the one platform — not separate products a customer installs or integrates. That single-substrate design is the reason a customer can, for example, run a Cortex agent over data that another team just shared into the account, all under one governance model.
Open formats & governance
Open formats & governance — Iceberg, Polaris, Horizon
What it is
This layer lets the same data live in an open, vendor-neutral table format — rather than only Snowflake's proprietary format — while keeping a single governance model over all of it. It is Snowflake's answer to the “lakehouse” and the lock-in concern: read and write the open format, let other engines touch the same tables, and still govern everything in one place.
Key capabilities
Apache Iceberg tables
Native support for the open Apache Iceberg lakehouse table format, so Snowflake reads and writes the same tables other engines use rather than locking data into a proprietary format. Generally available since June 2024. Iceberg on snowflake.com.
Apache Polaris / Open Catalog
An open-source Iceberg catalog Snowflake announced in 2024 and open-sourced under Apache 2.0 in 2025, donating it to the Apache Software Foundation as Apache Polaris. Snowflake Open Catalog is Snowflake's managed service for it — vendor-neutral, REST-API, multi-engine. Polaris is open source.
Horizon Catalog
The unified governance layer — security, compliance, privacy, lineage, and discovery across both relational and Iceberg tables. Introduced at Summit 2023 and since extended to govern assets that live outside Snowflake. Horizon on snowflake.com.
How it relates to the others
This layer sits directly on the engine. The data-sharing network and the Cortex AI layer both honor Horizon's governance, so a policy set once (a masking policy, a row-access rule, an object tag) applies whether the data is queried analytically, shared with a partner, or read by an AI agent. Polaris / Open Catalog is the piece that lets engines other than Snowflake read and write the same Iceberg tables under that governance.
Data sharing & collaboration · The network effect
The data-sharing network — Secure Data Sharing, Marketplace, Clean Rooms, Snowgrid
What it is
One of Snowflake's most-differentiated surfaces: the ability to share live, ready-to-query data without copying or moving it, across accounts, regions, and clouds. Because every account sits on the same platform, a provider grants access to data in place and a consumer queries it directly — no ETL, no file transfer, no stale copy. This creates a network effect: the more organizations on the platform, the more valuable the shared-data surface becomes.
Key capabilities
Secure Data Sharing
The foundational primitive, available since 2017: share live, governed data between accounts with zero copies. The consumer always sees the provider's current data.
Snowflake Marketplace
A marketplace of third-party datasets and native apps, grown out of the Snowflake Data Exchange launched in 2019. Providers list data products; consumers connect them in minutes. Marketplace on snowflake.com.
Data Clean Rooms
Privacy-preserving collaboration: two parties analyze each other's data together without either side exposing the underlying rows. The native clean rooms became generally available in 2024, built on Snowflake's December 2023 acquisition of Samooha, and run as Native Apps.
Cross-Cloud Snowgrid
The mesh that interconnects accounts across regions and clouds (AWS, Azure, Google Cloud). Sharing, replication, and consistent governance all ride Snowgrid, which also underpins cross-cloud business continuity. Snowgrid on snowflake.com.
How it relates to the others
The Marketplace, Secure Data Sharing, and Data Clean Rooms all ride Snowgrid, and all honor the Horizon governance layer. Native Apps distributed through the Marketplace (next layer down) are the delivery vehicle for clean rooms and for third-party applications, so the sharing network and the developer / app platform are tightly coupled.
Developer & app platform
The developer & app platform — Snowpark, Native Apps, Streamlit
What it is
The layer for building, running, and distributing code and applications next to the data, instead of exporting data to a separate processing tier. Developers write Python, Java, or Scala that runs inside Snowflake, build full applications that run in a customer's own account, and ingest data through managed connectors — all on the same engine and under the same governance.
Key capabilities
Snowpark
The developer framework for Python, Java, and Scala — DataFrame APIs and user-defined functions that run against the data in place. Snowpark for Python reached general availability in November 2022. Snowpark on snowflake.com.
Snowpark Container Services
Run containerized and GPU workloads — including custom and third-party models — fully managed inside Snowflake. Generally available across AWS, Azure, and Google Cloud commercial regions.
Native Apps Framework
Build and distribute applications that run inside the customer's Snowflake account, sharing data and logic without the data ever leaving. Previewed at Summit 2023 and generally available since 2024; distributed through the Marketplace. Native Apps on snowflake.com.
Streamlit in Snowflake
The popular Python framework for turning scripts into interactive data apps, which Snowflake acquired in 2022 and now hosts natively inside the platform. Streamlit on snowflake.com.
Notebooks
An interactive development environment for data and AI teams — write and run code, visualize results, and build pipelines and models in one place, governed by the platform.
Openflow
Managed data ingestion and integration with hundreds of ready-to-use connectors, for both batch and streaming. Built on Snowflake's November 2024 acquisition of Datavolo (from the creators of Apache NiFi). Openflow on snowflake.com.
How it relates to the others
Snowpark is the substrate for much of the platform's AI and data-engineering work; Container Services is where custom models and the heavier Cortex workloads run; Native Apps are distributed through the Marketplace; and Streamlit and Notebooks are common front ends for surfacing Cortex AI features to end users. Openflow feeds the engine itself, getting source data in so every layer above has something to operate on.
Transactional · Unistore
The transactional engine — Unistore / Hybrid Tables, and Snowflake Postgres
What it is
A transactional (OLTP) capability that lets a single Snowflake account serve transactional and analytical workloads, rather than forcing a separate operational database alongside the warehouse. Architecturally distinct from the analytical engine — it's optimized for fast, high-concurrency single-row operations — but it shares the same platform, security, and governance.
Key capabilities
Unistore / Hybrid Tables
Unistore is Snowflake's name for unifying transactional and analytical data; it is powered by Hybrid Tables, a table type for fast point operations. Announced at BUILD 2024 and generally available on AWS since October 2024 (and on Azure since October 2025). Unistore on snowflake.com.
Snowflake Postgres
An enterprise-grade managed Postgres running on Snowflake, aimed at mission-critical AI and transactional systems. It came from Snowflake's 2025 acquisition of Crunchy Data, an open-source PostgreSQL specialist.
How it relates to the others
Both options run on the same platform and under the same Horizon governance as the analytical engine, so transactional and analytical data can sit side by side without a separate system in between. This is the layer that lets Snowflake pitch itself as the home for operational and agentic application state, not just the analytics back end — which matters most for the Cortex agents in the layer above, since agents need somewhere to read and write their working state.
Cortex AI · The AI layer · The current growth narrative
Cortex AI — the AI layer
What it is
The top of the stack: a fully managed, serverless layer that brings large-language-model inference and other AI directly to governed data. AI functions are callable in SQL over data that never leaves the platform's security perimeter, and a family of higher-level surfaces — an analyst, a search service, agents, and a conversational interface — sit on top of those functions. This is the layer driving most of Snowflake's recent product narrative, and it evolves fastest: the Cortex family is renamed and expanded release-over-release, so what follows reflects the current company-side framing on snowflake.com and docs.snowflake.com rather than any fixed snapshot.
The Neeva origin
Much of this direction traces to Snowflake's 2023 acquisition of Neeva, an AI-powered search startup. The deal seeded the company's AI and search work and brought in Sridhar Ramaswamy, Neeva's CEO, who became Snowflake's CEO on February 27, 2024, succeeding Frank Slootman (who retired). Snowflake's AI-first repositioning has happened largely under Ramaswamy. Ramaswamy named CEO.
Key capabilities
Cortex AI (LLM functions)
Serverless access to industry-leading LLMs — including Anthropic Claude, OpenAI, Meta Llama, and Mistral — plus embeddings, classification, transcription, summarization, and translation, all callable directly in SQL. Cortex LLM functions reached general availability in May 2024.
Cortex Analyst
Natural-language-to-SQL analytics: ask a question in plain English and get an answer over structured data, grounded in a semantic model so the SQL is trustworthy.
Cortex Search
Retrieval and RAG over a customer's unstructured text — the search service that finds the relevant passages an LLM or agent needs to answer grounded questions.
Cortex Agents
Multi-step agents that orchestrate across both structured and unstructured data, using Cortex Analyst and Cortex Search as tools. Generally available since November 2025. Cortex Agents docs.
Snowflake Intelligence
The natural-language, agentic interface to the whole platform — “all your knowledge, one trusted enterprise agent” — powered by Cortex AI functions, Cortex Analyst, and Cortex Search. Generally available since November 2025. Snowflake Intelligence.
Cortex Code & Document AI
Cortex Code is a data-native AI coding agent that understands a customer's enterprise data context, launched in November 2025. Document AI extracts structured data from unstructured documents. Cortex Code on snowflake.com.
How it relates to the others
Cortex runs on the governed data already in the platform and honors the Horizon governance layer, which is Snowflake's central pitch for enterprise AI: the model comes to the data inside the security perimeter, rather than the data being shipped out to a model. Snowflake Intelligence and Cortex Agents are the conversational and agentic surface that ties the analyst, the search service, and the LLM functions together — and they increasingly lean on the transactional layer below for the working state an agent needs to act.
Every major product, by layer
The platform's major products, grouped by layer in stack order (engine at the top of the table, Cortex AI at the bottom). The “Origin” column flags whether a product grew organically or arrived through an acquisition; the “Key concept” column is the one-phrase mental model for each.
| Product | Layer | Shipped / acquired | Origin | Current state | Key concept |
|---|---|---|---|---|---|
| Snowflake platform | The engine | 2014–15 | Organic | GA | Separated storage & compute across AWS, Azure, GCP |
| Apache Iceberg tables | Open formats | 2024 | Organic | GA | Read & write the open lakehouse format |
| Apache Polaris / Open Catalog | Open formats | 2024; OSS 2025 | Organic (donated to ASF) | GA | Vendor-neutral open Iceberg catalog |
| Horizon Catalog | Governance | 2023 | Organic | GA | One governance, lineage & discovery layer |
| Secure Data Sharing | Data sharing | 2017 | Organic | GA | Zero-copy live data sharing |
| Snowflake Marketplace | Data sharing | 2019 | Organic | GA | Third-party data & native-app marketplace |
| Data Clean Rooms | Data sharing | 2024 | Acquisition (Samooha, 2023) | GA | Privacy-preserving collaboration |
| Cross-Cloud Snowgrid | Data sharing | — | Organic | GA | Cross-region / cross-cloud mesh |
| Snowpark | Developer | 2022 (Python) | Organic | GA | Run Python/Java/Scala next to the data |
| Snowpark Container Services | Developer | 2024 | Organic | GA | Containerized & GPU workloads |
| Native Apps Framework | Developer | 2023–24 | Organic | GA | Apps that run inside the customer's Snowflake |
| Streamlit in Snowflake | Developer | Acquired 2022 | Acquisition (Streamlit) | GA | Python scripts → data apps |
| Openflow | Developer | 2025 | Acquisition (Datavolo, 2024) | GA (AWS) | Managed data ingestion / integration |
| Unistore / Hybrid Tables | Transactional | 2024 | Organic | GA | Transactional + analytical in one table |
| Snowflake Postgres | Transactional | 2025 | Acquisition (Crunchy Data, 2025) | GA | Enterprise managed Postgres on Snowflake |
| Cortex AI (LLM functions) | AI | 2024 | Organic | GA | LLM inference callable in SQL |
| Cortex Analyst | AI | 2024 | Organic | GA | Natural-language → SQL |
| Cortex Search | AI | 2024 | Organic | GA | Retrieval / RAG over text |
| Cortex Agents | AI | 2025 | Organic | GA | Multi-step agents over all data |
| Snowflake Intelligence | AI | 2025 | Organic (Neeva-seeded) | GA | Natural-language agentic interface |
| Cortex Code | AI | 2025 | Organic | Available | Data-native AI coding agent |
Years are sourced to Snowflake's own product and engineering pages, the FY2026 10-K, Snowflake documentation release notes, and the relevant acquisition press releases (Streamlit 2022, Neeva 2023, Samooha 2023, Datavolo 2024, Crunchy Data 2025). “GA” means generally available; a few capabilities reach GA on one cloud before another (Hybrid Tables, Openflow), in which case the earliest GA cloud is shown. The Cortex family is renamed and expanded frequently — verify the current state against the linked primary sources.
Read these primary sources
Most of this page is paraphrased from the URLs below. They are the authoritative places to read what Snowflake says about each product on its own pages, the canonical platform description from the most recent SEC 10-K, and the acquisition press releases that establish the origins of the acquired products.
Snowflake's own product surfaces
Per-feature pages on snowflake.com/product/ — the canonical source for each capability's company-side framing — plus the technical documentation root.
# Platform & product hub
https://www.snowflake.com/en/product/
https://www.snowflake.com/en/product/platform/
# Open formats & governance
https://www.snowflake.com/en/product/features/iceberg/
https://www.snowflake.com/en/product/features/open-catalog/
https://www.snowflake.com/en/product/features/horizon/
# Data sharing · developer & app · transactional
https://www.snowflake.com/en/product/features/marketplace/
https://www.snowflake.com/en/product/features/data-clean-rooms/
https://www.snowflake.com/en/product/features/cross-cloud-snowgrid/
https://www.snowflake.com/en/product/features/snowpark/
https://www.snowflake.com/en/product/features/native-apps/
https://www.snowflake.com/en/product/features/streamlit-in-snowflake/
https://www.snowflake.com/en/product/features/openflow/
https://www.snowflake.com/en/product/features/unistore/
# Cortex AI
https://www.snowflake.com/en/product/features/cortex/
https://www.snowflake.com/en/product/snowflake-intelligence/
https://www.snowflake.com/en/product/features/cortex-code/
# Technical documentation
https://docs.snowflake.com/SEC filings — the canonical platform description
The FY2026 10-K (filed March 2026, for the fiscal year ended January 31, 2026) business section is the canonical description of the platform and the consumption-revenue model.
# Most recent 10-K (FY2026) — business section
https://www.sec.gov/Archives/edgar/data/0001640147/000164014726000008/snow-20260131.htm
# Snowflake on EDGAR — full filing history
https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001640147
# EDGAR submissions JSON — programmatic filing index
https://data.sec.gov/submissions/CIK0001640147.jsonAcquisition press releases & investor relations
The origin of each acquired product, plus where product-trajectory commentary (Cortex adoption, consumption trends) is described each quarter.
# Acquisitions that became products
https://www.snowflake.com/en/news/press-releases/snowflake-announces-intent-to-acquire-streamlit-to-empower-developers-and-data-scientists-to-mobilize-the-worlds-data/
https://www.snowflake.com/en/news/press-releases/sridhar-ramaswamy-named-chief-executive-officer-of-snowflake/
https://www.snowflake.com/en/blog/snowflake-to-acquire-samooha/
https://www.snowflake.com/en/blog/snowflake-postgres-enterprise-ai-database/
# Polaris open-sourcing & Unistore GA
https://www.snowflake.com/en/blog/polaris-catalog-open-source/
https://www.snowflake.com/en/news/press-releases/snowflakes-unistore-unifies-transactional-and-analytical-data-with-the-general-availability-of-hybrid-tables/
# Investor relations — earnings, events, trajectory
https://investors.snowflake.com/Sources: Snowflake's own product pages and per-feature pages for each capability's company-side framing; the FY2026 10-K business section for the canonical platform description; docs.snowflake.com release notes for general-availability dates; and acquisition press releases for the acquired-product origins (Streamlit, Neeva, Samooha, Datavolo, Crunchy Data). Reporter coverage is cited under fair use (linked, not republished). Last updated May 2026.
Mungomash LLC · More org pages · Snowflake on /orgs/ · Tech Filings