A brief history of Snowflake

Snowflake Inc. —founded in 2012 and based in San Mateo, California—is a cloud-based data-warehousing company. Its unique architecture separates data computation from storage, enabling businesses to scale up or down their data usage needs. Companies use Snowflake to optimize cost, performance, and flexibility.

The company experienced rapid growth since its inception, due to an innovative approach to data warehousing. Snowflake went public in 2020 in what was one of the largest software IPOs at the time. Among other factors, the company's success can be attributed to its robust product offering, user-friendly platform, and strategic leadership decisions.

Snowflake's impact on the data industry and data observability has been substantial. By offering a cloud-native, scalable, and easy-to-use data warehousing solution, Snowflake has democratized access to data analytics for businesses of all sizes. Newer features, like real-time data sharing and the ability to process diverse data types, have pushed the boundaries of what's possible in the realm of data analytics. Moreover, Snowflake's success has sparked increased competition in the cloud data warehousing space, leading to faster innovation and better products for customers. Here's a stroll through the company's history and a look at how they've impacted the data sector.

2012: Founding and public unveiling

Snowflake Computing, now known simply as "Snowflake", was founded in 2012 by three data warehousing experts: Benoit Dageville, Thierry Cruanes, and Marcin Zukowski.

Dageville and Cruanes had previously worked at Oracle as data architects, where they witnessed firsthand the limitations of traditional on-premise and cloud data platforms. Zukowski was the co-founder of Vectorwise, a database company that was acquired by Actian. All three recognized the constraints of existing solutions, including scalability issues, complex management, and the inability to effectively handle a growing volume of data generated by businesses.

The founders envisioned a data warehouse built for the cloud from the ground up. One that could leverage the flexibility of cloud compute and storage to provide dynamic and scalable data storage and analytics.

They wanted Snowflake to be accessible and user-friendly, ensuring that businesses of all sizes could analyze their data without needing to manage complex hardware or software configurations. Their goal was to separate compute from storage, allowing users to scale each independently based on their needs, a key differentiator for Snowflake.

Snowflake was kept in stealth mode until June 2015, when it unveiled its cloud data warehousing platform. The platform was met with enthusiasm from businesses seeking to leverage their data for insights, driving Snowflake's growth in the following years.

Separating storage and compute

Snowflake's architecture is unique because it separates storage and compute resources, a major departure from traditional data warehouse design where these are intrinsically linked. This separation is crucial for several reasons:

Scalability: By separating compute from storage, Snowflake allows users to scale each independently. Users can store an infinite amount of data (limited only by the capacity of their cloud provider) and can also scale up or down the compute resources as needed, depending on the power required for their queries. Users can process large data workloads quickly by adding more compute resources, and when not in use, they can be scaled down or turned off to save costs.

Performance: The separation also ensures that heavy queries don't slow down the system. Each user or workload gets its dedicated resources, preventing resource contention. This means that a large, complex query won't slow down a small, simple one, improving overall system performance.

Cost-effectiveness: With Snowflake, you pay for storage and computation separately. If your data is large but your queries are infrequent or simple, you can save money by keeping compute resources low. Conversely, if your data size is small but the complexity and frequency of queries are high, you can opt for more compute power. This flexibility results in cost savings and optimal resource utilization.

Concurrency and accessibility: Traditional data warehouses struggle with many users accessing the system concurrently. In Snowflake's model, multiple compute nodes simultaneously access the same data without performance degradation, supporting high levels of concurrency.

Flexibility and ease of use: Snowflake handles all aspects of operations, like data distribution, data partitioning, and query optimization, in a way that's transparent to the end-user. Users don't need to manage indices, partition data, or perform other administrative tasks common in other systems. They just load their data and start querying.

Early leadership

In 2014, Bob Muglia was appointed CEO of the company, a position he held until 2019. During his tenure at Snowflake, Muglia played a pivotal role in driving the company's growth and establishing its foothold in the cloud data warehousing market.

One of Muglia's significant contributions was shaping Snowflake's value proposition of the cloud-native data warehouse separating storage from compute resources. Muglia also championed the user-friendliness of Snowflake's platform, pushing for a system that required less maintenance and was easy for data analysts and scientists to use. The strategy proved successful, skyrocketing the amount of users that adopted Snowflake's technology.

Fundraising

In October 2015, shortly after the public unveiling of its product, Snowflake raised $45 million in a Series C funding round. The company kept attracting significant interest from venture capital, securing a $100 million Series D round in 2017.

Series A - 2012: Snowflake raised $5 million in its initial round of funding. The round was led by Sutter Hill Ventures.

Series B - 2014: The company secured $26 million in a Series B round led by Redpoint Ventures, with participation from Sutter Hill Ventures.

Series C - 2015: Snowflake raised $45 million in a Series C round led by Altimeter Capital, with participation from Redpoint Ventures, Sutter Hill Ventures, and Wing Ventures.

Series D - 2017: The company secured $100 million in a Series D round. ICONIQ Capital led the round, with participation from Madrona Venture Group, Redpoint Ventures, Sutter Hill Ventures, Wing Ventures, and Altimeter Capital.

Series E - 2018: Snowflake raised a hefty $263 million in a Series E round, which valued the company at $1.5 billion. This round was led by ICONIQ Capital, Altimeter Capital, and Sequoia Capital.

Series F - 2018: Later the same year, Snowflake raised another $450 million in a Series F round, bringing its post-money valuation to $3.5 billion. The round was led by Sequoia Capital, with participation from existing investors.

Series G - 2020: In its final funding round before its IPO, Snowflake raised $479 million at a valuation of $12.4 billion. Dragoneer Investment Group and Salesforce Ventures led the round, with participation from existing investors.

Product innovation and expansion

In 2016, Snowflake introduced Snowpipe, a service that loads data in a continuous, real-time manner as the data arrives in the cloud. A couple of years later, they launched Snowflake Data Sharing, enabling direct sharing of live data across Snowflake accounts without the need for ETL processes or copying data. Notably, in 2019, Snowflake introduced the Snowflake Data Exchange, a marketplace for data providers and data consumers.

Current leadership

In 2019, Muglia stepped down as CEO and was succeeded by Frank Slootman. While the reasons for the leadership change were not detailed publicly, it marked a shift in the company's trajectory, as Slootman was known for successfully leading companies through IPOs. Notwithstanding his departure, Bob Muglia's impact on Snowflake's early growth and market positioning remains a significant part of the company's history.

Slootman previously led ServiceNow and Data Domain to successful IPOs. This change indicated Snowflake's intentions to go public. Under Slootman's leadership, Snowflake has experienced remarkable growth. By 2020, it boasted several high-profile customers like Adobe, Sony, and Capital One, among others.

IPO

In February 2020, Snowflake raised $479 million in a funding round that valued the company at $12.4 billion. This funding round included Salesforce Ventures as a new investor, highlighting the company's impressive growth and potential.

On September 16, 2020, Snowflake debuted on the New York Stock Exchange under the ticker "SNOW." Priced at $120 per share, the company's shares began trading at $245, a more than 100% increase, making it the largest software IPO ever at that time.

Between 2015 and the 2020 IPO, Snowflake exhibited steady product innovation, rapid customer growth, and successful fundraising activities, establishing itself as a significant player in the cloud data warehousing space.

Key product expansions and partnerships

Snowpipe

Introduced in 2016, Snowpipe is a service that enables Snowflake users to load data in real time. It allows for automated and continuous loading of data as soon as it arrives in a cloud storage bucket. This makes it ideal for use cases where up-to-date data is critical, such as real-time analytics and monitoring.

Snowflake Data Sharing

This feature, launched in 2017, revolutionized how companies share data. With Snowflake's unique architecture, data can be shared between different Snowflake accounts instantly, securely, and without having to move or copy the data. This reduces the complexity, cost, and risk associated with traditional data sharing methods.

Snowflake Data Marketplace

Launched in 2020, the Snowflake Data Marketplace takes data sharing a step further. It provides a platform where providers can publish data sets, and consumers can access and analyze this data directly from their Snowflake account. The data remains in its original, live state, ensuring it is always up-to-date.

Snowpark

Announced in 2020, Snowpark allows data engineers, data scientists, and developers to write code in their languages (like Java, Scala, and Python) and execute workloads directly on Snowflake. This enables more types of data-intensive applications to leverage Snowflake's powerful platform, further expanding its capabilities beyond just data warehousing.

Polaris Catalog

Unveiled at Snowflake Summit 2024, Polaris Catalog is a vendor-neutral, open-source Apache Iceberg REST Catalog released under the Apache 2.0 license. Snowflake donated the project to the Apache Software Foundation in August 2024, where it became Apache Polaris. The catalog establishes an open standard for Iceberg table management that works across Databricks, Google Cloud, Microsoft Azure, Confluent, dbt Labs, and other ecosystem participants, allowing enterprises to manage Iceberg data without being locked into any single platform. Initial partners included Google Cloud, Microsoft, Confluent, Fivetran, and Collibra.

Cortex AI

Cortex is Snowflake's AI suite: a set of SQL-accessible large language model functions, retrieval capabilities, and agent orchestration tools that run natively on the Snowflake platform. Cortex LLM Functions reached general availability in May 2024, giving any Snowflake user access to hosted language models via simple SQL calls. Snowflake's own Arctic LLM, an open-weight enterprise model released under Apache 2.0 in April 2024 and optimized for SQL generation and coding tasks, was made available through Cortex. Cortex Analyst (natural language to SQL for business self-service) and Cortex Search (hybrid semantic and keyword retrieval) followed later in 2024.

Snowflake Intelligence

Snowflake's flagship agentic AI product reached general availability in November 2025. It gives business users a conversational interface to query structured and unstructured Snowflake data in natural language, powered by Cortex Agents orchestrating Cortex Analyst and Cortex Search. More than 1,000 customers deployed over 12,000 agents during the preview period. For a fuller account of the AI developments that led to Snowflake Intelligence and the changes announced at Summit 2026, see the "Becoming the AI Data Cloud" section above.

Partnerships

Strategic partnerships have been central to Snowflake's growth, enabling the platform to run across cloud environments, connect to enterprise software ecosystems, and more recently, power AI agent capabilities.

Cloud Providers (2020): Amazon Web Services, Google Cloud Platform, and Microsoft Azure are foundational partners for Snowflake. These partnerships enable Snowflake to offer its cloud-based data warehousing solution across different cloud platforms, giving customers flexibility in their infrastructure choices.

Salesforce (2020): In 2020, Salesforce became a strategic partner to Snowflake. Not only did Salesforce invest in Snowflake, but the two companies also announced integrations between Salesforce's Customer 360 platform and Snowflake. This allows mutual customers to unify and analyze their data in real time, and also provides access to Snowflake's Data Marketplace for additional data insights.

Alation (2021): Snowflake invested in data catalog provider Alation in 2021 and has awarded them partner of the year at the annual Snowflake Summit multiple years in a row. Their joint customers include large accounts such as Cisco, DocuSign, Expedia, and PepsiCo.

Anthropic (2025–2026): Snowflake's partnership with Anthropic deepened significantly in 2025 and 2026. Claude models power CoWork (formerly Snowflake Intelligence) and CoCo (formerly Cortex Code), Snowflake's two flagship AI agent products. Anthropic co-founder and President Daniela Amodei co-keynoted Snowflake Summit 2026, reflecting the strategic importance of the relationship to Snowflake's agentic AI roadmap.

AWS infrastructure commitment (2026): At Snowflake Summit 2026, Amazon Web Services announced a $6 billion, five-year infrastructure commitment to the Snowflake platform, cementing AWS as Snowflake's primary cloud infrastructure partner for its AI era and signaling the scale of enterprise AI workloads expected to run on the combined stack.

Becoming the AI Data Cloud: 2023 to 2026

The period from 2023 to 2026 marked the most significant strategic shift in Snowflake's history, from a cloud data warehousing company to an enterprise AI platform. The shift began with an acquisition, accelerated through a leadership change, and culminated in a suite of agentic AI products that redefined the company's market position.

Neeva acquisition and the AI leadership pipeline (May 2023)

In May 2023, Snowflake acquired Neeva, an AI-powered enterprise search startup co-founded by former Google SVP Sridhar Ramaswamy, for approximately $150 million. The deal brought Neeva's search and large language model technology into the Snowflake platform, and brought Ramaswamy himself into the company as SVP of AI. Within nine months, he would become CEO.

Summit 2023: the AI pivot begins (June 2023)

At Snowflake Summit 2023 in Las Vegas, with roughly 12,000 in-person attendees, the company unveiled its first wave of generative AI capabilities. Document AI (built on technology from Snowflake's 2022 Applica acquisition) enabled extraction of insights from unstructured documents directly inside the platform. Snowpark Container Services entered preview, allowing containerized workloads including custom ML models to run natively within Snowflake's security perimeter. Iceberg Table support was announced, establishing Snowflake's position on the open table format that would become strategically central the following year. Frank Slootman's keynote framing ("To have an AI strategy, you have to have a data strategy") defined the company's positioning through the transition that followed.

CEO transition: Sridhar Ramaswamy (February 2024)

On February 27, 2024, Snowflake announced that Frank Slootman was retiring as CEO. Ramaswamy, who had been building Snowflake's AI organization since the Neeva acquisition, was named his successor. Slootman remained as Chairman of the Board. The transition was explicitly framed as a move toward an AI-first product strategy, building on the infrastructure work Ramaswamy had been leading since joining the company.

Snowflake Arctic (April 2024)

In April 2024, Snowflake released Arctic, an open-weight enterprise LLM, under the Apache 2.0 license. The model used a Mixture-of-Experts architecture with 480 billion total parameters but activated only 17 billion at inference time, making it more computationally efficient than comparably performing models available at the time. Built in under three months at a fraction of the cost of comparable models, Arctic was optimized specifically for enterprise tasks (SQL generation, coding, and retrieval-augmented generation) rather than general-purpose conversation. It was made available through Snowflake Cortex for production AI applications.

Summit 2024: The AI Data Cloud and the Iceberg strategy (June 2024)

At Snowflake Summit 2024 in San Francisco, the company officially rebranded its platform from "The Data Cloud" to "The AI Data Cloud," reflecting the strategic direction Ramaswamy had been driving since taking the CEO role. The summit's most strategically significant announcement was Polaris Catalog: a vendor-neutral, open-source Apache Iceberg REST Catalog, released under Apache 2.0 and donated to the Apache Software Foundation by August 2024 (now called Apache Polaris). The timing was deliberate: Databricks announced its acquisition of Tabular, the company founded by Apache Iceberg's original creators, the same week. Snowflake's Polaris move signaled a commitment to open data formats independent of any single vendor.

Cortex Analyst, which enables business users to query Snowflake data in natural language, entered public preview at the same summit. Snowpark Container Services reached general availability on AWS in August 2024, followed by Cortex Search in October 2024.

The first billion-dollar quarter (May 2025)

For the quarter ended April 30, 2025, Snowflake reported product revenue of $1.004 billion, its first-ever billion-dollar quarter. The result exceeded analyst guidance by approximately $40 million and represented roughly a fourfold increase in quarterly revenue since the company's 2020 IPO.

Snowflake Intelligence goes GA (November 2025)

Snowflake Intelligence, the company's flagship AI agent product, reached general availability on November 4, 2025, alongside Cortex Agents. Snowflake Intelligence gives business users a conversational interface to query structured and unstructured data in natural language, powered by Cortex Agents orchestrating Cortex Analyst and Cortex Search. More than 1,000 customers had deployed over 12,000 agents in preview before the GA release, making it one of the fastest enterprise AI product adoptions in Snowflake's history.

Acquisitions extending the platform (2025–2026)

Snowflake completed a series of acquisitions that extended the platform beyond its data warehouse roots. In November 2025, Datometry (whose product enables workload migration from competing data warehouses without rewriting queries) was acquired from Dell Technologies Capital, with its technology integrated into Snowflake's SnowConvert AI. Also in November 2025, Snowflake acquired Select Star, a metadata intelligence and data lineage platform, integrating its capabilities into Horizon Catalog. In January 2026, Snowflake announced the acquisition of Observe, an AI-powered observability platform, for $1 billion, expanding into the broader IT operations management market. In May 2026, Snowflake announced the acquisition of Natoma, a startup building a governance gateway for the Model Context Protocol, specifically to enforce identity verification, access policies, and audit trails for AI agents taking actions across business applications.

Summit 2026: "Making AI Real for Business" (June 2026)

Snowflake Summit 2026, held in San Francisco with over 20,000 attendees, centered on the theme "Making AI Real for Business." The company announced more than 26 new capabilities, with governed agentic AI as the central thesis: that enterprise AI requires a semantic context layer and data governance infrastructure, not just compute. Snowflake Intelligence was rebranded as CoWork, with features including deep research, user memory, and integration with iOS, Slack, and Microsoft Office. Cortex Code was rebranded as CoCo, an autonomous coding agent. Cortex Sense was introduced as a runtime semantic context layer, benchmarked at 86% accuracy on structured business questions versus 24% for generic models. Snowflake also launched AI Agent Identity at general availability, providing cryptographic identity per agent, per-agent RBAC, and full audit trails. AWS announced a $6 billion, five-year infrastructure commitment to the Snowflake platform at the event.

Snowflake and Bigeye

As Snowflake Intelligence, Cortex Agents, and other agentic capabilities become embedded in enterprise data workflows, the question of whether those agents can be trusted to act reliably on business data has become central to how organizations run AI in production. An agent that queries a Snowflake table with stale or anomalous data, or accesses sensitive fields without appropriate controls, produces results that carry those problems forward, with no error signal in the agent's output.

Bigeye's Agent Trust Hub connects Snowflake Intelligence activity to the data trust signals that determine whether an agent's actions are reliable: data quality status, data classification, lineage, governance, ownership, and policy. Teams can see which agents are querying which tables, whether those tables had active quality issues or freshness failures at the time of the query, and whether sensitive data was involved in the interaction. AI Guardian provides enforcement at the query layer, blocking agents from accessing fields classified as restricted before those queries execute.

For teams evaluating the trust posture of their Snowflake Intelligence deployment, or looking to bring AI agent activity into a registry, a 30-day free trial of Agent Trust Hub is available at bigeye.com/agent-trust-hub. No waitlist required.