Snowflake’s Acceleration Delivers High-Quality Growth Across Broad Customer Base in Q2 2027

Snowflake’s Acceleration: Quality, Durability, and the Role of Coco and CoWork

Overall assessment

The acceleration appears higher quality and more durable than a narrow, AI-speculation-driven spending surge, although the evidence is still predominantly operational and qualitative rather than a fully disclosed financial bridge. Management says growth is coming from a broad customer base rather than being concentrated in AI-native companies, which remain a small portion of revenue, and that customers are adding workloads across migrations, core data operations, applications, AI functions, and agentic products.¹²

The strongest evidence of quality is the apparent flywheel between Coco/CoWork adoption and Snowflake’s core platform consumption: accounts using Coco consume more of the core platform, while AI users overall consume more across Snowflake’s data platform.³⁴ Management also reports that the time for new-logo customers to reach 80% of purchased consumption has “very visibly improved” for newer cohorts, suggesting faster realization of value rather than merely larger contractual commitments.⁵

However, the durability case is not yet conclusive. Management explicitly said it had observed only “a couple of quarters” of the relevant behavior, did not disclose exact uplift figures, and provided no detailed cohort economics or retention data specific to Coco and CoWork.³⁶ The appropriate conclusion is therefore encouraging evidence of durable consumption expansion, but not yet proof of a mature, long-cycle revenue stream.

Quality of the acceleration

1. Broad-based rather than concentrated

The acceleration is not being attributed primarily to a small group of AI-native or venture-backed customers. Management described it as coming from a “very broad swath” of customers and said AI-native companies remain a small part of Snowflake’s overall revenue stream.¹ Snowflake also reported 14.6 thousand customers, 692 net new customers in the quarter, and 65 customers with more than $10 million of trailing 12-month product revenue, indicating activity across both new and large established accounts.⁴

That matters because broad-based adoption generally carries less concentration risk than growth dependent on a narrow set of early-stage companies with uncertain budgets. It also appears that the acceleration is extending into large enterprises: examples cited include BlackRock and Block increasing their mission-critical workloads, an Australian bank processing 17 billion transactions on Snowflake, and a large network-equipment manufacturer completing a Teradata migration in less than three quarters rather than the previously expected two to three years.⁷⁸⁴

2. The acceleration is partly core-platform growth, not only new-product revenue

Management estimated that AI products—including Coco, CoWork, AI functions, and the AI Gateway—contributed approximately half of the acceleration, while other areas such as notebooks, Streamlit or React applications, migrations, and workloads coming onto the platform also showed robust growth.² This is important: the investment case is not simply that Snowflake is selling two new AI interfaces, but that those interfaces may increase the velocity and volume of work performed on the underlying data platform.

The reported 89% year-over-year increase in customer use cases or individual projects deployed on Snowflake further supports the idea that expansion is occurring at the workload level, not merely through product trials.⁹ Management also said customers using AI on Snowflake consume more across the broader data platform, describing AI activation as a “structural multiplier” for platform consumption.⁴

3. Evidence of faster customer value realization

The reported improvement in the time required for new-logo cohorts to reach 80% of purchased consumption is one of the more persuasive indicators of quality.⁵ Faster time to value can improve the likelihood of renewals, expansions, and additional workload migration because customers are using their purchased capacity sooner.

There is also evidence that adoption is progressing beyond experimentation. CoWork surpassed 9.1 thousand accounts and added more than 2,000 net new accounts during the quarter, while Cortex Code reached 5.8 thousand accounts and grew nearly 11% sequentially.¹⁰ These figures demonstrate substantial adoption, although account counts alone do not establish usage intensity, gross-margin contribution, or long-term retention.

4. Durability indicators are positive but incomplete

Management cited several factors supporting durability:

• Snowflake has observed the relevant consumption behavior for multiple quarters and bases its guidance on observed behavior rather than only pipeline assumptions.³
• Accounts adopting Coco are consuming more of Snowflake’s core platform, creating a reported product-to-platform flywheel.³
• Gross retention was described as relatively flat across the last several quarters, rather than deteriorating as adoption expands.³
• Management believes the breadth and depth of use cases argue against merely pulling projects forward from future periods.⁶
• Snowflake is selling to a broader set of personas, including CFOs, CROs, CMOs, and CEOs, which can expand the number of budget owners and use cases supporting the platform.³

The limitations are equally important. Management did not provide exact uplift numbers, only cohort-level observations, and the period of observed behavior is still relatively short.³⁶ The excerpts also do not provide net revenue retention, dollar-based expansion specifically for Coco/CoWork users, product-level gross margins, customer payback periods, or evidence of renewal behavior through a complete annual cycle. Consequently, the current evidence supports strong early durability, but an analyst should not treat it as definitive proof that current acceleration will persist at the same rate.

Why Coco and CoWork fit supply-chain use cases

Coco and CoWork appear well suited to supply-chain work because supply chains combine large and fragmented data estates with operational decisions that require business context, cross-functional analysis, and sometimes direct action.

1. They reduce the distance between enterprise data and operational decisions

Management’s core argument is that AI has “massively shrunk the distance between data and value,” allowing business users to obtain useful answers and take action more quickly.¹¹ Snowflake positions Coco and CoWork as a governed layer through which knowledge workers and builders can use enterprise data through conversational language, including for supply-chain operations.⁸

Supply-chain analysis often requires users to combine data from inventory, suppliers, logistics, orders, production, and demand forecasts. The excerpts do not provide a detailed supply-chain product specification, so it would be excessive to claim that Coco or CoWork automatically solves every such workflow. The more supportable point is that Snowflake’s architecture places the data, business context, models, governance, and workflows in one environment, which is strategically relevant to cross-functional supply-chain questions.⁸¹²

2. Coco is useful for exploration, debugging, and technical acceleration

Coco can be directed at a slow query, asked to debug the 10 longest-running queries, or used to identify the most idle warehouses; management said cost management is one of its top 10 skills.¹ These capabilities are directly relevant to supply-chain data environments, where performance, data freshness, query efficiency, and the cost of repeatedly processing large operational datasets can affect the economics of analytics and AI workloads.

Coco also accelerates migrations and development. Snowflake cited an example in which a large network-equipment manufacturer was executing a Teradata migration in less than three quarters, versus an estimated two to three years previously.⁷ Faster modernization can bring supply-chain data into a common platform sooner, enabling downstream planning, forecasting, and operational use cases.

3. CoWork is better suited to repeatable business workflows

CoWork’s positioning is broader than a developer assistant: it allows users to interact conversationally with governed enterprise data and enables actions such as sending emails, summarizing Slack conversations, and opening Jira tickets through integrations.¹³ That action orientation is relevant to supply-chain processes because the value is not only in producing an analysis; it is also in routing exceptions, communicating with stakeholders, and initiating workflow steps.

Management cited supply-chain optimization as an example of a use case that would not previously have been considered by Snowflake and said AI can now support sophisticated actions that might previously have required a dedicated application, a multiquarter implementation, and a staged rollout.⁶ This suggests CoWork can function as a faster path from an operational question to a usable workflow, while Snowflake’s governed data layer reduces the need to move sensitive enterprise data into a separate, less-controlled environment.⁸⁶

Why Coco and CoWork fit finance use cases

Finance is an especially credible reference market because Snowflake says it is using these products internally across finance-related functions, not merely selling them externally.

1. Finance has broad internal adoption

Snowflake reports more than 150 internal “Snowflake on Snowflake” use cases, with Coco used in deal desk, tax and accounting, internal audit, FP&A, and treasury.¹⁴ Management also said adoption within the finance organization is “almost at 100%,” which provides stronger evidence of practical utility than a marketing demonstration alone.¹⁴

Snowflake further said its finance organization reduced long-range planning from a three-person team using more than 50 spreadsheets to one analyst using a series of models reflecting pricing structure and consumption dynamics.⁹ This example indicates that the products can help finance teams consolidate fragmented planning processes and connect models more directly to operational and consumption data.

2. Finance use cases benefit from governed, contextual data

Finance decisions depend on consistent definitions, auditability, permissions, and institutional knowledge. Snowflake’s stated product direction includes Cortex Sense, which captures business definitions and institutional knowledge and provides that context when an agent answers a question.¹³ The platform also emphasizes enterprise security, governance, and observability, including visibility into what AI is doing, how it performs, and what it costs.¹³

That combination is strategically important for finance because a fast answer is not sufficient if the underlying metric definitions, data lineage, or authorization are unclear. The excerpts support the view that Snowflake is attempting to make Coco and CoWork useful within governed financial processes rather than positioning them solely as generic chat interfaces.¹³⁸

3. Finance is a credible expansion path beyond traditional analytics

Management said it is now regularly speaking with CFOs and that CFOs, along with CROs, CMOs, and CEOs, are increasingly involved in purchasing decisions.³ Finance therefore matters in two ways: it is an internal proof point for Snowflake’s own transformation, and it represents a senior budget-owning function that can expand adoption beyond the data team.

The company’s own use of CoWork in marketing, sales, finance, and other functions also gives its sales team a practical demonstration of business transformation rather than an abstract AI pitch.¹¹⁹¹³ That should improve credibility in enterprise sales, although customer references and actual production outcomes remain more important than internal adoption claims.

Are there signs of inefficient Coco or CoWork spending?

Evidence against widespread inefficient spending

The excerpts contain several positive signals against irrational or poorly controlled consumption:

1. Optimization is built into the product and sales motion. Coco can identify slow queries, long-running queries, and idle warehouses, while cost management is described as one of its top 10 skills.¹
2. Snowflake actively encourages efficient customer spending. Management said the company stresses efficient spend with every customer and that sales representatives view identifying waste as a trust-building exercise that can lead to additional projects.¹
3. The platform includes cost and governance controls. The Cortex AI Gateway is designed to route tasks to appropriate models based on policies and performance data, with cost and governance controls built in.¹³
4. Usage appears connected to business outcomes. Examples include eliminating $400 thousand of annual agency spend, reducing keyword research from approximately 10 hours to 20 minutes, reducing content production from approximately 24 hours to two hours, and reducing long-range planning from a three-person team to one analyst.⁹
5. Management is measuring penetration and adoption methodically. Snowflake says it tracks Coco penetration from legal approval through daily usage and uses in-product guidance and customer labs to drive outcomes.¹⁵

These points suggest that the company is not simply encouraging customers to consume resources indiscriminately. Its stated model is to help customers optimize workloads, achieve measurable labor or external-spend savings, and then expand into additional projects.¹¹⁵⁹

What is not yet known

There is no explicit evidence in the excerpts of irrational customer behavior, poor operational hygiene, or material waste in Coco or CoWork consumption. Management did not identify a pattern of customers overbuying capacity, running uneconomic agents, or abandoning pilots after spending heavily.¹¹⁶

But the absence of a disclosed problem is not the same as proof that spending is efficient. The excerpts do not quantify:

• Coco or CoWork revenue and gross margin;
• inference or model costs per customer;
• customer-level payback periods;
• the proportion of accounts that are daily active users;
• the frequency of unused or underutilized agent deployments;
• whether high adoption is concentrated among a few heavy users;
• whether customers are consuming more because of productive workloads or because AI workloads are intrinsically more compute-intensive.

Management’s statement that cost optimization is a top Coco skill is encouraging, but it also implicitly acknowledges that Snowflake workloads require active cost management.¹ Similarly, the AI Gateway’s model-routing and cost-control functionality suggests that economics and governance remain ongoing operating considerations rather than solved problems.¹³

Bottom line

Snowflake’s acceleration appears to be high quality relative to a narrow AI-led spending spike because it is broad-based, includes established enterprises and core migrations, is associated with faster customer time to value, and appears to increase consumption of the underlying data platform.¹⁸²⁵⁴

Coco and CoWork are credible solutions for supply chain and finance because they combine conversational access to governed enterprise data with technical acceleration, business context, workflow execution, and the ability to move from analysis to action.¹¹¹³⁸⁶ Supply-chain value comes primarily from connecting complex operational data to faster decisions and workflows, while finance value is supported by Snowflake’s own internal adoption across FP&A, treasury, accounting, tax, audit, and planning.¹⁴⁹

There are no clear signs in the excerpts of inefficient Coco/CoWork spending, and Snowflake describes a deliberate emphasis on optimization, cost controls, and measurable business outcomes.¹¹⁵¹³ Nevertheless, the durability thesis remains a developing one: the company has disclosed strong adoption and favorable cohort behavior, but not enough product-level financial and usage data to establish long-term returns or rule out pockets of uneconomic consumption.³⁶