cimt

Technology Partner · Snowflake

Snowflake partner for implementation, migration and operations

cimt sets up Snowflake, migrates your existing data warehouse and takes over operations. A partner since 2018, with agreed budgets and auto-suspend so costs stay predictable.

Snowflake AI Data Cloud Services Partner, Select tier Snowflake logo Snowflake Partner

cimt is a Snowflake services partner and is listed, as part of the cimt group, in the Snowflake Partner Directory. Our consultants design, build and run data platforms on Snowflake for clients in the Netherlands, Belgium and the UK, from lakehouse architecture and Data Vault 2.0 modelling to migration and managed services.

Snowflake is a cloud-native data platform that combines the benefits of a data warehouse (ACID, governance, SQL performance) with those of a data lake (cost, flexibility, unstructured data) in a single lakehouse architecture. cimt positions Snowflake as the storage- and-compute layer in our reference architecture, combined with Data Vault 2.0 for modelling flexibility and Qlik Talend Cloud for data integration. Multi-cloud (AWS / Azure / GCP), consumption-based pricing, and near-zero maintenance. We implement, model, migrate and manage end-to-end.

See also: A good data platform: why it pays off, What goes into an SLA for data environments? and What is a retainer in managed services?

Core benefits

Why Snowflake

Four properties that set Snowflake apart from legacy data warehouses and generic cloud database instances.

Consumption-based pricing

Pay only for storage and compute you use: no upfront investment, no tuning per warehouse.

Multi-cloud

Run on AWS, Azure or GCP and replicate data across clouds without vendor lock-in. Snowflake is working with STACKIT on a European option.

Near-zero maintenance

No indexes, no vacuum, no tuning. Snowflake manages the infrastructure.

Data sharing

Share data safely with partners and customers via Marketplace and Private Data Exchange, without creating copies.

AI on Snowflake

AI agents on your data, within agreed rules

At Expedition 2026 in November, Snowflake announced capabilities that let AI agents work with approved definitions and within your existing permissions. Two of them bear directly on data governance.

Public preview coming soon

Business Ontology

Data stewards record what terms like 'revenue' or 'active customer' mean and link those definitions to tables. AI agents use the approved definition first, so three teams asking the same question get the same answer.

Generally available

Cortex AI Gateway

One entry point for every agent. Snowflake records what an agent did and attributes AI spend to a user, team or agent. Checking each tool call against your existing permissions is in public preview.

Where cimt helps

The definitions themselves

The technology is there. The work is in the agreements: who owns 'revenue', which definition applies and who approves a change. We record them following DAMA DMBoK, so agents work with the same numbers as your reports.

See also: Semantic layer: why AI guesses without agreed definitions, EU AI Act: what does this mean for your data? and our approach to AI data readiness.

Data in Europe

Snowflake and STACKIT

Snowflake runs in EU regions, so your data can stay in Europe. In September 2026, Snowflake and STACKIT, the European cloud of the Schwarz group, announced a step-by-step integration of their platforms: data in the open Apache Iceberg format on STACKIT, with the keys held in Europe.

cimt already builds a sovereign data platform along these lines. Your data sits in Iceberg on STACKIT storage that you manage yourself, Snowflake processes it in an EU region and language models run on STACKIT AI Model Serving in Germany. Revoke the keys and Snowflake can no longer read your Iceberg tables.

See the sovereign data platform →

Data Vault 2.0

Scalable modelling on Snowflake

We implement Data Vault 2.0 as the modelling methodology on Snowflake: Hubs, Links and Satellites that admit new sources without breaking existing models, with full audit trail.

Hubs, Links, Satellites

Standardised modelling structure that admits new sources without breaking risk.

Full audit trail

Every change traceable, essential for compliance and governance.

Parallel loading

Inherently parallelisable, optimal for Snowflake multi-cluster compute.

Source-agnostic

Integrate ERP, CRM, IoT, APIs and flat files into one uniform model.

Our Data Vault implementations are supported by erwin Data Modeler for visual model design and Qlik Talend Cloud for automated pipeline generation.

cimt Snowflake services

From first design to ongoing operations

Service Description More
Architecture & design Lakehouse reference architecture, Data Vault 2.0 modelling, zone layout (Raw / Curated / Consumption) View →
Data integration ETL/ELT pipelines via Qlik Talend Cloud, CDC, API integration, Snowpipe streaming View →
Migration Migration of on-premises data warehouses (Oracle, SQL Server, Teradata) to Snowflake: parallel run, validation, rollback View →
Analytics & BI Qlik Sense dashboards on Snowflake data, direct query and import mode View →
Managed Services Ongoing management, monitoring, cost optimisation and warehouse tuning View →

Snowflake for your organisation

Request Snowflake licence advice

In a first conversation we look at your current data warehouse situation and which Snowflake deployment model fits, including a first cost indication.

cimt + Snowflake

Frequently asked questions about Snowflake

What is Snowflake?

Snowflake is a cloud data platform for storing, processing and sharing data. It combines the traits of a data warehouse (SQL, transactions, governance) with those of a data lake (low-cost storage, semi-structured data) and runs as a service on AWS, Azure or Google Cloud. You manage no servers: storage and compute are separate and scale independently. Organisations use Snowflake for reporting, analytics, data sharing with partners and, increasingly, AI on their own data.

What does Snowflake cost?

On demand, Snowflake costs $2.60 to $5.20 per credit, plus $23 per TB of storage per month (Azure West Europe region, Snowflake price list, November 2026; Snowflake prices in USD). The price per credit depends on the edition: $2.60 for Standard, $3.90 for Enterprise and $5.20 for Business Critical. An XS warehouse uses 1 credit per hour, each size up doubles that, and you pay per second the warehouse runs, with a one-minute minimum each time it starts. Smaller items such as data transfer and serverless features can come on top. A capacity contract, where you buy in advance, can bring lower rates. The bill depends mostly on how long and how large your warehouses run. cimt keeps it under control with the right warehouse size, auto-suspend and resource monitors that hold each warehouse or team to a budget.

Is Snowflake a data warehouse or a lakehouse?

Both. Snowflake started as a cloud data warehouse and now also supports lakehouse patterns: semi-structured and unstructured data, and tables in the open Apache Iceberg format that other tools can read too. Whether you use Snowflake as a warehouse or as a lakehouse is an architecture choice. cimt makes that choice per organisation, usually with layers that run from raw source data to trusted data for reporting and AI.

Is Snowflake suitable for smaller organisations?

Yes. On demand you pay per second of use, with no minimum commitment. An XS warehouse that runs two hours each working day uses about 44 credits a month: roughly $170 of compute in the Enterprise edition in the Azure West Europe region, plus storage. For a smaller organisation the risk lies less in the price per credit than in an environment without agreements on operations and cost. That is why cimt pairs a small start with fixed agreements on auto-suspend, budgets and who may do what.

Can I migrate my existing data warehouse to Snowflake?

Yes. cimt has extensive experience with migrations from Oracle, SQL Server, Teradata and other platforms. Our approach covers schema conversion, data movement, query rewrite and parallel validation. We use Data Vault 2.0 to lay a future-proof modelling layer during the migration: not just "as-is" relocation but modernising as we move.

How does Snowflake compare to Databricks?

Both are powerful cloud data platforms with different focus. Snowflake excels at SQL-based analytics, data warehousing and data sharing. Databricks is stronger in data engineering with Spark and ML workloads. cimt picks Snowflake when the primary need is analytics and BI, and when a SQL-first approach fits your organisation and team skills.

Snowflake or Microsoft Fabric: which fits us?

That depends mostly on your existing landscape. Microsoft Fabric fits well if your organisation already runs largely on Microsoft, with Power BI and Azure as its base. Snowflake fits better if you combine several clouds or sources, want to share data with customers or partners, or want to separate and charge back compute per team. The two do not exclude each other: Power BI also works on Snowflake. cimt makes that assessment with you based on your sources, teams and costs.

Do we have to use Data Vault 2.0 on Snowflake?

Not strictly. Star schemas and dimensional models work fine on Snowflake too. Data Vault 2.0 pays back especially with multiple source systems, long historical retention, compliance requirements (audit trail) and teams that want to develop in parallel. For simple use cases we often recommend star schema; for enterprise scope almost always Data Vault.

Can Snowflake run in Europe, for example on STACKIT?

Snowflake runs in EU regions, so your data can stay in Europe. In September 2026, Snowflake and STACKIT announced a step-by-step integration of their platforms, with data in the open Apache Iceberg format and the keys held in Europe. cimt already builds a sovereign data platform along these lines: data in Iceberg on STACKIT, processing in Snowflake. More on our data sovereignty page.

Can AI agents work safely with our Snowflake data?

Yes, once two things are in place. The first is permissions: with Cortex AI Gateway every agent goes through one entry point, Snowflake checks each tool call against your existing permissions (in public preview), and you see what an agent did and what it cost. The second is definitions: an agent that does not know what 'revenue' means in your organisation makes up its own answer. Business Ontology, with public preview coming soon, lets you record those definitions in Snowflake. That second part is mostly governance work, and it is usually where cimt starts.