Power BI vs Looker: DAX vs LookML, Pricing (2026)

Quick verdict

Short answer

Choose Power BI if you are a Microsoft organisation, want a low published per-user price, and your report authors are comfortable with a desktop tool, Power Query and DAX. Choose Looker if you want metrics defined once in version-controlled code by data engineers, queried live in a cloud warehouse such as BigQuery or Snowflake, and reused by every dashboard, embedded app and even other BI tools. Looker suits SQL-fluent teams with a modern warehouse; Power BI suits broader self-service reporting across many sources.

How we know: This comparison is research-based: licensing, editions, published prices, semantic modelling languages, SQL access, embedding and connectors were checked against Microsoft's Power BI pricing page and Microsoft Learn, and against Google Cloud's Looker pricing page, Looker documentation and the Google Cloud blog, in October 2026. We have not built models in either product for this page and make no performance claims.

Power BI is Microsoft's BI product and a workload of Microsoft Fabric. Authors build a semantic model in Power BI Desktop (Windows only): Power Query loads and shapes data, tables are related in a model, and measures are written in DAX. The model usually imports data into an in-memory engine, though DirectQuery and Direct Lake leave it in the source. Reports are published to the Power BI service and shared with Pro or Premium Per User licences, or through Fabric capacity.

Looker is Google Cloud's enterprise BI platform. Developers describe tables, joins, dimensions and measures in LookML files stored in a Git repository; business users then explore through a browser, and Looker writes SQL against the database for every query rather than storing extracts. Google sells it as Looker (Google Cloud core), managed from the Google Cloud console, and the older Looker (original), which can be Looker-hosted or customer-hosted. Looker is a different product from Google's free Data Studio (formerly Looker Studio, renamed in April 2026).

Side by side

Aspect	Power BI	Looker
Semantic layer	Semantic model built in Power BI Desktop: Power Query plus DAX measures	LookML code files: views, explores, dimensions and measures
Where the data is queried	Import (in-memory) by default; DirectQuery, Direct Lake and composite models available	In the database at query time; Looker generates the SQL
Authoring	Power BI Desktop (Windows only) plus browser editing in the service	Browser-based IDE for LookML; Explores and dashboards in the browser
Version control	Fabric Git integration at workspace level (Azure DevOps, GitHub); reports and semantic models are still in preview there	Git built in: each project is a repository, Development Mode works on branches
Pricing model	Published per-user prices (Pro, PPU) or Fabric F SKU capacity	Platform edition (Standard, Enterprise, Embed) plus user licences (Developer, Standard, Viewer); quote only
Free option	Free Power BI Desktop; free account cannot share	90-day trial instances of Looker (Google Cloud core)
Databases	Many connectors: relational databases, warehouses, files (Excel, CSV, Parquet) and online services such as Salesforce and Dynamics 365	More than 40 SQL dialects (BigQuery, Snowflake, Redshift, PostgreSQL, MySQL, SQL Server, Databricks, Oracle and others)
Reuse by other tools	XMLA endpoint on Premium, PPU and Fabric (Excel, SSMS, Tabular Editor and others)	Looker BI connectors for Power BI, Tableau Desktop, Excel, Google Sheets and Data Studio
Main trade-off	Cheap to start and broad self-service, but Windows authoring and models that can multiply per report	Strong governed metrics in code, but needs LookML developers, a capable SQL database and a sales quote

Key differences

Semantic layer: LookML code versus a Power BI semantic model

In Looker, the semantic layer is code. A LookML view maps a table, dimensions map columns or SQL expressions, measures define aggregates, and explores define the joins users can query. This is the pattern from Looker's reference documentation:

# LookML (view file)
dimension: id {
  primary_key: yes
  sql: ${TABLE}.id ;;
}

measure: total_revenue {
  type: sum
  sql: ${sales_price} ;;
  value_format_name: usd
}

Because the sql parameters contain your database's own SQL dialect, SQL users can read and review a LookML model easily, and every dashboard built on an explore uses the same definition of revenue.

In Power BI, the semantic model is built in Power BI Desktop. Measures are DAX expressions evaluated in a filter context; this example comes from Microsoft's CALCULATE reference:

-- DAX (Power BI measure)
Blue Revenue =
CALCULATE ( SUM ( Sales[Sales Amount] ), 'Product'[Color] = "Blue" )

A published Power BI semantic model can be shared by many reports and opened by other tools through the XMLA endpoint, so central modelling is possible. In our assessment the difference is in defaults: Looker forces a central, reviewed model before anyone explores, whereas Power BI lets each author build a model, which is quicker to start but can lead to several versions of the same metric unless a team curates shared models.

Query execution and SQL access

Looker does not hold a copy of the data. Each Explore query, dashboard tile or API call is translated into SQL and run on the connected database, so performance and cost depend on that database. Google lists more than 40 supported dialects. SQL users can also run ad hoc queries in SQL Runner (for Developer users; Viewer users do not get it).

Power BI imports data into its in-memory engine by default, which Microsoft recommends for interactivity, and uses DirectQuery or Direct Lake when data must stay in place. DirectQuery has documented limits, including a one million row cap on intermediate results and a four-minute query timeout in the service. Many connectors accept a hand-written SQL statement, and Power Query's View Native Query shows the SQL generated from your steps.

For a team with a cloud warehouse such as BigQuery or Snowflake, Looker's live model keeps one copy of the data; for data spread across files, SaaS apps and on-premises databases, Power BI's import model and connector library are more flexible. On a pay-per-query warehouse, remember that Looker's live queries are billed by the warehouse.

Licensing and pricing

Power BI publishes per-user prices for Pro and Premium Per User, and sells capacity as Fabric F SKUs. On capacities of F64 or larger, free-licence users can view content; on smaller capacities, every viewer needs Pro or PPU. Microsoft is retiring Power BI Premium P SKUs in favour of F SKUs.

Looker (Google Cloud core) pricing has two parts, per Google: a platform fee for the instance, by edition, and user licences by type. Standard is for organisations with fewer than 50 users; Enterprise adds private connections, customer-managed encryption keys, VPC Service Controls and higher API limits; Embed adds signed and cookieless embedding and the highest API limits. Each edition includes 10 Standard users and 2 Developer users. User types are Developer (LookML, SQL Runner, administration), Standard (dashboards and content, without LookML development) and Viewer (view, filter, drill, schedule and download). Google does not publish list prices; you contact sales, with annual commitments.

Embedding and serving other tools

Looker offers private embedding in all editions, while signed embedding (for customer-facing apps) is limited to the Embed edition. Its BI connectors, which Google documents as generally available, let Power BI, Tableau Desktop, Excel, Google Sheets and Data Studio read Looker models, so LookML can be the governed layer even where other tools draw the charts. Some of these connectors work only on Looker (Google Cloud core) or Looker-hosted instances.

Power BI embeds through Power BI Embedded (A SKUs) or F and P capacities. In the embed for your customers pattern, end users need no Power BI licence; in embed for your organisation, viewers need Pro or PPU unless the capacity is F64 or larger.

Platforms and ecosystems

Looker is browser-based, so authors can use any operating system. Looker (Google Cloud core) runs on Google Cloud; customer-hosted Looker (original) is the option for running it on your own infrastructure. Power BI Desktop is Windows only, and the Power BI service lives inside Microsoft Fabric, with Direct Lake access to OneLake (see Databricks vs Microsoft Fabric). In our view, the cloud your warehouse runs in is often the deciding factor: BigQuery shops lean to Looker, Azure and Fabric shops to Power BI.

Pricing and licensing

Power BI. Microsoft's Power BI pricing page, checked 7 October 2026, lists Power BI Pro at USD 14.00 per user per month, paid yearly. Premium Per User is a higher per-user tier, Fabric capacity is sold separately through Azure at regional prices, and Power BI Desktop is free. Prices exclude tax and vary by country.

Looker. Google's Looker pricing page, checked 7 October 2026, describes platform pricing by edition (Standard, Enterprise, Embed) plus user pricing by type (Developer, Standard, Viewer), and asks buyers to contact sales; no list prices are published. Trial instances of Looker (Google Cloud core) last 90 days. Instances created after 1 August 2026 no longer include the 50 free Data Studio Pro licences that earlier Looker instances received. Your database or warehouse bill for the queries Looker runs is separate.

Pricing checked on the vendors' official pages on 7 October 2026. Prices change; confirm before buying.

Where each one leads

Power BI strengths

Published, low per-user price, and Pro is included in Microsoft 365 E5
Very wide connector library, including files, SaaS apps and on-premises sources
In-memory import mode does not depend on the source database for each interaction
Part of Microsoft Fabric, with Direct Lake reports over OneLake tables

Looker strengths

One governed semantic layer in LookML, readable by SQL users
Git version control and Development Mode built in
Queries the database live, with no extracts to refresh
BI connectors let Power BI, Tableau, Excel and Sheets reuse Looker models
Browser-based authoring on any operating system

Limitations

Power BI limitations

Power BI Desktop is Windows only
Models can multiply across reports unless a team curates shared semantic models
DAX filter context is hard for SQL users to learn
Free-licence viewing needs an F64 or larger capacity

Looker limitations

No published prices; every purchase goes through Google sales
Needs LookML developers before business users can explore new data
Performance and cost depend on the database, because every query runs there
Signed embedding for customer-facing apps requires the Embed edition

When to choose each

Choose Power BI if

Your organisation runs on Microsoft 365, Azure or Fabric
You need many data sources, including Excel files and SaaS applications
Analysts in business teams build their own reports
You want a published per-user price without a sales process

Choose Looker if

Your data is already in a cloud warehouse such as BigQuery or Snowflake
You want metric definitions in version-controlled code reviewed by data engineers
You are building embedded analytics into a product at scale
Different teams use different BI tools and you want one governed model behind them

When neither is right

You want visual exploration with desktop authoring on macOS: see Power BI vs Tableau and Tableau vs Looker.
You want free, self-hosted SQL dashboards: see Power BI vs Metabase, Metabase vs Apache Superset and the best open source BI tools.
You only need ad hoc reports on Google data: Data Studio (formerly Looker Studio) is free; see Looker vs Looker Studio.

Final recommendation

Bottom line

Pick Power BI for broad, affordable self-service BI in a Microsoft organisation, accepting Windows-only authoring and the work of keeping shared models tidy. Pick Looker when governed metrics matter more than speed of first report: it rewards a team that writes SQL, owns a cloud warehouse and wants every number defined once in Git. The two can also coexist, since Looker's Power BI connector lets Power BI users report on LookML models. Whichever you choose, the warehouse underneath matters; see data warehouse vs database.

Frequently asked questions

What is the difference between LookML and DAX?

LookML is a modelling language that describes tables, joins, dimensions and measures, with SQL snippets inside; Looker turns it into SQL that runs on your database. DAX is a formula language for measures and calculations inside a Power BI semantic model, evaluated by Power BI's own engine (or translated to SQL in DirectQuery mode).

How much does Looker cost?

Google does not publish Looker list prices. Pricing combines a platform fee by edition (Standard, Enterprise or Embed) with user licences by type (Developer, Standard, Viewer), and you get a quote from Google sales. A 90-day trial instance is available for Looker (Google Cloud core).

Can Power BI use a Looker model?

Yes. Google documents a generally available Looker BI connector for Microsoft Power BI, alongside connectors for Tableau Desktop, Excel, Google Sheets and Data Studio, so Power BI reports can query LookML models.

Is Looker the same as Looker Studio?

No. Looker is the enterprise BI platform with LookML. Looker Studio was Google's free report builder, and Google renamed it back to Data Studio in April 2026, with Data Studio Pro as the paid tier.

Does Looker store a copy of my data?

Looker queries the connected database directly and generates SQL for each request, rather than importing data into its own engine. Power BI imports data by default, with DirectQuery and Direct Lake as alternatives.

Sources

Checked October 2026.

How we research comparisons: our editorial method.