Databricks entering the Excel connectivity market is meaningful market validation. It confirms what Exponam has built a company around: governed, business-user access to lakehouse data from Excel is not an edge case or a transitional requirement. It is a permanent, strategically important capability that enterprises need.

We have spent three years in production enterprise deployments solving exactly this problem. Databricks has spent considerably more on engineering and market resources than we have. So we want to give an honest account of where the new product performs, where it falls short, and what the right choice is depending on your situation.

What Databricks built

---

The Databricks Excel Add-in uses the SQL warehouse endpoint — the same compute layer that powers Databricks SQL notebooks and BI tool connections — to query data and return results into Excel. Users browse Unity Catalog, apply filters, and import data into worksheets through a task pane interface. They can also write SQL directly using =DATABRICKS.SQL() cell functions. Unity Catalog governance applies throughout. The product is a web-based Office Add-in, running on Windows, macOS, and Excel for the web using the same cross-platform JavaScript framework.

For a first-version public preview, it is a credible product. It will serve some organizations well — particularly small teams already operating inside a Databricks workspace who need occasional data in a spreadsheet.

Where Exponam.Connect is different

---

The foundational architectural difference is the data retrieval path. Exponam.Connect offers users a choice: access data via Delta Sharing — the open protocol that retrieves compressed Parquet files directly from cloud object storage, bypassing Databricks compute entirely — or via a SQL warehouse endpoint when full SQL syntax is required. The SQL endpoint path, currently in private preview, gives users arbitrary joins, window functions, and subqueries. The Delta Sharing path gives them speed and zero compute cost.

A note on Delta Sharing availability. The zero-compute path requires that Delta Sharing be enabled in your Databricks environment. Most current enterprise deployments have it active, but it is subject to internal approval processes and is not universal. Where it has not yet been approved, the SQL endpoint path is available as an alternative.

Databricks’ add-in uses the SQL endpoint only — there is no zero-compute access mode. Every data pull consumes DBUs.

On Windows, Exponam.Connect uses the VSTO framework — the native Microsoft COM-based integration layer — rather than the web-based Office Add-in model Databricks uses on all platforms. The practical result: data writes into Excel are significantly faster for large datasets, as the trial results below demonstrate.

A direct comparison

---

The table below covers the dimensions that matter most in an enterprise evaluation. We have tried to be accurate about both products, including areas where they are genuinely equivalent or where Databricks holds an advantage.

Dimension	Exponam.Connect	Databricks Excel Add-in
Installation	Installer link at exponam.com. Under five minutes to first data access. ▲ Exponam advantage	Download manifest XML → edit file → create shared folder → configure Trust Center → restart Excel. 10+ steps; fails silently in many corporate IT environments.
Data retrieval	User choice: Delta Sharing (zero DBUs, direct Parquet from cloud storage) or SQL endpoint (full SQL syntax, compute consumed). Both Unity Catalog governed. ▲ Exponam advantage	SQL warehouse endpoint only. Every import consumes DBUs.
Performance & data volume	~11 seconds per 1M rows on Windows. Full 2,879,789-row trial dataset retrieved successfully. ▲ Exponam advantage on Windows	Trial testing returned 948,650 rows from the same 2.88M-row dataset before producing a non-descriptive error. ▲ Validate before deployment
Cost structure	Transparent, volume-tiered license: $10/user/month at 100 users, scaling to $0.50/user/month at 100,000. Fixed and predictable. ▲ Exponam advantage at scale	No add-in license fee. But trial testing observed 16 DBU ($14.58) consumed in a single day of casual analyst use. ▲ Databricks advantage for very small teams only
SQL query	Dynamic SQL editor (private preview) via SQL endpoint. Unity Catalog Views always available. — Parity achieved	SQL query editor in task pane; =DATABRICKS.SQL() and =DATABRICKS.Table() cell functions.
External access	.share files work without Databricks workspace accounts. Partners and clients access governed data without workspace provisioning. ▲ Exponam advantage	Requires a Databricks workspace account for every user.
ML model execution	Exponam.AI runs Databricks ML Serving Endpoints as native Excel formulas. ▲ Exponam advantage — no equivalent	Not available.
UC governance	Full UC governance on both Delta Sharing and SQL endpoint paths.	Full UC governance. — Parity
Mac / web	Supported. Exponam.AI and advanced SQL query not yet available on Mac. — Comparable	Full support on macOS and Excel for web.
UI / usability	Designed for business users. Controls mirror Excel’s own conventions. ▲ Exponam advantage for business users	Designed by technologists. Comfortable for technical users, learning curve for business users. ▲ Databricks advantage for technical audiences
Platform scope	Databricks today. Snowflake and Azure Fabric on roadmap. ▲ Exponam advantage for multi-cloud	Databricks only.

What we observed in direct testing

---

In testing conducted for this comparison, both products were run against the same large enterprise-scale dataset. The results were unambiguous.

Product / path	Rows returned	Outcome
Exponam.Connect — Delta Sharing	2,879,789	✓ Complete. Full dataset retrieved successfully.
Exponam.Connect — SQL endpoint	2,879,789	✓ Complete. Full dataset retrieved successfully.
Databricks Excel Add-in	948,650	Incomplete. Retrieval stopped at ~one-third and returned a non-descriptive error.

Dataset: a large-scale enterprise transaction table available as a Delta Sharing trial dataset, containing approximately 2.88 million rows. Results from a single test run on standard hardware; individual results may vary.

Designed for different users

---

Exponam.Connect was designed by business users for business users. The ribbon integration, task pane layout, and interaction model are all built to mirror what Excel users already know. Filters are applied exactly as they are on a native Excel sheet — same gestures, same mental model. A finance analyst or operations manager can be productive within minutes of installation, without training or documentation.

The Databricks Excel Add-in was designed by technologists. Even the basic table-select workflow resembles a BI platform’s report configuration interface. It is parameter-driven, query-centric, and structured around concepts that are second nature to a data engineer but unfamiliar to the typical Excel business user. A Databricks-credentialed analyst who spends time in notebooks will feel at home. Someone who does not will not.

The practical deployment implication. In most enterprises, the population of Excel business users outnumbers the technical Databricks user base by a wide margin. Exponam.Connect is the product they will actually use without hand-holding. The Databricks add-in will require change management and training investment that its “no additional license” pricing does not account for.

The cost question, with real numbers

---

The Databricks add-in carries no license fee, which looks attractive on first comparison. During testing, we ran four to five query imports in a single working day — casual, representative analyst use. The result: 16 DBU consumed, equating to $14.58 at the $0.70/DBU SQL Serverless list rate. For one analyst. One day. Light load.

Extrapolating to a realistic monthly cost. At a conservative 5 DBU/analyst/day and $0.70/DBU, that is $3.50/analyst/day — approximately $70/analyst/month on a standard 20-working-day basis. At 10 DBU/day the figure doubles to $140/analyst/month.

User count	Exponam ($/mo)	Databricks @ $25/user †	Databricks @ $70/user ‡
100 users	$1,000	$2,500	$7,000
1,000 users	$5,000	$25,000	$70,000
10,000 users	$10,000	$250,000	$700,000

† Conservative floor estimate. ‡ Based on observed trial rate of 5 DBU/analyst/day at $0.70/DBU (SQL Serverless list) × 20 working days. Actual costs vary by query frequency, warehouse type, dataset size, and contracted DBU rates.

The shared-workbook recalculation risk. Data imported via =DATABRICKS.Table() or =DATABRICKS.SQL() lives as a formula in the worksheet. Excel formula recalculation can silently re-execute those queries against a running SQL warehouse. In workbooks that circulate widely, this adds unbudgeted compute spend. Exponam.Connect imports data as static cell values; refresh is managed entirely through the ribbon and is not affected by any Excel recalculation event.

Where Databricks has a genuine advantage

---

Small teams with existing Databricks credentials. If your team is fewer than 100 people, already operating inside a Databricks workspace, and primarily running SQL-scale queries rather than million-row imports, the Databricks add-in is reasonable.

Mac-first organizations. On macOS, both products use the Office Web Add-in framework and the picture is more balanced. Exponam.Connect’s Delta Sharing cost advantage still applies, but the VSTO performance advantage does not.

Platform confidence. Databricks is a $62B company releasing a first-party product. Enterprises weigh vendor stability and support structure, and Databricks carries weight in that evaluation.

What comes next from Exponam

---

Databricks’ entry does not change our roadmap — it confirms it. Three capabilities currently in development:

Natural language / AI query. Users will be able to describe their data need in plain English and receive governed, reproducible results — with a private or BYO LLM option for regulated environments.

Automatic path optimization. AI-driven routing will select between Delta Sharing and the SQL endpoint on each query, optimizing automatically for cost and performance.

Multi-cloud expansion. Snowflake and Azure Fabric support are on the near-term roadmap. Databricks will build a good connector for Databricks data. They will not build one for Snowflake. We will.

The bottom line

---

Databricks entering the space is good news for the market. It validates the problem and raises awareness that governed, no-code lakehouse data access from Excel is available.

For organizations with more than 100 users, Windows-primary deployments, cost sensitivity at scale, external data sharing needs, or multi-cloud environments: Exponam.Connect is the stronger choice on the merits. For small teams already inside Databricks’ ecosystem who need light-duty access: the Databricks add-in is a reasonable starting point — with the caveat that data volume limits and DBU consumption should be validated before any broader rollout.

The full technical comparison — covering installation, architecture, performance, cost, governance, usability, and roadmap in depth — is available below.

---

© 2026 Exponam, LLC · exponam.com · info@exponam.com · +1.646.360.0110

Exponam is a Databricks Validated Technology Partner