Review · Read 13 October 2026
DataPelago review: the DataPelago Accelerator for Spark
ETL Compare staff · Figures checked September 2026 · Sourced from vendor docs, project pages and public benchmarks
In brief
The DataPelago Accelerator for Spark (DPA-S) is a commercial accelerator built on the DataPelago Nucleus engine that states existing Spark code runs unchanged and is sold pay as you go through the Google Cloud and AWS marketplaces. It scores 3.14 out of 5 and ranks 8th of 9 on our lead ranking, because specific managed services are not named, shuffle, spill and write behavior are not described and its evidence is vendor-published.
3.14 / 5
ETL fit score (editorial assessment, 0-5)ETL fit score (editorial assessment, 0-5)
Rank 8 of 9
Compute-heavy Spark SQL and DataFrame jobs on Parquet or ORC in Google Cloud or AWS
| Approach | Commercial Spark accelerator (DPA-S) built on the DataPelago Nucleus engine, using CPU, GPU or FPGA hardware |
|---|---|
| Runs on | Multiple Spark versions on Google Cloud and AWS; Amazon EMR, EKS and Databricks not named on the DPA-S page |
| Instances | Any available hardware, CPU to GPU (vendor statement) |
| License | Commercial |
| Price | Pay as you go through the Google Cloud and AWS marketplaces; the AWS Marketplace listing shows a vCPU-hour dimension and a $100,000 per month one-month contract entry |
| Code changes | None stated; standard Spark settings |
| Read | 13 October 2026 |
Sources: datapelago.ai, DPA-S page, DataPelago press page, DataPelago company page, AWS Marketplace listing · Fetched 27 Sep 2026
How does it score on each criterion?
| Criterion | Weight | Score | Why |
|---|---|---|---|
| Adoption effort | 20% | 3.8 | DataPelago states existing Spark code runs unchanged with no data migration, configured through standard Spark settings and bought through the Google Cloud or AWS marketplace. |
| Stage coverage | 20% | 3.0 | Its page names good fits (compute-intensive jobs, Parquet, ORC, SQL and DataFrame) and says to avoid heavy RDD and UDF use; shuffle, spill and write behavior are not described. |
| Tuning and operating burden | 20% | 3.0 | A vendor-run accelerator configured with standard Spark settings; tuning guidance and support terms are not published on the pages we reviewed. |
| Platform and instance portability | 10% | 3.2 | Supports multiple Spark versions on Google Cloud and AWS and any hardware from CPU to GPU, by the vendor's statement; specific managed services such as Amazon EMR, Dataproc or Databricks are not named. |
| Cost model transparency | 5% | 3.0 | Pay as you go through cloud marketplaces; the AWS Marketplace listing shows a vCPU-hour dimension and a $100,000 per month contract entry, but no per-job rate you can model without a quote. |
| Published evidence | 15% | 3.0 | Vendor claims of 10x faster at 80% cost reduction, customer quotes with figures (Akad Seguros, ShareChat) and a benchmark page against NVIDIA cuDF, all vendor-published. |
| Maturity and community | 10% | 2.6 | The Accelerator for Spark launched on 5 August 2025, on the Nucleus engine unveiled in October 2024; offices in Mountain View and Hyderabad; no open-source community of its own. |
What is DataPelago?
DataPelago sells the DataPelago Accelerator for Spark (DPA-S), built on its Nucleus engine, which uses heterogeneous CPU, GPU and FPGA hardware. DataPelago states it runs on any available hardware, CPU to GPU, supports multiple Spark versions on Google Cloud and AWS, and is configured through standard Spark settings. It builds on Substrait-based open-source frameworks such as Gluten, also targets Trino, and supports PySpark, Scala and Spark SQL.
The Accelerator for Spark launched on 5 August 2025, on the Nucleus engine unveiled in October 2024. The company lists offices in Mountain View and Hyderabad.
Reported (secondary): NetApp announced on 16 July 2026 that it acquired DataPelago, which will operate as a wholly owned NetApp subsidiary (NetApp press release seen as a search result; not opened). It does not affect the score.
Sources: datapelago.ai, DPA-S page, DataPelago press page, DataPelago company page · Fetched 27 Sep 2026
What does DataPelago claim?
Vendor states
10X faster at 80% cost reduction.
Source: datapelago.ai · Fetched 27 Sep 2026
Vendor figure. Not measured by ETL Compare.
This is DataPelago's figure. We have not measured it. DataPelago also publishes customer quotes with figures (Akad Seguros, ShareChat) and a benchmark page against NVIDIA cuDF, all vendor-published.
What does it cost?
Pay as you go through the Google Cloud and AWS marketplaces. The AWS Marketplace listing shows a vCPU-hour dimension and a $100,000 per month one-month contract entry; there is no per-job rate you can model without a quote.
Sources: AWS Marketplace listing · Fetched 27 Sep 2026
Where is it strong?
- DataPelago states existing Spark code runs unchanged with no data migration.
- Bought through the Google Cloud or AWS marketplace rather than a separate contract process.
- A stated fit you can check against your own jobs: compute-intensive, Parquet, ORC, SQL and DataFrame.
What are the watch-outs?
- Its page says to avoid heavy RDD and UDF use.
- Shuffle, spill and write behavior are not described.
- Amazon EMR, Dataproc and Databricks are not named on the DPA-S page.
- Tuning guidance and support terms are not published on the pages we reviewed.
Who should shortlist it?
Teams in Google Cloud or AWS whose slow jobs are compute-heavy Spark SQL or DataFrame work on Parquet or ORC, with little RDD or UDF code.
Read 13 October 2026.
Which job stages does it address?
| Accelerator | Read | Transform | Shuffle | Spill | Write |
|---|---|---|---|---|---|
| DataPelago | Documented | Partial | Not stated | Not stated | Not stated |
- Documented: the vendor's own documentation says it addresses this stage
- Partial: indirect or experimental, or covered only by an end-to-end claim
- Not stated: not found on the pages we reviewed
This shows what each vendor says, not what we measured. Sources are listed on each review.
Good fit: compute-intensive, Parquet, ORC, SQL DataFrame; avoid heavy RDD and UDF use.
Source: DataPelago Accelerator for Spark page · Fetched 27 Sep 2026
Frequently asked questions
Is DataPelago open source?
No. It is a commercial product; DataPelago states it builds on Substrait-based open-source frameworks such as Gluten.
How is DataPelago priced?
Pay as you go through the Google Cloud and AWS marketplaces. The AWS Marketplace listing shows a vCPU-hour dimension and a $100,000 per month one-month contract entry.
Compare with
- Apache Gluten with Velox review
- RAPIDS Accelerator review
- DataPelago alternatives
- Apache Auron vs DataPelago
- DataFusion Comet vs DataPelago
- DataPelago vs DualBird
- DataPelago vs Flarion
- DataPelago vs Apache Gluten
- DataPelago vs Databricks Photon
- DataPelago vs RAPIDS Accelerator
- DataPelago vs Speedata
Before you buy
Find your slowest stage with Profile a slow Spark job, then work out the money with Spark cost per job, explained.
Sources
- DataPelago homepage · Fetched 27 Sep 2026
- DataPelago Accelerator for Spark page · Fetched 27 Sep 2026
- DataPelago press page · Fetched 27 Sep 2026
- DataPelago company page · Fetched 27 Sep 2026
- DataPelago on AWS Marketplace · Fetched 27 Sep 2026