Create an account →
Create an account →
// Pricing

Pay for data processed.
Not for hours burned.

Quanton bills on a single number: the volume of data your jobs read and write. Every speedup we ship cuts your infra bill while keeping your Quanton bill flat. Run in your own Kubernetes — or directly on VMs (EC2, GCE, Azure VMs, on-prem). Reserved and spot discounts stay yours.

Run the calculator → See the rate card
YOUR ENTIRE INVOICE
cost = GiB_processed × rate
First 100 GiB / month FREE
Read · 1–10 TiB band $0.0007 / GiB
Write 2× read
No DBU markup. No per-cluster premium. No support tier. Tier price drops as volume grows.
// Why per-GiB is fair
Faster jobs = less money for the vendor VENDOR REVENUE High Low Slow Fast JOB PERFORMANCE
Compute billing — 4× faster = 75% less revenue for them.
Per-GiB — speed is free to give. We make money by getting better.
01
Our incentive is your compute bill going down.
Every release we ship that runs Spark faster cuts your EC2 hours — but charges you the same fee. We win when you process more, not when each job drags on.
02
Volume tiers reward growth, not punish it.
Six tiers, from $0.0012/GiB at small scale down to $0.0003/GiB above 1 PiB. The more you process, the lower your per-GiB rate — automatically.
03
No DBU. No per-cluster premium. No support tier.
One line item. The EC2 you'd already buy sits in your own account at your own discount. Quanton sits on top — and that's the entire invoice.
// Plans
Free
$0 forever
Try Quanton on real workloads. No credit card. Falls back to free, open-source acceleration past the monthly allowance.
  • 100 GiB / month of Quanton-accelerated processing
  • After the cap — automatic fallback to OSS Spark with Apache DataFusion Comet and Apache Gluten pre-installed, also free
  • AI Spark Engineer included — works across every engine
  • Community support via Slack
  • Free on your laptop, free in a cluster — small data, no charge
Start free →
MOST POPULAR
Team
Volume-tiered $ / GiB
Every job on Quanton, fully accelerated. Pay only for the data you process — and the rate drops as you grow.
  • Unlimited Quanton-accelerated read/write — no throttling
  • First 100 GiB / month free, carried over from the Free tier
  • Six volume bands · from $0.0009 down to $0.0003 / GiB (see rate card)
  • AI Spark Engineer included
  • Community support via Slack
See the rate card →
Enterprise
Flat fee / year
One annual number — no per-GiB metering, no surprise overages. For orgs running petabytes and standing on compliance and procurement contracts.
  • Unlimited data volume for a single flat annual fee
  • Unlock Onehouse Managed Lakehouse
  • Dedicated support & CSM
  • Custom contracts
  • Volume discounts
Contact sales →
// Rate card
Volume tier
Read / GiB
Write / GiB
0 – 100 GiB / month free tier
FREE
FREE
100 – 1,000 GiB / month
$0.0009
$0.0018
1,000 – 10,000 GiB / month most workloads
$0.0007
$0.0014
10,000 – 100,000 GiB / month
$0.0005
$0.0010
100,000 – 1,000,000 GiB / mo
$0.0004
$0.0008
> 1,000,000 GiB / month
$0.0003
$0.0006
How bands stack

Each band's rate applies only to the GiB that lands inside it. The first 100 GiB of every month is free; the next chunks unlock the next band's lower rate — automatically, no upgrade needed. Volumes measured in GiB (1,024 MiB), matching what Spark reports. Write is 2× the read rate.

A 5 TiB month, priced in plain English
First 100 GiB free $0.00
Next 900 GiB · 100–1 K 900 × $0.0009 $0.81
Next 4,000 GiB · 1 K–5 K 4,000 × $0.0007 $2.80
Total read fee · 5,000 GiB $3.61
// Interactive · TCO calculator

Estimate your bill, side by side.

Inputs default to one TPC-DS 10 TB ETL run. Edit anything — read volume, runtime, cluster size, EC2 discount, runs per month — and the four columns update in real time.

Quick presets:
QUANTON · BYOC
$0.00
per run
Quanton metered fee$0.00
EC2 (your VPC)$0.00
Monthly:$0
DBX CLASSIC PHOTON
$0.00
per run
Photon DBU fee$0.00
EC2 (your VPC)$0.00
Monthly:$0
—
DBX SERVERLESS PHOTON
$0.00
per run
Serverless DBU$0.00
EC2 (bundled)included
Monthly:$0
—
OSS SPARK
$0.00
per run · 4.87× longer
Engine feefree
EC2 (4.87× hours)$0.00
Monthly:$0
—

Assumes identical hardware on Quanton, Photon Classic, and Serverless (workload is compute-bound). OSS Spark runtime projected from the TPC-DS 10 TB benchmark below (5.12× Quanton on the same cluster). DBU rates: Classic $0.15, Serverless $0.35, Premium tier, US-East-1. Estimates only.

// About these comparisons
Quanton vs the rest, on identical iron.

Quanton and Photon Classic run on the same hardware in the same VPC — your EC2 discount applies to both, so it cancels. Serverless bundles compute at the higher $0.35/DBU rate. OSS Spark is free, but burns 4.87× the wall-clock on the same cluster.

Click a preset above to see the worked breakdown for a real workload.

// Benchmark proof · TPC-DS 10 TB
Total runtime · 99 queries · TPC-DS 10 TB
11 × m8gd.4xlarge · any open format · same hardware throughout
Quanton Photon Open / accelerated Spark
OSS Spark
12,200s
5.12×
baseline
Comet (Tuned)
9,122s
3.83×
OOMs on q67, q93
Gluten (Tuned)
8,563s
3.59×
OOMs on q67, q93
Databricks Photon 18.2
2,550s
1.07×
closed, paid
Quanton
2,384s
1.00×
BYOC, per-GiB
Quanton and Photon converged on identical hardware over six quarters of engine work — Quanton runs 6.5% faster at this checkpoint, with no per-cluster premium.
2.0×
TPCx-BB · 1 TB
SF1000 Iceberg · 10 r8g.4xl executors · Spark 3.5. Quanton vs Apache Spark.
2.05×
TPC-DI · 1 TB
Python transformations + TPC-H-like SQL. Native rollup avoids 9× row replication.
4.1×
LakeLoader · 1 TB
MERGE INTO / CDC updates. Low-shuffle columnar MERGE scales with rows changed.
~4×
Iceberg / Hudi compaction
Native columnar compaction · ~75% less wall-time vs OSS row-wise rewrite.

Source: Inside Quanton — Storage-aware Spark design-partner deck. Independent re-runs welcome — Quanton runs in your own VPC.

// Side by side

Quanton vs every other way to run Spark.

Quanton
DBX Classic
DBX Serverless
OSS Spark
Apache DataFusion Comet
Apache Gluten
Pricing model
Per-GiB processed
DBU × compute hours
DBU × compute hours, EC2 bundled
Free engine, pay for EC2 hours
Free engine
Free engine
Runs in your VPC
Yes
Yes (BYOC)
No — Databricks-hosted
Yes
Yes
Yes
EC2 discounts (RI/spot) stay yours
Yes
Yes
No
Yes
Yes
Yes
Vectorized execution
Optimized and reimplemented Velox operators
Photon (closed)
Photon (closed)
No
OSS DataFusion operators
Vanilla OSS Velox
Storage-aware planning
Iceberg + Hudi metadata
Delta-focused
Delta-focused
No
No
No
Scan speedup
Yes — optimized I/O, lower scheduling overhead
Yes
Yes
No
No
No
Index-aware joins
Yes — new relational operator that cuts join cost
No
No
No
No
No
Query plan optimization
Advanced plan reshaping
Yes
Yes
Limited
Limited
Limited
Native columnar MERGE / compaction
Yes (~4× faster)
Delta only
Delta only
No
No
No
Dynamic acceleration
Yes — dynamic index maintenance
No
No
No
No
No
Memory-pressure resiliency
Optimized memory allocator, smart spilling
Yes
Yes
Spark default
OOMs on q67/q93
OOMs on q67/q93
AI Spark engineer in Spark UI
Yes — free, works across all engines
Limited
Limited
No
No
No
Reversible — point back to OSS Spark
One config line
Locked in by Databricks SQL extensions
Locked in by Databricks SQL extensions
N/A
Yes
Yes
Quanton Yes / strong Partial / limited No / blocker
// FAQ

Questions a skeptical platform team would ask.

I run in my own VPC and already get an EC2 discount — what's the edge over Databricks Classic?

EC2 cancels on Classic: same instances, same VPC, your discount applies to both sides. The honest comparison is platform fee only. On one TPC-DS 10 TB run, that's Quanton $5.27 vs Photon DBU $15.67 — 66% lower, on identical iron and identical runtime.

Why per-GiB instead of per-DBU or per-vCPU-hour?

Compute-time billing pays the vendor more when jobs run slowly. Every optimization we ship — better SIMD, smarter planning, native MERGE — would cut our own revenue if we billed by the hour. Per-GiB ties our income to the value delivered, not to inefficiency. The faster we make Spark, the better the deal you keep.

Does the rate change if I land on a different volume tier?

Yes — tier pricing is progressive. The GiB that lands in the 1–10 TiB band is billed at $0.0007; the next chunk in the 10–100 TiB band is billed at $0.0005; and so on. You don't fall off a cliff between tiers. Most single-run ETL workloads sit in the 1–10 TiB band; aggregate monthly volume often crosses into $0.0005 or $0.0004.

How does this compare to OSS Spark, Comet, or Gluten — all free?

Free engine, paid time. On TPC-DS 10 TB with the same 11-node cluster, OSS Spark takes 12,200s vs Quanton's 2,384s — 5.12× the EC2 hours. Comet and Gluten are 3.6–3.8× slower and OOM on q67/q93. The Quanton fee is smaller than the extra EC2 you'd burn waiting for OSS to finish — and that's before counting engineer time.

What if I want to walk away?

One config line points your orchestration back at OSS Spark, Comet, or Gluten. Same Scala / PySpark / SparkSQL jobs, no rewrites. Quanton ships as a Docker image and a Kubernetes operator that mirrors the Kubeflow SparkApplication spec. Your data never leaves your VPC; your jobs stay portable.

Is the AI Spark engineer extra?

No. Quanton AI lives in your Spark UI and is free — it works across Quanton, OSS Spark, Comet, and Gluten. Your Claude or OpenAI key lives in your browser; prompts and AI data never touch our servers.

Why GiB and not GB?

Spark UI reports read/write bytes in binary (1 GiB = 1,073,741,824 B). Billing on GB instead of GiB would silently change the fee by ~7%. We use the same unit Spark reports — no decoder ring required.

Bring a workload. We'll show you the diff.

Spin up Quanton in your own Kubernetes in under 10 minutes. Compare against your current Databricks bill in your own VPC.

Create an account → Talk to us