Pay for what your queries actually use.
Query Engine: pay for consumption, for work, or for a guaranteed outcome - you choose. Answer Engine takes that same plan and adds agentic token cost on top.
Two ways to pay for Query Engine. Pick the one that matches how predictable you need cost to be.
Pay for Compute
$0.175 / vCPU-hour
Metered on the compute you actually consume, with scaling in increments as small as 1 vCPU-hr. Available on pay-as-you-go with no minimum commitment. No step-jump warehouse sizing with wasted capacity.
- Scales based on load
- Easiest way to get started
Outcome-Based Pricing
Custom · contact us
Shifting from a traditional warehouse or lakehouse platform? Lock in a 30-50% savings for 1 to 3 years. We work backwards from your current usage and offer outcome-based pricing that guarantees the savings amount.
- Guaranteed % reduction, secured for mid- to long-term contracts (6 months to 3 years)
- Best for teams averse to bill unpredictability, price shocks, and complex credit-based pricing models
Need pricing structured a different way - fixed cost per user instead of per query, for example?
Contact us →Send a query to Query Engine and it returns a set of results - you're charged for that execution. Send a question to Answer Engine and it does whatever it needs to get you an answer - you're charged for the tokens behind the agentic operations, plus the query (or queries) it runs along the way.
Agentic Token Cost
Custom · scoped to usage
Billed on top of whichever Query Engine plan you choose, metered on tokens consumed by the agent's reasoning, retries, and orchestration steps.
Talk to us about Answer EngineChoose your Query Engine plan on the first tab, then talk to us about layering Answer Engine on top - pricing is scoped to your expected agent volume.
COST QUESTIONS
Answered plainly
How many vCPUs will I need?
Get started with the minimum, and auto-rightsizing will figure it out to solve for two things - no waste, and making sure your queries run as fast as possible. Or, if you know your sizing on Databricks, Fabric, Snowflake, Redshift, or BigQuery, our AI concierge will understand your workload and goals (cost savings, latency, etc.) and get you auto-configured and ready to go. If you'd like us to run a pilot, POC, or bake-off, feel free to reach out and our team will get in touch.
What does this include?
Costs towards all the physical infrastructure the engine runs on - e.g. EC2 on AWS or VMs on Azure, NVMe disk caches, networking, Kubernetes, and so on - plus license, bundled into a single all-inclusive price. If you're running BYOC (bring your own cloud), contact us for pricing: you pay your cloud provider directly for compute, and pay e6data only for the license - similar to a bring-your-own-compute tier on platforms like Databricks Classic/Pro Compute.
Do you have enterprise pricing? What about discounts for pre-committed spend?
Yes. Reach out and we'll put together something that fits your needs based on your expected usage volume, contract length, and payment terms.
Do you support pay-as-you-go and bite-sized prepaid packs? Is there a minimum commitment?
Yes - you can start here with no minimum commitment, and switch to a plan with pre-committed spend later if you'd like a discount on the listed rate card.
What's the difference between Query Engine and Answer Engine pricing?
Send a query - SQL, Python, or otherwise - to Query Engine and it gives you a set of results; you're charged for that query's execution. Send a question to Answer Engine and it does whatever it needs to get you an answer; you're charged for the cost of tokens across the various agentic operations, plus the cost of the query (or queries) executed along the way. In short: Answer Engine takes whichever Query Engine plan you pick and adds a metered charge for agentic token usage - the reasoning, retries, and orchestration steps an agent runs on top of the query layer.
Book a demo on your own workloads
Reach out to book a demo, share challenges you're facing, and tell us how this fits into what you're currently working on or thinking about.
Prefer to self-serve? Problems we're solving →