Rethinking Streaming Costs: How we simplified pricing for Condense

Written by
Sudeep Nayak
.
Co-Founder & COO
Published on
Jun 30, 2025
Product

Share this Article

Share this Article

In the world of data streaming, pricing is often where clarity goes to die. 

Some platforms charge per connector. Others per task. Some even break it down per schema, per byte, or per million events. And if that wasn’t enough, it’s often bundled with infrastructure, billed in mysterious “data units,” and peppered with unpredictable thresholds and premium gates. What starts as an experiment in real-time often ends in a budgeting nightmare. 

We studied over a dozen commercial Kafka-based platforms and managed streaming services, and here’s what stood out: 

The industry has overcomplicated what should have been simple: compute equals cost. 

At Condense, we’ve done things differently. 

One Simple Metric: vCPU-Hours 

Condense pricing is grounded in reality: what you run is what you pay for. Every connector, every transform, every enrichment, every alert pipeline, under the hood is powered by compute. 

So, we decided to make that the billing unit: vCPU-hours 

No markup for “premium connectors.” No surcharge for “custom schemas.” No asterisk beside “advanced transforms.” If it runs, it consumes compute. If it’s idle, you don’t pay. 

Real-time should scale with actual workload, not with line items in a billing sheet. 

The Core Model 

Condense offers three simple tiers, each with included vCPU-hours to make planning predictable. Here’s the breakdown: 

Standard Tier – Built for Growing Teams 

 

Rate beyond included vCPU-hours 

Line Item 

Base Price 

Included vCPU-hours 

PAYG Rate 

1-Year Rate 

3-Year Rate 

Standard Tier 

$800 

1,440 vCPU-hours 

$0.40/hr 

$0.36/hr 

$0.32/hr 

What this means

You get 1,440 vCPU-hours upfront with a flat $800 license. That’s roughly 48 hours of parallel execution at 30 vCPUs, plenty to run full streaming workloads across multiple pipelines and connectors. 

After that, you move to pay-as-you-go, with discounts for longer commitments. 

Evaluation Tier – For Testing, Prototyping, and Pilots 

 

Rate beyond included vCPU-hours 

Line Item 

Base Price 

Included vCPU-hours 

PAYG Rate 

Eval Tier 

$0 

5,760 vCPU-hours 

$0.30/hr 

No tricks. No credit card. Just a full-featured cluster with compute to explore, test, and build. 

Enterprise Tier – For Custom Workloads, Higher Throughput 

 

Rate beyond included vCPU-hours 

Line Item 

Base Price 

Included vCPU-hours 

PAYG Rate 

1-Year Rate 

3-Year Rate 

Enterprise 

Contact us 

1,440 vCPU-hours 

$0.50/hr 

$0.45/hr 

$0.40/hr 

Tailored for high-security environments, edge-to-core integrations, or use cases exceeding 100MBps throughput. 

What’s Included? 

Everything. There are no limits per connector, per schema, or per job. 

Condense pricing does not penalize scale, flexibility, or creativity. You want 10 different Kafka topics feeding 20 different business logic paths? Go ahead. You’ll only be billed for the vCPU time your cluster uses. 

This contrasts sharply with other platforms we evaluated: 

Platform Feature 

Others Charge Separately 

Condense Model 

Per connector instance 

✅ Yes 

❌ Included 

Per running task 

✅ Yes 

❌ Included 

Schema registry tiering 

✅ Often 

❌ No charge 

Managed Kafka per GB 

✅ Yes 

❌ BYOC, your cost 

Autoscaling surcharge 

✅ Hidden costs 

❌ Simple compute 

Custom logic / enrichment 

✅ Often extra 

❌ Run freely 

When teams move to Condense, they don’t just get cost transparency, they get permission to build without fear of being billed for every experiment. 

Let’s Talk Infrastructure

Because Condense supports Bring Your Own Cloud (BYOC) deployments, you control where it runs GCP, AWS, Azure, under your own billing and security policies. 

Here’s what a base deployment looks like on GCP (Mumbai) for ~1MBps sustained streaming: 

Component 

Qty 

Unit Rate (Monthly) 

Subtotal 

GKE Nodes (n2d-highmem-4) 

$121.84 

$609.20 

Kubernetes Engine control plane 

$73.00 

$73.00 

Artifact Registry 

$10.00 

$10.00 

Persistent Disks (1TB PVCs) 

$49.15 

$147.45 

Compute Node (n2d-standard-2) 

$41.91 

$41.91 

Static IPs 

$8.75 

$17.50 

Estimated Total Infra (Monthly) 

~$899 

This estimate scales linearly with your throughput. No surprises. No bundling. No cloud lock-ins. 

Predictable Scaling, No Surprises

As you scale throughput or increase the number of pipelines: 

  • Infra costs scale based on actual resources (under your cloud). 

  • Compute costs scale linearly through vCPU-hours. 

  • Billing visibility remains clear across teams and environments. 

There’s no rate card to decode. No “pricing FAQ” to interpret. Just one thing to remember: 

The more your pipeline computes, the more vCPU-hours you use. That’s it. 

Final Thoughts: Simple Pricing That Just Works 

Condense pricing is made for engineers. It’s clear, predictable, and doesn’t get in the way of building. 

You don’t need to talk to sales just to run a new pipeline. There are no extra charges for adding connectors, writing custom logic, or scaling to more use cases. 

Whether you're processing event streams, enriching data in real time, or powering alert systems, the pricing stays the same: based on how much compute you actually use. 

While other platforms bury costs in layers of tasks, features, and usage thresholds, Condense keeps it simple, transparent, and built for real workloads. 

Ready to Start? 

  • Start with Evaluation tier and 5,760 free vCPU-hours. 

  • Scale via Standard and pay only as you grow. 

  • Deploy securely in your cloud, on your terms. 

  • Talk to us for Enterprise scale and support. 

Book a meeting with us to discuss you custom use case here

On this page

Get exclusive blogs, articles and videos on Data Streaming, Use Cases and more delivered right in your inbox.

Ready to Switch to Condense and Simplify Real-Time Data Streaming? Get Started Now!

Switch to Condense for a fully managed, Kafka-native platform with built-in connectors, observability, and BYOC support. Simplify real-time streaming, cut costs, and deploy applications faster.