Why I built Hardcharger

I’ve spent 25+ years in data engineering, at P&G, HP, Dunnhumby, Kroger, and Voltron Data. Over that time, the cost of analytics has climbed faster than the value most teams get from it. Warehouse bills grow every quarter. Clusters take minutes to wake up. Pricing pages need a spreadsheet to decode.

Then DuckDB came along and changed what I thought was possible. One modern machine running DuckDB can handle queries that used to need a fleet of servers. For a large share of real-world workloads, it’s faster than the big platforms at a fraction of the cost.

But every time I tried to bring DuckDB into a business, I hit the same walls:

It’s an embedded, in-process database. That’s great on a laptop. It’s much harder when a whole team, a BI tool, and several apps all need the same data at once.
It has no built-in role-based access control. There are no users and no roles, so you can’t say “finance can see this table but marketing can’t.” For any company that cares about security or compliance, that’s a dealbreaker.
Running it as a shared, always-on service means building a lot of plumbing yourself.

I know how much RBAC matters because I’ve lived it. As a DBA (I was named Oracle DBA of the Year for North America in 2016), access control was the job. Businesses don’t adopt a database just because it’s fast. They adopt it because it’s fast and they can trust it with their data.

So I built Hardcharger.

Hardcharger takes the raw speed of DuckDB and adds what businesses actually need:

Client-server architecture. Analysts, dashboards, notebooks, and applications across your organization all share one high-performance engine.
Built-in RBAC. You control who can see what and who can do what.
Chargewire. Our new open wire protocol, built from the ground up to move query results to clients as fast as possible.
Open source under Apache 2.0. You’re never locked in.
A fully managed option on GizmoData Cloud. You can skip the ops entirely and just run queries.

The goal is simple: warehouse-class performance and governance without warehouse-class bills. If you’re paying Snowflake or Databricks prices for workloads that don’t need that scale, Hardcharger should be a relief. You get fast queries, transparent pricing, and no surprise invoices.

Why the name? I’m a former U.S. Marine. “Hard charger” is what Marines call someone who attacks every task with relentless energy and doesn’t quit until it’s done. That’s what I wanted this database to be: fast, tough, and dependable.

GizmoData is bootstrapped. There’s no VC money and no growth-at-all-costs pressure. That means I get to build the product I’d want to buy, with honest pricing, real support from the person who wrote the code, and a roadmap shaped by customers instead of investors.

I’d love for you to give Hardcharger a try, and I’d really value your feedback. Tell me what works, what doesn’t, and what you’d need to move your workloads over.

— Phil