D-02 · Rewriting performance- and cost-critical systems in Rust
Orders of magnitude cheaper to run. Provably correct states.
Microsoft, AWS and the Linux kernel reached the same conclusion for the same reasons. Resource cost falls by orders of magnitude. State guarantees become provable rather than hoped for. The savings compound for every hour the system runs.
Orders-of-magnitude cheaper
Rust applications routinely consume a fraction of the memory and compute of their Python, Node or Java predecessors.
State guarantees
The type system makes impossible states impossible — a whole class of production bugs is prevented at compile time.
Full-stack Rust
Backend APIs, frontend via WASM, embedded devices — one language across the whole surface. (This site is Rust + Axum, by the way.)
Energy-aware by construction
Efficient runtimes aren't a marketing line; they're how we design from day one.
Auditable
Agentic systems move real money and make real decisions — we use the language that takes that seriously.
Questions
Is a Rust rewrite worth it?
It pays off for long-running, compute- or memory-heavy services and for correctness-critical state. Rust applications routinely use a fraction of the memory and compute of their Python, Node or Java predecessors, and the savings compound every hour the system runs.
Do you rewrite everything at once?
No. We migrate incrementally — strangler-fig style, one service at a time, measured before and after — so there's no big-bang cut-over.
Can Rust cover the whole stack?
Yes: backend APIs, frontends via WebAssembly and embedded devices. This website is Rust and Axum, with no Node.js in the build.