We design the chips AI runs on, build the tools used to design them, and ship applied AI. So we see the whole stack.
Most people who analyze AI see a single slice of it: the market, the models, or the cloud bill. We have worked the entire stack, from individual transistors and the energy they burn, up through the accelerators, memory, and datacenters that run a model, to the applications that turn it into a product people pay for.
That perspective is rare, and it is the whole reason WaferZero exists. We understand what a token truly costs before anyone sets a price on it, we can see which workloads are wasteful at the level of the hardware itself, and we can follow one line of reasoning from the silicon all the way to the business case. When you need to know what is actually happening in AI, we can answer it from the metal up.

Usman Zia
Co-founder
Previously Senior Engineer at AMD, designing silicon for AI accelerators.

Gurinder Garcha
Co-founder
Previously Staff Engineer at AMD, designing silicon for AI accelerators.
Worked at every layer we write about.
AI accelerator design
Prior experience designing the chips that train and run frontier models.
AI for chip design
Building the ML and agentic tooling used to design those chips.
AI control-plane prototype
A working prototype that meters token spend, governs agents, and maps cost to output.
Tribunus Labs
Applied AI shipped into a regulated, document-heavy real-estate workflow.
The clearest sense of how we think is the work itself.