Skip to content

IN DEVELOPMENT · NO CHIP HAS BEEN TAPED OUT

What has not been decided

Three decisions about BramHA are open: the process and the foundry, the published specification, and how robot makers would eventually obtain the part.

· 3 min read

IN SHORT

Three things about BramHA are undecided or unannounced: the process and the foundry, what a published specification will contain, and how robot makers would eventually obtain the part. Until each is settled and stated, the site makes no claim about it. No chip exists and nothing can be ordered.

Some things about BramHA are not done. Others are not decided. The difference matters: an unfinished task has a known shape, while an open decision could go more than one way.

This article covers the open decisions. BramHA is in development, no chip has been taped out, and nothing is for sale. For the list of unfinished work, see what is not done yet.

The process and the foundry

The program plans to use a mature process node. Which one, and at which foundry, has not been announced.

This choice carries others with it. The foundry's process sets the manufacturing rules a layout has to obey. It also sets how much area and power the design costs, since mature nodes cost more of both per unit of compute than the leading edge.

Until the process is announced, none of those consequences can be stated. That is one reason the site has no figures.

The published specification

The first design's specification is not published. What a published specification would contain, and in what form, has not been decided.

The specification is one of the three gates a named engineer signs. A published specification is a further step: the version a robot maker's engineers could read and design against.

Until one exists, the description on the site is the whole of what is public: a small real-time controller with a root of trust, not a vision processor and not an AI accelerator.

How robot makers would eventually obtain it

BramHA is intended for robot makers: the vendors and integrators who build the machine. It is a part they build in, not a product an end facility buys. Silicon for the vendors who build the machine describes that audience.

How such a part would reach them has not been decided. There are no terms, no prices, no quantities, and no order process. There is also no partner intake, so there is nothing to join and nothing to sign.

How to treat an open decision

Treat it as open. It is not a hint about the likely answer, and the order of this list is not a ranking or a sequence.

The site is intended to state each decision when it is made, in the same plain form, and to leave it listed as open until then. It publishes no dates for any of them; the one date on the site is the target for first silicon, November 2027, and it is a target, not a promise. What in development means explains how to read statements of this kind.

A robot maker weighing the program today can evaluate the reasoning: that a robot's own silicon needs certainty, not density, and that a mature process node is enough for it. The three decisions above are what stand between that reasoning and a part that could be assessed. The page at /program carries the current list.

Questions this raises

When will these decisions be made?
The site publishes no dates for them. First silicon is targeted for November 2027, which is a target and not a promise, and it is the only date on the site. Each decision is intended to be stated when it is made.
Why is the foundry not named?
Because the process and the foundry have not been announced, and the site states only what is settled. Naming a process before it is fixed would be a claim the program could not stand behind.
How will robot makers get the chip?
That has not been decided. The silicon is intended as a part robot makers build into their machines, but there are no terms, no price, and no order process, and there is no partner intake.

KEEP READING