I. Traditional branding is dead.

Your company will live or die in rooms you aren't present in.

A truth as old as time. But these days, there are more rooms than ever, they fill up quickly, and no one quite has the time to figure out what you actually do.

On a Sunday night, VCs skim your deck with the weight of hundreds of unanswered messages. Senior hires weigh up your offer against two other companies also “building the future of X”. Regulators decide whether you’re serious or unserious by looking at your Twitter. Journalists set the frame that every publication adopts. LLMs summarise your company in three sentences, that is, if they don’t hallucinate and confuse you for somebody else. Years of work. Years of your life. Compressed into sixty seconds of skimming.

You’re not in any of these rooms. But your surfaces are. 

These days, decisions are made on what's visible. It is easy to be misread. You feel it, even when you're not in the room. The VC stops replying. The senior hire passes on your offer. The journalist dismisses your story. You learn about the decision after it’s been made, and you can’t name what went wrong. Because nothing went wrong in the room you were in. Something went wrong on a surface, somewhere else, while you were doing the work. 

The market either sees a project or a category-defining company. These two are not the same. They raise from different investors. They attract different talent. One gets a shot at a generational company. The other struggles to survive. 

This wasn't always the way. For a long time, your surface could be small. The technology spoke for itself to the few people who could read it, and the rest was left for trusted institutions to translate: journals, journalists, conferences, tradeshows. Brand was the cosmetic layer. A logo, a set of colours, a site people could click on. Brought in at the very end. Executed by people who didn't need to understand the technology.

That world is gone. The people deciding whether your company gets capital, talent, distribution, or permission now form their views the same way everyone else does. They look at whatever they can find, scan, and understand in under a minute. They make a judgment. They move on. Your surface can no longer be small. It can no longer be contradictory. It can no longer be outdated. There is no second layer. The real case is made on the surface.

Over five hundred founder conversations, the pattern is the same. The technology is real but the surface decides whether you live or die. The result is a category-wide failure mode: technology that is good for the world and hard to explain. And in the rooms you aren’t present in, hard to explain reads as not yet real.

To be illegible is to disappear. 

II. Legibility starts with the founder.

Only you can tell the truth about your company. No one else has the full picture. When you stay quiet, others fill in the surfaces. And they fill it with far less. 

Legibility is when capital, talent, and customers can read your company clearly. In deep tech or AI, legibility can't be delegated. The narrative is the technical case, made legible. Only the founder can hold the full architecture. From the biology to the physics to the ML. To why it matters, to what changes if it works, to why now. No one knows the architecture as well as you do. Legibility starts with the founder. With you.

It shows up at the sentence level. Your speech is a surface. The CFO who hears "optimised Kubernetes orchestration" hesitates to fund the project. The same CFO who hears "we cut our AWS bill by thirty percent" knows immediately. Both sentences are true. One works for the room. One doesn't.

The same happens with every surface. A line on the website that signals research rather than company. A slide that buries the unlock. A tweet that reads as hedging. A surface fails, in real time, even when you're not standing on it.

This is translation in action. 

Translation is finding the sentence that survives the room whole. It's not simplification. Done well, it makes the technical truth more itself, not less. The same insight, rebuilt in language the room can carry. It’s the architecture: What to say, what to leave out, what to reinforce. These choices can't be made by someone who only half understands the technology.

They’re made by you. This doesn't mean founders should write every sentence or design every surface. It means the architecture of legibility belongs to the founder, the same way the technical insight does. You can hand off the building. You can't hand off the blueprint.

If you don't engineer your brand, the market will engineer it for you.

III. Brand is infrastructure.

Marketing optimises for reach. Branding, as most of the industry practises it, optimises for aesthetic coherence. Neither is the problem deep tech or AI founders have. The real problem is that the technical truth of your company doesn't make it to the surfaces of the right rooms. And the gap between the truth and the perceived truth widens as the company scales. 

What closes the gap is engineering. Load-bearing, structural, precise work, applied to every surface where the company is read. The deck a VC opens in an airport lounge. The site a candidate sends to their partner. The data room a skeptic combs through on a weekend. The founder's posts. The first conversation. The second one. The sentence that’s repeated in IC. The line the journalist uses in the headline. Each is a component. Each has tolerances. Each can be designed for the load it carries.

This work has a name. Brand engineering. The discipline of making the technical truth carry across every surface it's being read, without losing its depth.

This is the work the next decade of deep tech and AI will be won and lost on. Narrative has become the primary input to what companies can build. The capital they access. The talent they attract. The permission they earn. The category they rule. The companies that treat it that way will win. The ones that don't will keep building extraordinary technology the world won't recognise. Cures that don't reach patients. Infrastructure that never makes it out of blueprint. Breakthroughs that die in the room where the wrong word was used.

By 2030, every deep tech and AI company will have a brand engineer the way they have a fractional CFO. The ones who figure it out early will already be three years ahead. The ones who don't won't just lose to competitors. They'll lose the work they came to do.

Make your technology feel inevitable.

Only the legible survive.