Dark Matter Industries

The Shape of What You Cannot See

Late in the last century, cosmologists noticed something embarrassing. The galaxies were spinning too fast. Take every star, every gas cloud, every speck of luminous material you can actually see, do the math on the gravity it should produce, and the outer arms of a spiral should fling themselves into the void. They don't. Something holds them. Something with enormous gravitational presence is curled around every spiral, dragging on every orbit, bending light around clusters that — by all visible accounting — are far too light to bend it.

We call it dark matter. We have never seen it. We know it is there from the shape of what it pulls.

Industries are no different. There is a luminous economy — the part that emits light, fills the conferences, draws the tutorials and the YC batches and the LinkedIn carousels and the venture decks. And there is a gravitational economy — the part that moves the money, employs the people, processes the transactions, runs the substrate of an ordinary day. The two overlap far less than anyone admits. The galaxy of developer attention spins around the luminous half. The dark half does the work.

That gap is the most reliable map there is. Wherever economic mass and developer mindshare diverge — violently, measurably — a whole category of work is sitting in the open, unbuilt. Not because nobody noticed. Because the people best equipped to build it are looking somewhere brighter.

---

The Power Law of Mindshare

Developer mindshare is not a bell curve. It is a power law, and a steep one — near vertical — where a handful of categories swallow nearly all the visible light. Fintech. Developer tools. Infrastructure. Productivity SaaS. And, these last three years, anything that so much as brushes a foundation model. The names recite themselves: Vercel, Linear, OpenAI, Anthropic, Datadog, Supabase, Cloudflare. SDKs pulled down millions of times a week. Documentation with a blast radius wider than the GDP of a small nation.

This is not a complaint. Those companies are good. The tooling is good — in places it is the finest human-computer interface we have ever built for the work of making software. They have earned their pull.

The problem is what the pull distorts. A generation of engineers grows up inside the gravity of four or five categories and learns — without ever quite deciding to — to want only the problems those categories already solved. Problems that deploy in seconds. Problems that wrap cleanly around a model call. Problems that flatten into a primitive tidy enough to fit on a landing page. Gravity bends every trajectory that passes near it, and so a generation ends up in orbit around one small bright cluster, mistaking the orbit for the whole sky — while the rest of the economy runs on where nobody is looking, on bespoke integrations and FTP transfers and faxes and the heroic exhaustion of in-house IT teams nobody will ever write an essay about.

Set the two inventories side by side and the asymmetry turns almost comic. The light pools around payments and code editors and CRMs, vector databases and auth providers and CI runners — the tools developers build for other developers. It pools nowhere near exhibition. Agriculture. Construction. Logistics. Hospitality. Public transit. Mineral extraction. Port operations. Healthcare back-office. Any one of them moves more money in a year than the entire vector-database category will see in a decade. Any one of them employs more people than the whole fintech-tooling layer combined. Enormous galactic mass. Almost no developer signal at all.

The gap between gravitational pull and visible light is not an inefficiency to lament. It is a map.

---

Why the Dark Stays Dark

A few properties keep these industries invisible, and they are worth naming, because the same properties are what make the opportunity inside them so durable.

Start with the unglamorous part, in the precise way developer culture means the word. There is no leaderboard for the best inventory system for a regional grocery chain. No Hacker News thread climbing to a thousand points on freight-rate optimisation. The categories that hold the light are the ones where the work is legible to other developers — a clean API, a satisfying primitive, an abstraction you can admire. The work that runs the dark economy is legible mostly to the people doing it, and those people are not, by and large, writing blog posts about Rust.

Deeper than glamour is specificity. These domains are irreducibly specific. Horizontal SaaS — the dominant playbook of the last fifteen years — works by stripping the domain away until a thin universal primitive is left: a payment, a message, a document. The big dark industries will not strip that thin. Seat-level inventory at a multiplex on a Friday night — overlapping showtimes across nine auditoriums, three premium formats, six loyalty tiers, two food-and-beverage models, a concurrent IMAX release whose distribution deal dictates minimum run length down to the day — does not flatten into a generic primitive. It has the texture of the actual business. Generic tooling slides off it like water off glass.

Then there is the seal on top. The dark industries are owned and run by people who do not speak at developer conferences. They speak at exhibition trade shows, freight expos, hospitality forums, agritech field days held inside actual fields. What these industries need does not travel through the channels developers watch. It travels through channels developers have never heard of, in the language of operators who would frankly rather not learn about developers either.

And underneath all of it, the property that matters most: the gap cannot be closed from outside. A founder who has never sat through a Friday-night peak, never watched a session-time conflict ripple across an inventory system in real time, never seen a concession queue spill into the lobby because one terminal died during the trailers — that founder builds the clean abstraction and misses the part that mattered. The category is protected by the cost of understanding.

That is why the dark stays dark. The visible industries reward generalists. The dark ones reward people who have lived inside them. And lived experience does not scale the way Y Combinator scales.

---

Cinema, Measured

Cinema is one of the cleanest dark matter industries a contrarian could ask for. And the numbers are unkind to anyone inclined to wave it away.

The largest cinema technology platform on earth runs more than NZ$3 billion in annualised gross transaction value through its cloud alone. Every ticket. Every concession. Every loyalty point earned and burned. One substrate of software, across cinemas in more than a hundred countries, handling real-time seat concurrency at a scale most ecommerce platforms have never had to imagine. That single figure comfortably exceeds the lifetime processed volume of many celebrated developer platforms.

Now go looking for the developer ecosystem. Search npm for cinema SDKs. Search GitHub for typed clients against the exhibition stack. Search the conference circuit for a talk on cinema platform engineering. Search for the component libraries, the test-mode sandboxes, the language clients, the community type definitions, the cookbook integration guides any fintech or messaging developer treats as the baseline air they breathe.

There is almost nothing.

Here is where the careful reader has to be careful, because cinema throws two very different shadows, and a fast objection waits in the gap between them. There is cinema-as-content — what a film is, who made it, what it is about, what it is rated, what it grossed — and that half is luminous. More than a million developers build against open movie-metadata platforms; the community libraries pile up in every language anyone has bothered to write, from AS3 to a C99 wrapper nobody asked for. Point at the dark and someone will point back at that and call the thesis wrong. But all that light pools on the discovery side — the business of describing the film. Cross to cinema-as-operating-business — selling the seat, running the screen, pricing the show, moving the money through the building — and the developers vanish. Same cultural object. Two opposite densities of attention. The dark matter was never the movie. It was the business of showing it.

The platform exists. The APIs exist. They serve the operator and a thin population of enterprise integrators well enough. But access stays gated and consultancy-mediated — talk to your local representative, integrate through a specialist — and so the ambient layer, the second-order weave of community tooling that turns a capable API into a productive category, never formed. There is no public, self-serve sandbox a curious developer can treat as air. Tens of billions of dollars move through cinema exhibition technology worldwide every year, and the developer mindshare attached to that operational mass is, conservatively, something on the order of three orders of magnitude smaller than what attaches to a single well-funded vector-database company. Take it as an order-of-magnitude intuition, not a measured statistic — but the shape of the gap is not in dispute.

Real gravity. No light. A category-defining piece of work, waiting on whoever is willing to build the layer that lets the rest of the galaxy see it at all.

---

How to See the Dark

There is a tired venture exercise — which giant industry has the lowest software penetration? Right instinct, wrong question. The honest version: where is the gap between economic mass and developer mindshare widest, and what does it take to see it before anyone else?

A working definition will do. A dark matter industry is one where the transaction volume measured at the platform layer runs into tens of billions of dollars a year, and where the developer community attached to that layer would fit, comfortably, in one mid-sized conference room.

By that test the list runs long and uncomfortable. Cinema exhibition. Commercial agriculture. Building automation. Port and freight logistics. Hospitality property management. Construction project management. Public transit. Mineral extraction and commodities. Healthcare back-office. Funeral services. Every one runs on software. Every one moves enormous volumes of money. Every one carries a developer footprint thin enough that a single builder with real domain proximity could become, inside five years, the default name the whole category is built around.

The pattern holds across all of them. Old enough to have been digitised, not loud enough to have been re-platformed. Domain-deep enough to shrug off generic abstraction. Run by people who were never in the room where the conferences happen. And protected — almost perfectly — by the cost of understanding, which quietly hands the advantage to whoever already paid it by accident.

The companies that define the next decade will not, on the whole, come from the visible categories. The visible categories are saturated. Every founder sprinting toward them runs against thousands of equally equipped peers running the same play. The next decade comes out of the dark — out of the industries with measurable gravity and no light — and the founders who build there will be the ones who can see the dark because they have lived inside it.

That is the contrarian thesis hiding in the open: the largest opportunities of the next decade are not where the attention is. They are where the mass is.

---

Coda

You will not find dark matter through market research. Market research surveys the visible spectrum. It tells you what people already talk about — which is, by definition, the luminous half. The dark half never shows up in the survey, because the people in it are too busy running the business to answer it.

Dark matter is found by proximity. By accident. By the summer spent working a multiplex, never forgetting the geometry of a Friday rush. By the family that ran a regional freight company. By the founder who grew up in a hospital, watching the back-office software fail in slow motion for twenty years. By the engineer who took an unglamorous first job because it was the only one going, and stayed long enough to understand what was actually broken.

The advantage is not intelligence. Not capital. Not technical skill in the abstract. It is the irreversible imprint of having been close to something, for long enough, that you can feel its gravity without instruments. That kind of seeing cannot be bought in a quarter. It is recognised, late, as the strange gift of having spent your time somewhere most people would never choose to spend theirs.

The galaxies that spin too fast were always spinning too fast. The mass was always there. The only question was who would be holding the right instrument when the universe finally became obvious.