Skip to main content

Industries

We have watched bridges close.

An industry runs on data it cannot own, delivered through pipes it does not control. Then a protocol arrives. The window opens — two, maybe three years — where a team with the right stack can become the pipe. Then an incumbent wakes up, or a regulator moves, or a better-funded competitor deploys first. The window closes.

This matrix names 28 industries. Each one has a window in some state of open or closing.

You have a team. You have a stack. You have a finite window. The matrix below names the 28 plays. Three patterns survive contact with reality. DePINs, permissionless protocols, and tokenized assets are how the window stays open — until it doesn't. Data flows through every industry. The question is who controls the pipes. Healthy ecosystems depend on the answer.

The Five Layers Of Industry​

Good trade is built on five layers. Each layer rewards a different investment. From atoms to bits and back again — the loop that standards compound on top of.

  • Substrate — atoms, energy, physical laws. Everything else rents this.
  • Bits — data, schema, processing. The translation layer from physical to digital.
  • Pipes — rails, networks, supply. The connective tissue that moves value between layers.
  • Works — physical production and transformation. Where atoms get rearranged into outputs.
  • Stories — attention, belief, identity. The setpoint of desire that pulls demand through every layer below.

Each layer pulls hardest on a different one of the Inner Tight Five build buckets — Laws, Standards, Property, Machinery, IT. The industry you pick determines which buckets your team must fund.

The Matrix​

Every industry generates data. AI and robots consume it. The gap between what AI can do and what an industry has adopted IS the opportunity — but the window has a date.

The window for data-dense industries closes when the flywheel locks in: whoever deploys sensors first owns the training data, and whoever owns the training data owns the prediction. The window stays open in midstream-fragmented industries — where no incumbent has yet built the toll bridge between the prediction and the action.

Read the matrix for gaps, not scores. A 5/5/2 row (high data, high AI, low readiness) is a positioning window. A 5/5/5 row at Phase 6.0 is either a frontier or a door already shut. Mind the gaps — the empty cells are the opportunity.

Read the scores through the layers. A high-Robot + low-Ready row is a Works bet (your team must fund Machinery + Property). A high-Data + low-Ready row is a Bits or Pipes bet (Standards + IT). A high-AI + high-Ready row is Stories already — the position is gone unless you own the attention layer. Filter by layer first; the scores tell you which row inside the layer is ripe.

  • Data — Value of the industry's data for AI training (1 = generic, 5 = irreplaceable ground truth)
  • AI — How much AI transforms operations and value (1 = marginal, 5 = existential)
  • Robot — How much physical automation reshapes the industry (1 = minimal, 5 = dominant)
  • Phase — Evolution stage (3.0 = automation, 4.0 = smart systems, 5.0 = augmented workforce, 6.0 = autonomous ecosystems)
  • Ready — How prepared the industry is (1 = analog, 5 = native)

Filter by layer

Filter by category

28 of 28 industries

Three patterns:

  1. High data + high AI + low readiness = positioning window. Healthcare (5/5/2), energy (5/4/2), mobility (5/4/2). The gap between what AI can do and what the industry has adopted IS the opportunity. Watch for midstream toll bridges that make these windows artificially narrow.
  2. High robot + low readiness = physical frontier. Agriculture (5/1), mining (5/1), construction (4/1), manufacturing (5/2). Whoever deploys DePIN devices captures the data moat before midstream interference.
  3. High everything + frontier phase = convergence. Robotics (5/5/5). AI, data, and physical automation collide. Every mature industry was once frontier — telecom (1900), computing (1970), internet (1995), crypto (2015). Position at frontier before commoditization.

What every pattern shares: the team that owns the pipes owns the call. Data value, physical frontier, convergence — three different plays, one underlying fact. The pipe is the position.

Where to Place the Bet​

The matrix names 28 industries. A single team places one bet. Three filters pick it:

  • Stack-native fit — Does the work you already do match the work the industry needs? A team built on programmable settlement, verifiable provenance, and agent-to-agent commerce starts on third base in industries where trust and traceability ARE the product.
  • Open midstream — Is there a gatekeeper between prediction and customer? Healthcare has EHRs. Gaming has the 30% app store cut. The industries with no entrenched toll bridge let a new entrant become the bridge.
  • Readiness gap — Low readiness with high data + high AI scores is the positioning window. The incumbents have not deployed; the field is open.

Run the filter against the matrix. The example below is scored for a software-first team — programmable settlement, verifiable provenance, agent-to-agent commerce in the stack. If your primary build is hardware (Machinery-heavy), domain operations (Works-heavy), or regulated property (Property-heavy), reweight: a hardware-first team flips the Healthcare and Robotics rows from ruled-out to in-scope.

Stack-native + open midstream — Supply Chain (#11) + Technology (#28). Provenance, attestations, and agent-to-agent commerce are the product, not a feature. The midstream is unbuilt — agent procurement has no incumbent.

Already-arrived — AI Data (#15) + AI Compute (#16). Phase 5→6, readiness 5. Substrate every other industry rides on. Better as foundation than destination.

Moat-creating — Energy (#7) + Manufacturing (#12). Veracity 5, sensor-data flywheel. Whoever deploys DePIN devices first owns the data. Capex heavy.

Ruled out (software-first) — Healthcare (#1) + Robotics (#24). High scores, but EHR midstream toll (Healthcare) and hardware integration cost (Robotics) block a software-first team. A Machinery-heavy team reads these rows as in-scope.

The bet most teams should place: the industry where the work you already do is the work the customer needs, and the toll bridge has not yet been built.

The wrong bet is the industry where you can build the prediction but cannot deliver the action. A model that hits the EHR wall, the app store cut, or the regulated settlement layer cannot convert precision into revenue. Midstream matters more than the model. (Conviction: HIGH — directly supported by the Healthcare and Gaming examples above.)

Data Intensity​

Not all data is equal. Five dimensions determine how hard a data problem is — and where the cost of guessing wrong compounds fastest.

A tier-three veracity miss in healthcare is a regulator and a casket. A tier-one volume miss in advertising is a slow Tuesday. The same engineering mistake has different consequences depending on which industry you picked. Know which tier you are in before you build the pipeline. (Conviction: HIGH on the healthcare consequence; MEDIUM on the advertising asymmetry.)

  • Volume — How much data is generated (scale of storage and processing)
  • Velocity — How fast data changes (real-time vs batch tolerance)
  • Variety — How many types of data (structured, unstructured, sensor, behavioral)
  • Value — How much a correct prediction is worth (economic or physical consequence)
  • Veracity — How critical data accuracy is (regulatory, financial, or safety stakes)

Sort the 28 industries into three tiers. The tier — not the score — is the strategy.

Data IS the product — Advertising, Banking, AI Data, Telecom, Gaming. High volume and velocity, and the BI tool moat is already dissolving. AI natural language interfaces are replacing dashboard complexity. The competitive advantage shifts from tool expertise to schema quality and governance. Advertising is the original case — Google and Meta don't sell software, they sell targeting precision built on behavioral data. Banking, Telecom, and Gaming face the same shift now.

Data creates the moat — Healthcare, Manufacturing, Energy, Supply Chain. Proprietary sensor data from operations. The moat is physical: whoever deploys DePIN devices first owns the data flywheel. The data gap between incumbents and challengers widens with every sensor deployed.

Data determines trust — Real Estate, Mining, Space. Veracity is highest. You cannot throw an LLM at the query layer when a wrong answer has regulatory or physical consequences. Trust scoring and governance are the product, not the pipeline.

The industries where veracity is highest are where data engineering standards — repository quality, schema governance, trust scoring — compound most. See Data Analysis for how AI is attacking the BI tool layer across these tiers.

Evolution​

We sit at the Industry 4.0 → 5.0 seam — augmented workforces, agent collaboration, and tokenization are deployable today; Industry 6.0 (closed-loop AI + DePIN, network states) lands from 2027.

The distribution gap is the bet. Find the industry where Industry 5.0 tools exist but the field is running on Industry 3.0 norms — that is where the margin lives. The future is already here. It is not evenly distributed. Yours is to deliver it.

Full Industry 3.0 → 6.0 reference

Industry 3.0 (1970s–2000s) — Automation, computers, electronics, IT systems.

Industry 4.0 (2010s–present) — Smart systems, cyber-physical systems, IoT, networks.

Industry 5.0 (2020s–present) — Augmented workforce, agent collaboration, decentralized identity, tokenization.

Industry 6.0 (2027+) — Closed-loop AI and DePIN feedback systems, autonomous self-healing ecosystems, network states.

Value Chain​

Disruption maps to three layers of the digital supply chain:

  • Upstream (Moat): Can you defend the raw material? (Collection cost + data exclusivity)
  • Midstream (Scale): Is the pipeline open or monopolized? (AI advantage + pipeline dependency)
  • Downstream (Wedge): Can predictions trigger direct action? (Time to ACV + actuator potential)

The highest risk is the midstream toll bridge — EHRs in healthcare, the 30% app store cut in gaming. A prediction model loses all value if it cannot pass through legacy gatekeepers.

Value migration: Science discovers → Protocols standardize → Standards industrialize → margins compress → value moves to edges.

The loop: DePIN captures → Clean/Fast/Open data → AI learns → Better predictions → More value → Better devices.

Platform Stack​

A platform is the integral of investment in five build buckets — Laws, Standards, Property, Machinery, IT. Each bucket compounds independently. Each also degrades independently.

  • Laws — jurisdictions, licenses, protocol rules. The regulatory surface that determines what the platform is permitted to do at scale.
  • Standards — schemas, governance, interoperability contracts. The connective tissue that lets disparate machines speak the same language.
  • Property — land, spectrum, data, IP, permits. The resource positions that cannot be copied or commoditized quickly.
  • Machinery — physical infrastructure: rigs, devices, sensors, plants. The footprint that creates defensible data density.
  • IT — apps, agents, orchestration, contracts, pipelines. The coordination layer that converts data into decisions.

One weak bucket does not just slow the platform — it caps the ceiling of every other bucket above it. A software team that hits a property wall (no spectrum, no data rights, no permits) cannot build the model it designed. A machinery team that hits a standards wall produces data nobody else can read.

Build in the order that unblocks the next bucket. (Conviction: MEDIUM on sequence — order depends on industry; the unblocking principle is HIGH.)

Each bucket carries its own core metric, failure risk, and upside:

  • Laws — Metric: compliance cost as % revenue, time-to-approve. Risk: adverse regulation, classification risk. Upside: first-mover in clear regimes, trust advantage.
  • Standards — Metric: adoption rate, schema coverage. Risk: fragmentation, competing standards. Upside: default interoperability, compounding network.
  • Property — Metric: resource share controlled, yield, duration. Risk: expropriation, low utilization. Upside: compounding advantage, access pricing.
  • Machinery — Metric: utilization, uptime, unit economics. Risk: capex intensity, hardware obsolescence. Upside: footprint defensibility, economies of scale.
  • IT — Metric: SLOs, gross margin, automation %. Risk: technical debt, commoditization. Upside: high-margin coordination, default API.

The test: if one of these five failed badly, would you still invest? If yes, the platform thesis is not tight enough. See Tight Five for how this nests into the 5P framework.

The Convergence​

These industries aren't separate verticals — they're a convergence of data-centric systems that determine who navigates and who gets navigated.

Data owned

Connectivity

ControlLearn more →
Risk if lost: They control the signals between you and the world

Data owned

Movement

ControlLearn more →
Risk if lost: They know where you go and when

Data owned

Training data

BeliefLearn more →
Risk if lost: They train the brain that makes your predictions

Data owned

Transactions

ValueLearn more →
Risk if lost: They record what you value enough to pay for

Data owned

Financial records

ValueLearn more →
Risk if lost: They custody your stored value

Data owned

Attention + identity

BeliefLearn more →
Risk if lost: They shape what you see and believe

Data owned

Behavioral

BeliefLearn more →
Risk if lost: They design the systems you inhabit

Data owned

Sensor + actuator

ControlLearn more →
Risk if lost: They command the agents that act on your behalf

Own the data, own the navigation. Lose it in any one industry and the matching system fails — you stop steering and start being steered.

DePIN isn't a cheaper way to run infrastructure. It's who owns the call. Own the data, own the navigation. Lose it in any one industry and the matching system fails — you stop steering and start being steered.

Which one fails first if someone else owns the data?

Software Strategy​

Data sovereignty is not a compliance checkbox — it is a build decision that fixes your ceiling. The industries with the highest veracity requirements (Healthcare, Finance, Supply Chain) are also the ones where a wrong architecture choice locks in technical debt at the worst possible moment. See Buy or Build before committing the stack.

  • Healthcare — Very High data sensitivity (PII, PHI). SaaS: CRM, Scheduling. Crypto opportunity: secure EHRs, patient-owned records.
  • Real Estate — High data sensitivity (transactions). SaaS: CRM, Legal. Crypto opportunity: property tokenization, smart contracts.
  • Finance — Very High data sensitivity (regulated). SaaS: Analytics, BI. Crypto opportunity: DEXs, verifiable compliance.
  • Gaming — Medium data sensitivity (player data). SaaS: Community, Loyalty. Crypto opportunity: NFT assets, play-to-earn.
  • Supply Chain — High data sensitivity (provenance). SaaS: BPM, Analytics. Crypto opportunity: DePIN tracking, attestations.

See Vertical RaaS for the playbook and SaaS Toolkit for feature specs.

Context​

  • Data Engineering — How pipelines, schemas, and repositories are built for data-heavy industries
  • Data Analysis — How AI is dissolving the BI tool moat across industry tiers
  • AI Data Industry — The industry that sells data infrastructure to every other industry
  • Culture — Music, sport, fashion, food shape identity and adoption
  • Matrix Thinking — Cross verticals with forces to find gaps
  • Navigation System — Data sovereignty is navigation sovereignty
  • Tight Five Prompts — The 5P lens each industry follows
  • Business Development — The playbook for finding and closing deals
  • DePIN Devices — Where physical infrastructure meets token incentives
  • Data Flow — Clean, fast, open data principles

The Decision​

You have read the matrix. You have seen the three patterns. The bet is yours.

Which industry's data are you positioned to capture before the toll bridge closes?

Two questions to test the pick before you commit:

  • When the cost of intelligence goes to near zero, what stays scarce in your chosen industry — and do you already own it?
  • If you lose control of the baseline data in that vertical, which part of your navigation (Belief, Control, Value) degrades first?

Commit the bet​

Reflection is not a build. A bet becomes a build when you write it down with a name, a layer, a bucket, and a date.

  • The play: <one-sentence industry bet>
  • The layer: Substrate · Bits · Pipes · Works · Stories — <pick one>
  • The bucket that funds it this cycle: Laws · Standards · Property · Machinery · IT — <pick one>
  • The window closes by: <date — be specific>

Fill those four lines. Then open a workchart against them. That is how the bet becomes a build. (Conviction: HIGH — commitment devices outperform reflection at converting intent into action.)

Name the answer aloud. Open a workchart against it. That is how the bet becomes a build.