Sigraphs Intelligence - Sample Dispatch - Confidential Deterministic signal report - Tesla - June 2026
Sigraphs
Integrated Signal Report

Tesla

The platform is real. The plan is not yet routed.
The verdict, before the detail

The platform is real. The plan is not yet routed.

Tesla has an exceptional platform, a genuine technology moat, and a $120B revenue goal that is currently built on one path , and that path goes through a regulator.

78
Decision readiness, out of 100
Standard
94
Strongest: Technology & Integration
58
Watch: Regulatory Exposure
The six-domain scorecard

What the engine sees, tested domain by domain.

Full-stack advantage. Under-priced.

Tesla owns more of its own stack than any competitor: the drivetrain, the battery chemistry, the software, the charging network, and the retail channel. OTA updates mean every car in the fleet improves overnight without a service visit. The problem is not the capability , it is that none of this appears in the pricing logic. The premium charged for a Tesla does not reflect the compounding value of a car that gets better after purchase. That gap is a missed revenue lever, not a brand decision.

Un-buyable moat. Un-instrumented.

Tesla has logged more real-world miles of autonomous driving data than any other company on earth. No competitor can buy this, license it, or approximate it at scale , it only exists because millions of owners have been driving production vehicles with sensors running for years. The moat is real and it compounds with every mile. The failure: it does not appear in the financial model, the investor deck, or the pricing rationale as a tracked metric. An asset that is not measured cannot be defended, priced against, or capital-allocated toward. It is currently a story, not a number.

Brand strong. Value gap closing.

Tesla's brand is still the dominant reference point in the EV category , the name most people say when asked to name an electric car. That is a real and durable advantage. The problem is directional: BYD has reached near-price-parity at scale in multiple markets, and Chinese entrants are closing the feature gap in ways that matter to buyers outside North America. Tesla's premium has historically been justified by a genuine advantage delta. That delta is narrowing, and no standing mechanism exists to track it, defend it, or communicate it. A brand that leads but does not measure its lead is a brand on borrowed time.

Loyalty real. Churn fuse lit.

Approximately 60% of US Tesla owners repurchase within the brand , a loyalty rate that outperforms most automotive brands. That base represents a significant annuity of services revenue, FSD subscription attach, insurance, and upgrades that has not been fully activated. The same base is also generating the earliest signals of churn: service wait times, build quality regressions, and FSD performance inconsistencies are appearing in review data and support queues. The annuity and the churn risk are the same asset. One is being ignored on offense, the other on defense, and there is currently no unified model watching both.

Service friction drags recovery.

Tesla's most consistent customer complaint is not about the car , it is about what happens when the car needs service. Wait times, limited center capacity, and poor communication during the service process are documented at scale in reviews, forums, and social data. This matters beyond customer satisfaction because it sits directly upstream of the loyalty number. An owner who has a bad service experience is an owner whose likelihood to repurchase, to attach a subscription, and to refer a friend all decline simultaneously. Service is currently funded as an operational cost. It belongs on the loyalty model as a retention system.

One approval gates the number.

Tesla's $120B revenue target is built substantially on the commercialization of Full Self-Driving and Robotaxi products. Both require regulatory approval , not from one body, but from regulators across multiple jurisdictions, each with its own timeline, political environment, and safety standard. These are not dates Tesla controls. The plan currently models this as a single event on a calendar rather than a probability distribution with multiple outcomes. If approval slips by 12 months in the two largest markets, no contingency plan currently exists that gets Tesla to the stated number through other means. The risk is not that FSD fails , it is that the plan has no alternative.

The findings

What the signals surface, and who owns each.

Every finding routes to three readers: the one who decides, the one who sequences, and the one who builds. That routing is the accountability layer.

Finding 01

The plan has one scenario. It needs two.

Strategic success probability , fails when leadership intent is mistaken for execution evidence.

Growth that depends on a regulatory event you cannot control is not a plan , it is an optimistic scenario.

Who owns this finding
Leadership
Owns the decision
Needs to know
  • Your growth number and your biggest stated risk are the same bet.
Owns

Split the goal into a base plan and an upside plan. Assign the pricing decision an owner, a model, and a date.

Manager
Owns the sequence
Needs to know
  • A fused plan cannot be tracked.
Owns

Two separate revenue models by end of month.

Technical IC
Owns the build
Needs to know
  • Eight quarters of guidance-vs-actual is the primary measure.
Owns

Decompose $120B into two P&L lines. Build price-elasticity model at 3-5 price points.

Done when

Done when two plans exist, tracked independently, pricing model decision-ready

Finding 02

The loyalty base is an annuity and a churn risk at the same time.

Untapped value + customer loss , offense and defense must share a dashboard.

The 60% loyal US customer base is generating repeat revenue but also generating complaint signals that are the first inputs to customer loss.

Who owns this finding
Leadership
Owns the decision
Needs to know
  • The most defensible near-term revenue is not the next car sold.
Owns

Shift resources toward installed-base activation as the primary revenue bridge.

Manager
Owns the sequence
Needs to know
  • Running offense and defense separately produces contradictory metrics.
Owns

Two tracks, one weekly review. Track A: monetization ladder. Track B: churn early-warning feed.

Technical IC
Owns the build
Needs to know
  • The churn model escalates at defined probability bands months before loss.
Owns

Health score per owner cohort. Wire escalation to defined thresholds.

Done when

Done when every at-risk cohort scored, routed, actioned before renewal opens

Finding 03

Service friction is upstream of the loyalty the recovery depends on protecting.

Service visibility , documented failure mode: silence during delivery.

Tesla's most common complaint , service wait times , is the first step in a chain running from service experience to trust to repeat revenue.

Who owns this finding
Leadership
Owns the decision
Needs to know
  • Service wait times appear on the P&L as an operational cost. They belong on the loyalty model as a churn predictor.
Owns

The reframe. Service capacity is a retention system.

Manager
Owns the sequence
Needs to know
  • Most complaints are about not knowing , not the wait itself.
Owns

Convert service KPIs from cost-based to retention-linked.

Technical IC
Owns the build
Needs to know
  • An owner in a service flow should always be able to see status, next step, and ETA.
Owns

Milestone notifications per service stage. SMS/email triggered by state change.

Done when

Done when any owner can self-serve status at any moment without calling anyone

Finding 04

The premium is charged against a value gap that is not being tracked or defended.

Competitive displacement , when competitors reach feature parity, switching cost becomes the primary retention lever.

BYD is closing the price gap at scale. Chinese entrants are closing the feature gap. The premium is supported by brand momentum rather than a tracked delta.

Who owns this finding
Leadership
Owns the decision
Needs to know
  • You are charging a premium against an advantage you are not measuring.
Owns

Make the competitive value gap a tracked, standing number.

Manager
Owns the sequence
Needs to know
  • A standing index reviewed monthly is the output.
Owns

Competitive value-gap index across price, features, trust, and switching cost per major rival.

Technical IC
Owns the build
Needs to know
  • When parity is reached, the retention lever shifts to switching cost , ecosystem lock, Supercharger dependency.
Owns

Build the value-gap index. Update monthly.

Done when

Done when the premium maps to a measured delta, not a brand narrative

Finding 05

The largest risk in the plan is modeled as a date.

Policy-to-business exposure , a regulatory item is a plan input only when source, entity, exposure, and confidence are all present.

A multi-jurisdiction regulatory event is not a date , it is a set of outcomes with different likelihoods. The contingency does not yet exist for the case where approval does not come on schedule.

Who owns this finding
Leadership
Owns the decision
Needs to know
  • The question the board will ask: what revenue looks like if approval slips. That answer is not in the plan.
Owns

A board-level answer: revenue if unsupervised autonomy slips 12-18 months.

Manager
Owns the sequence
Needs to know
  • A contingency built after a denial is a crisis response, not a strategy.
Owns

Regulatory dependency register: each jurisdiction, approval type, status, probability band, and revenue exposure.

Technical IC
Owns the build
Needs to know
  • A regulatory item enters the revenue model only when source, entity, exposure, and confidence are all present.
Owns

Build the register. High-confidence items feed the revenue model.

Done when

Done when no revenue line depends on an unmodeled, undated regulatory approval

Finding 06

The most defensible asset in the business is a narrative, not a number.

Lifetime value allocation , when the loop is narrated but not measured, the asset cannot be priced or defended.

The real-world driving data loop cannot be purchased or approximated. Every mile generates behavioral data that feeds model improvement. It does not appear in the financial model as a metric.

Who owns this finding
Leadership
Owns the decision
Needs to know
  • You have an asset no competitor can buy, and it is not in your numbers.
Owns

The directive to convert the data loop from a narrative asset to a reported metric.

Manager
Owns the sequence
Needs to know
  • Data, Finance, and Product cannot produce this metric alone.
Owns

Cross-functional effort: three stakeholders, eight weeks, one recurring metric tracking data loop contribution.

Technical IC
Owns the build
Needs to know
  • The chain: miles to edge-case coverage rate to model performance delta to FSD attach rate to LTV uplift.
Owns

Instrument each link. Build the metric. Get it into investor materials.

Done when

Done when the data loop has a number, a trend, and an owner

The ladder out

The exposures, ordered as a climb.

Each rung is a step that unlocks the one above it. This is the sequence, not a menu.

1

Split the $120B target into two separate P&L lines: controllable revenue and autonomy-dependent revenue effort low impact high

Separate reporting into revenue the company controls today and revenue contingent on autonomy milestones, so each can be planned and judged on its own. Define the classification rules before splitting to avoid moving goalposts later. Use it internally first before any external disclosure.
unlocks Expect clearer internal accountability and planning rather than a market reaction, since this is a management lens first. Track forecast accuracy per line as the leading indicator.
References: Mosaic, Pigment, Workday
2

Build a monetization ladder for the 60% loyal installed base: insurance, FSD subscription, upgrades effort low impact high

Map the loyal base and design a staged set of recurring offers across insurance, software subscription, and upgrades, sequenced by adoption ease. Start with the offer that has the lowest friction for existing owners. Measure attach and retention per rung before adding the next.
unlocks Expect a growing recurring-revenue stream from existing owners that compounds slowly, not a step change. Track per-owner recurring attach as the leading indicator.
References: Stripe, Recurly, Chargebee
3

Instrument the data loop as a tracked board metric: miles to model improvement to subscription attach to LTV effort high impact high

Define and instrument the chain from fleet miles to model improvement to subscription attach to lifetime value, then track it at board level. Start by wiring the two links with the cleanest data before completing the full chain. Treat gaps in the data as the first thing to fix.
unlocks Expect better visibility into whether the data advantage actually converts to value, surfacing over a few quarters. Track end-to-end loop instrumentation coverage as the leading indicator.
References: Snowflake, Databricks, Amplitude
4

Create a regulatory dependency register mapping each jurisdiction, approval type, probability band, and revenue exposure effort low impact high

Build a living register that ties each autonomy-dependent revenue stream to its jurisdiction, approval type, probability, and exposure. Populate the highest-exposure jurisdictions first. Review it on a fixed cadence with legal and finance.
unlocks Expect earlier warning on regulatory risk to revenue rather than a direct financial gain. Track exposure-weighted approval probability as the leading indicator.
References: Notion, Airtable, LogicGate
5

Wire service complaint stream into a churn scoring model that escalates months before customer loss effort high impact high

Feed service and complaint signals into a churn model that flags at-risk owners early and routes them to intervention. Validate the score against historical churn before triggering automated outreach. Start with the highest-value segment.
unlocks Expect earlier, cheaper retention saves on high-value owners rather than broad churn elimination. Track save rate on flagged accounts as the leading indicator.
References: Snowflake, Gainsight, Zendesk
6

Build a competitive value-gap index across price, features, trust, and switching cost reviewed monthly effort low impact high

Construct a monthly index scoring the company against rivals on price, features, trust, and switching cost from consistent inputs. Start with the segments where competition is sharpest. Keep the input methodology fixed so trends stay comparable month to month.
unlocks Expect sharper, faster competitive response rather than an immediate share change. Track month-over-month index movement as the leading indicator.
References: Crayon, Klue, Similarweb
The prioritized roadmap

Where to start, and in what order.

Move 01 · Immediate

Split the target into two plans

Separate $120B into revenue achievable through controllable levers and revenue that requires autonomy approval.

Move 02 · 30 Days

Activate the installed base as a revenue system

Build the monetization ladder: insurance to FSD subscription to upgrades. Wire the complaint stream into a churn model.

Move 03 · 60 Days

Put the moat in the model

Instrument the data loop: miles to model improvement rate to subscription attach to LTV. Report it as a metric.

Signal proof

The questions a board should ask next.

Weakness map

Risk 01Critical

The plan has one scenario

The $120B goal requires ~27% growth off flat revenue. The primary path runs through FSD/Robotaxi regulatory approval , an event Tesla cannot schedule.

Risk 02High

The loyalty base is showing churn signals before it has been monetized

~60% US customer loyalty is under-monetized. The same base is generating complaint signals that are the first inputs to customer loss.

Risk 03High

The strongest asset has no number

The real-world driving data loop cannot be purchased or approximated by any competitor. It does not appear in the financial model as a metric.

Board-level questions

If no one can answer this, the board is approving a single-scenario plan.

The answer to this question is the contingency.

This is the baseline. Without it, attach-rate targets are guesses.

An undated, un-owned strategic decision defaults every quarter.

This reframes service as a retention cost, not an operational one.

Without a metric, the moat is a narrative.

Reproducibility

Sigraphs is deterministic by construction. The pre-compute layer decides which signals fire, how they compose, and who owns them. The language model is demoted to a renderer: it may translate the computed skeleton into prose, but it cannot add, remove, rename, or re-rank any signal, owner, or score. Run the same inputs twice and the verdict and routed actions are identical.

This report is reproducible as a timestamped artifact. Any regeneration under the same inputs returns the same verdict and the same routed actions.