GrayStak is an independent press operation and an institutional signal product, drawn from the same field network, on the ground and online. Our field teams cover political and social dynamics as they unfold in video journalism, photojournalism, and written reporting, across institutional politics at the city, state, and federal level, immigration enforcement, protests and civic action, and investigative work spanning corporate governance, public finance, national security, and AI regulation. That same field network feeds DPVI, a scored signal for trading desks measuring Political Velocity: escalation by its rate of change, not the probability of outcomes.
GrayStak measures Political Velocity: the rate at which political escalation moves, in real time, from publicly available signals. One field network feeds both pillars, independent press on one side and a scored signal product on the other. They share a domain and a standard, and are governed apart.
An independent press operation documenting political and social dynamics in real time. Our field teams work in video journalism, photojournalism, and the written word, covering institutional politics at the city, state, and federal level, immigration enforcement, protests and civic action, and investigative work across corporate governance, public finance, national security, and AI regulation.
Independent documentation of events as they unfold. We document; we do not advocate. We do not produce sponsored or commissioned event media.
A scored signal product that measures the velocity of U.S. domestic political escalation. It is built from public signals and scored deterministically, read as one number through a five-state machine.
The engine is live. Validation work is in progress. State and score deliver to trading desks ahead of where attention indices measure and ahead of where news products fire.
The field network is a GrayStak Media capability, under editorial direction. What it observes on the ground informs independent coverage, and the same publicly observable conditions feed the DPVI Engine as one signal layer. The engine scores those public conditions, not GrayStak Media's published stories. Shared domain expertise, separate governance: coverage is not set by what the engine needs, and the engine is not steered by commercial pressure on the reporting.
A reporter in the DC metro covers an unusual pre-noon gathering. The same publicly observable events inform both the reporting and the measurement:
GrayStak Media is press. We cover American political movements, elections, and political conflicts in real time: video journalism and photojournalism from the field alongside written reporting, interviews, commentary, and analysis. Our work spans institutional politics at the city, state, and federal level, immigration enforcement, protests and civic action, and investigative journalism on matters where politics meets the economy.
We document ICE operations, courthouse activity, and community impact in real time, with primary video, photography, and on-scene reporting. High-risk field work, documented firsthand.
This is some of GrayStak's most circulated coverage and the work that has built our readership.
Hearings, votes, executive actions, policy moves, and ward-level political mechanics. Direct interviews with officials. The intersection of local power and national outcomes.
Direct on-camera interviews with candidates running for congressional, gubernatorial, and U.S. Senate seats. Documentation of campaign strategy and the political infrastructure behind elections.
Live field documentation of political and social dynamics as they unfold across U.S. metros. Primary video, photography, and first-person reporting from moments that often don't reach wire reports in real time.
Investigative reporting on matters of corporate governance, public finance, national security, defense industry geography, and AI regulation. Long-form work that goes beyond the daily news cycle to examine the structural intersections of business, policy, and political action.
This is reporting that takes time to develop and asks the questions a daily cycle can't.
Our primary coverage markets are Washington D.C., New York, and Chicago. We are actively expanding to additional markets and growing our media team.
GrayStak Media operates independently. We do not produce sponsored or commissioned event media. We do not accept advertising from politically-interested parties. Commercial activity in the engine pillar does not influence editorial decisions in the media pillar.
Editorial independence is structural. The two pillars run under co-founder authority: Christopher leads the engine and Ryan leads GrayStak Media, and a coverage decision cannot be commercially overridden.
Our reporting and analysis is published as GrayStak Media. Read recent work, including investigative pieces and ongoing coverage of the markets we work in.
DPVI measures the velocity of U.S. domestic political escalation. It is built from public signals, scored deterministically, and read as one number through a five-state machine. It is a measurement system, not an AI product: domain expertise produces the signal, and machine learning is bounded to upstream relevance filtering, which keeps the output auditable.
The "D" is a scope claim: domestic, U.S. The buyer base is global, trading desks running U.S. exposure overnight, where systematic measurement matters most.
Country-risk and geopolitical-risk indices run on multi-year, sovereign horizons. Attention indices measure attention once consensus has formed. Breaking-news and sentiment products fire coincident with events.
DPVI sits upstream of where attention indices measure and ahead of where news products fire, inside the consensus-gap window. It is a different measurement category that complements the attention indices a desk already runs without replacing them.
The engine moves the desk through five operating states (Baseline, Elevated, Critical, Resolving, Anticipatory) based on composite trajectory. A scheduled sweep maintains continuity between cycles; event-driven re-scoring fires when underlying signals move materially inside a single cycle.
Every score is sourced and every state transition logged. Component weights, state thresholds, dual-trigger parameters, and cadence are part of the Buyer Briefing, available to qualified prospects on request.
Markets price the probability of political outcomes well once a narrative forms, but they price velocity poorly. The window between when escalation starts moving and when markets reach a shared reading is the consensus gap, and DPVI is built to measure it.
The edge is not hidden information. The signals are public; the methodology reads them at finer resolution and faster than consensus forms.
The intellectual basis for DPVI is set out in Political Velocity (Sweat and Lazarus), the framework paper behind the velocity-over-probability claim and the consensus-gap construction. The paper is shared with qualified evaluators alongside the Buyer Briefing.
A measurement product is credible only if it can be wrong in measurable ways. GrayStak runs a structured validation program against the live engine: sensitivity testing on confirmed historical cascades, specificity testing on held-out non-events, identical architecture across both sets, and frozen calibration. The work is in progress. We share the matched results with qualified evaluators in a controlled way.
Confirmed historical political-escalation cascades scored on the production architecture with no event-specific tuning. Multiple cascade categories. The question: did the score register the cascades that actually happened?
Held-out non-events, moments that carried elevated political tension but did not cascade, run through the identical architecture on the same calibration. They are genuinely out-of-sample. The question: did the score correctly stay below the threshold?
Architecture frozen during the runs. Same model version across both sets. Identical floor configuration applied to both. The validity of the contrast depends on it.
DPVI is a single architecture, but the value it delivers depends on how a desk uses it. The matrix below maps where DPVI fits across institutional buyers, and where it doesn't. Trading desks are the beachhead; adjacent buyers come later, after the trading credential is earned.
Most quant firms could build a signal like this. The reason to buy is focus, not capability: political-escalation regime classification is rarely core enough to a firm's alpha thesis to justify rebuilding the cascade architecture, weighting and trigger logic, case-library calibration, and field-observation protocol in-house. The pilot lets a desk answer build-versus-buy on its own economics.
The reporting and the score answer to one standard: verified before it runs, sourced so anyone can check it, and kept free of commercial influence. The principles below govern what we let into our systems, what we publish, and what we do when we're wrong.
Field observations enter the system only after a documented verification protocol. Conflicting reports are flagged and held; we don't quietly average or merge them.
Every score component and every published claim is tied back to its source. We treat the audit trail as part of the product.
GrayStak Media operates behind a wall from engine subscriptions. We do not produce sponsored or commissioned event media. Institutional pilots receive no editorial preference, and editorial decisions are not visible to commercial functions until publication.
When automation does the work, like DPVI scoring or signal aggregation, we say so. When humans do the work, like frontlines coverage, fact-checking, or observer triage, we say so.
Errors are corrected at the source with a visible correction record. We don't quietly edit history. Every correction is timestamped and linked from the original.
Field observers operate under a documented safety and consent protocol. No observer is named without explicit permission. Anonymous source protections are maintained.
The full Editorial Standards document and the detailed DPVI methodology brief are available to qualified prospects, journalists, and institutional reviewers on request.
Request Standards Brief →These are absolute commitments that define the line our work does not cross, by design and as a matter of policy. They protect the people we document, the integrity of our signal, and the buyers and readers who rely on us.
We do not collect, store, or process personally identifying information about the people in the events we document. The unit of analysis is the event and the pattern rather than the individual.
Everything in our systems comes from publicly available information observed in public space. No private accounts, no leaked data, no purchased identity datasets, no covert observation. What enters GrayStak is what any member of the public could lawfully see.
GrayStak is not sold to, designed for, or made available for government intelligence, military targeting, law enforcement surveillance, or any classified or kinetic application. Our work serves civilian audiences only.
These limits are an ethical position before they are a commercial one. Public-signal measurement and intelligence work are different products with different ethics: one operates on what is visible and accountable, the other against subjects without their knowledge. GrayStak does not do intelligence work. We measure what is publicly observable, and we are accountable to the people we document.
The limits also protect the product: GrayStak's edge comes from being faster and more systematic on the same public sources everyone has, and crossing any of these lines would forfeit it.
We will not accept work, contracts, or partnerships that require us to cross these lines.
GrayStak is built and run by two co-founders. Pilot conversations and methodology questions go through us directly, without a sales layer in between.
CS · 01
Co-author of Political Velocity, the framework paper behind DPVI and the consensus-gap construction. Leads the DPVI Engine pillar: commercial direction, institutional sales, and product. The view that anchors the work: political velocity, not probability, is the right unit of risk for markets pricing escalation.
christopher@graystak.com →
RL · 02
Co-author with Christopher of Political Velocity, the framework paper behind DPVI. Background in international business, journalism, and media, with a creative practice that shapes GrayStak Media's editorial voice. Leads the GrayStak Media pillar: editorial direction, coverage standards, and the field network.
ryan@graystak.com →GrayStak started from a gap in how political risk gets measured. The standard methods react to national consensus, or register an event only once it has already broken, and they work on slow horizons. Material political events, meanwhile, keep arriving more often and with more intensity, and the cost of waiting for consensus keeps climbing.
That lag is an information asymmetry, and it puts people at a disadvantage exactly when they have important decisions to make. GrayStak exists to narrow it: to measure political escalation by its rate of change, before consensus forms, and to put that measurement, and the reporting behind it, in front of the people deciding on it. That conviction, that political escalation is better read by its rate of change than by the odds of any single outcome, is set out in Political Velocity, the framework paper behind DPVI.
The two pillars share one field network, on the ground and online. GrayStak Media reports the events; the DPVI Engine measures how fast they escalate. They are governed apart on purpose, and a coverage decision is never overridden for commercial reasons.
A paid evaluation, opening soon, with success criteria agreed upfront and calibrated to the buyer's own framework. The per-desk cohort is deliberately small; evaluation quality matters more than headcount. Founder conversations are available now to align scope, criteria, and timing ahead of the formal opening.
Our reporting and analysis is published as GrayStak Media. Independent press in video journalism, photojournalism, written reporting, interviews, commentary, and investigative work across institutional politics, immigration enforcement, corporate governance, public finance, national security, and AI regulation.