Validation Path
We are not replacing your systems. We are building an intelligence layer.
Synestria identifies economic losses between systems your current tools manage separately. The pilot is how we validate that thesis against your actual operating data — in a controlled, reversible engagement.
Synestria is the operational intelligence layer for AI factories.
The pilot in one sentence
Thirty days to your EA baseline. No data leaves your facility.
We connect to the telemetry your power, cooling, and server stack already generate — no new sensors, no configuration changes. We establish a baseline. We produce a report showing exactly what we found and what it is worth. You decide what happens next.
Pre-product stage
We are looking for our first design partner.
Synestria is pre-product. The thesis is grounded in research. The architecture is designed. The MVP is in development — targeted for delivery within the next six to nine months. What we are doing now is finding the first operator — landlord or tenant — willing to partner with us to build and validate the platform against real operating data.
That partner shapes how the platform is built. They get the deepest engagement we will ever run — an EA baseline produced specifically for their infrastructure during the validation period. There is no production commitment and no obligation to continue. The operator who says yes first sets the standard the platform is built to.
Start a conversationWhat this is
One path. You are in or you are not.
We are pre-product. We are not asking you to evaluate a finished platform. We are looking for the operator who wants to be the first — who commits to the first site, shapes how the platform is built, and starts receiving EA recommendations the moment baseline is established. The output is not a report. It is the platform working on your infrastructure, continuously.
The path from here to live
Four steps. One commitment.
There is no menu of options. You commit to being the first site, or you do not. If you commit, this is what happens.
Step 1 — Weeks 1–2
Letter of intent
You sign a short letter of intent with a scoping deposit. This confirms you as the first design partner and locks the timeline. The deposit applies toward the Phase 1 contract. The LOI is a short, plain-language document — no complexity.
You are now the first. You will shape decisions no other operator ever gets to make about this platform.
Step 2 — Months 1–9
Build
Synestria builds the MVP with your infrastructure profile in mind. You have direct input on use cases, integrations, and reporting format. This is the only time the platform will be built to a single operator’s specifications.
No ongoing involvement required during the build. We come to you with questions, not the other way around.
Step 3 — Month 9–10
Go live
A pre-configured Synestria validation rack arrives at your facility and rolls into the approved location in your building. Connection to your existing power and cooling telemetry takes a single day. No new sensors. No changes to your control systems. Telemetry reaches the edge servers via hardware data diodes — physical devices that allow data to flow in one direction only, with no return path into your OT network at the hardware level. On large campuses we deploy building by building, so you never need to provision space across the whole campus at once. All data is stored on the edge hardware inside your building.
What we need from you: an approved rack location, a power circuit, and a technical contact for the installation window. We provide everything else.
Step 4 — Month 10 and ongoing
Baseline and EA recommendations
Synestria establishes an Economic Availability baseline over thirty days — observing power, cooling, and compute signals in parallel and identifying the losses no single-domain tool can see. Once baseline is complete, EA recommendations begin continuously.
Advisory outputs only. No automation unless you authorize it. You decide what to act on. The data is yours.
Platform maturity
Advisory today. Autonomous over time.
Step 4 is where the engagement begins, not where it ends. The platform is designed to progress through three levels of operational authority as the design partner validates each stage.
Level 1 — From baseline
Advisory
Synestria identifies EA losses, quantifies their economic impact, and delivers recommendations continuously. Your team decides what to act on and when. The system observes and advises. Nothing changes without a human decision.
Level 2 — Months later
Assisted
Synestria recommends and can apply approved fixes with explicit human authorization on each action. Faster response time, same level of control. The system does not act without sign-off — but sign-off now takes seconds, not hours.
Level 3 — After proven track record
Autonomous
After a validated human-in-the-loop period, Synestria acts within agreed operating parameters without manual approval on each action. The scope of autonomy is defined by you. Transitions between levels require your explicit sign-off.
How your team is alerted
An alert that tells you what to do. A record that builds itself.
Synestria does not surface anomalies for your team to diagnose. Every alert carries the full consequence chain — what is happening, what it will cascade into, what the recommended action is, and how long you have before impact. Your team decides whether to act. The system does not.
Alert delivery
Where your team already works
Alerts reach your team through the Synestria dashboard, email and SMS for on-call staff, and directly into whatever ticketing system you already run — PagerDuty, ServiceNow, or similar. Synestria opens the ticket automatically. Your team responds the same way they respond to any other operational event. No new workflow. No new tool to check.
Audit trail
The record builds itself
When Synestria creates a ticket, the ticket lifecycle becomes the audit record. Alert sent, acknowledged, action taken, resolved — all timestamped automatically. The time between ticket creation and closure is the actual resolution time. Compared against baseline, that is the time Synestria saved. No manual logging. No disputed numbers. Your ticketing system becomes the corroborating record for every line in the monthly value statement.
How the value is calculated
Your numbers. Our measurements. Nothing else.
The financial case for Synestria does not rest on estimates. It rests on a calculation that uses your own data at every step. Here is exactly how it works.
Step 1 — At onboarding
You provide the financial inputs
Before Synestria establishes a baseline, you supply three numbers in writing: your compute value per MW-hour, your fully-loaded operations labor rate, and your SLA penalty exposure per hour. These become the fixed multipliers for every calculation in the engagement. Synestria does not estimate or substitute them.
Step 2 — Days 1–30
The baseline sets the counterfactual
For thirty days, Synestria reads the telemetry your power, cooling, and compute infrastructure already generates — no new sensors, no configuration changes. The system measures how often events occur, how long they take to resolve, and what the EA degradation looks like during each one. That record becomes the baseline — countersigned by you — and defines what the campus costs without Synestria.
Step 3 — Every month after
Value credited is time saved times your rate
For every event Synestria catches and helps resolve, the credit calculation is: time saved versus baseline × your compute value × affected capacity. The formula is Synestria’s. The inputs are yours. The product of the two is a number neither party can dispute — and the monthly statement shows every event that produced it.
Conservative by design
Synestria credits only the delta between how long events historically took to resolve and how long they took with Synestria direction. If a new event class has no baseline comparator yet, it is excluded from the financial calculation until sufficient data exists. When in doubt, an event is not credited. The goal is that every number in the monthly statement is defensible without qualification.
What we are asking for
A commitment. Not a check.
We are not selling a finished product. We are looking for an operator — landlord or tenant — who understands the problem, sees the upside of being first, and is willing to sign an LOI so we can build toward their site. The commitment is mutual: we build to your infrastructure, you get the deepest engagement this platform will ever offer.
If the baseline does not show value
If the thirty-day baseline period does not identify a material EA gap in your infrastructure, the Phase 1 production contract does not activate. Data access is revoked and all collected data is deleted within five business days. No obligation is created beyond what the LOI specifies.
The first site sets the standard.
If you are a landlord or tenant operating at hyperscale and you see the problem Synestria solves — start here. If the 30-day baseline finds no material EA gap, the engagement ends. No obligation.