Put gas, wind, solar, grid and batteries into one portfolio. Get the maximum supportable load, in reliability metrics your lenders and utilities already use.
Have an access code? Use Create account at the top of the page. Already have an account? Sign in.
Snapshot demo using binomial unit outages (6% forced outage rate, 60 MW units); wind, solar and battery availability are simple assumptions here. The full engine runs hourly Monte Carlo with weather-driven renewables.
Start with the question you have to answer.
How much IT load can this site commit to a tenant, and what would raise it?
Find max IT load and Site feasibilityDo our generation and storage perform as one firm supply, within permits and grid terms?
The limits that bindIs the reliability number defensible, and what does a bad year cost?
Costs and contractsCan we rerun scenarios in minutes and hand over a documented, repeatable result?
Method and validationThree steps, the same for every analysis. The explainers below show each one.
Gas units, wind and solar sites, the grid connection and the battery, with the IT load and your reliability target.
See an hour-by-hour dayEvery hour is dispatched with random outages, planned maintenance and resampled weather, so rare bad stretches show up.
Why weather mattersSupply availability in nines, LOLH, EUE and LOLEV, with a 95% interval that shows how settled the result is, plus reports and an Excel workbook.
Watch an estimate settleScroll through one day. The load stays at 240 MW while the sun sets and the wind drops.
Illustrative dispatch for a single day. Nine Firm simulates thousands of these days, with real outages and real weather.
Each column is one day and each row is one hour of the day. Solar follows the seasons. Wind follows nothing: it stalls for days, then returns. Nine Firm resamples whole weeks of weather so these lulls stay in every simulated year.
Illustrative synthetic weather, generated in your browser. Hover a cell to read it. In the product, you load your own site data.
Most simulated years have no shortfall at all, and a few are bad. One year tells you almost nothing. Nine Firm runs thousands of years, optionally repeated from independent random seeds, and shows you the estimate settle: its 95% interval narrows as years are added, so you can see when you have run enough.
Illustrative. Each thin line is an independent run of the same simulation. The bright line is the one the readout follows.
Your results will be questioned by engineers and lenders. The method is documented, the terminology is standard, and the engine is checked against known answers.
Every analysis runs on the same simulation, so answers stay consistent when you change an assumption.
Enter your thermal units, renewables, grid connection and battery. See availability, loss-of-load hours, expected unserved energy and loss-of-load events against your target.
The tool searches for the largest IT load the portfolio supports at your reliability target, with uncertainty ranges from independent random seeds.
Model the grid as an asset with its own outage statistics. Compare islanded, grid-as-primary and grid-as-backup cases side by side.
Dispatch is by energy cost, with true tolling economics: fee, variable O&M and fuel kept separate. Find the cheapest mix that clears the target.
The largest load the site can serve with generation alone, the grid alone and both, for each grid size you list, as a low, typical and high range. Then every single equipment failure, and pairs of failures, are screened against your target.
Suggests the week each maintenance batch should start for the fewest loss-of-load hours, within the months and crew limits you set. The suggestion is checked against your current schedule on fresh simulated years before you apply it.
Steps through the quarters as resources come online and retire, and solves the supportable IT load in each. See the timeline.
A plant that looks firm on nameplate can fall short once its permits, its grid agreement and the weather have their say. Each of these is optional and applied hour by hour.
Annual caps on operating hours, fuel and NOx and SO₂ tons, for one plant or several. When a cap is used up, a plant stays off or runs only to avoid load loss, and the results show when that happens.
Split the grid into firm and conditional import that the utility may withdraw during curtailment events, drawn at random or from an hourly series you upload, with the hours and unserved energy during curtailment reported.
Gas turbine output and the datacenter's PUE can follow an hourly temperature series, so heat waves reduce supply and raise load in the same hours. Upload your site's data, or start from built-in hourly temperatures for seven US datacenter regions.
Peaking units can use a demand-based repair time, so they fail as often as their outage statistics imply. Maintenance schedules longer than the year are flagged instead of quietly dropped.
Reliability and cost come from the same run, so every number traces back to the same simulated years.
Capacity fees, variable O&M and fuel kept separate, with cost per MWh delivered and curtailed renewable energy charged as a PPA would charge it.
Capital cost, financing and annual recovery for each owned resource, built up to a levelized cost of energy, starting from cited public cost studies that you replace with your own figures.
Energy services and power sales agreements: payments, availability guarantees and damages for shortfalls, with a lender view of debt service coverage in an average and a bad year.
Studies live in your organization's workspace, with the reference data your team agrees on.
Invite colleagues or issue access codes. Projects are saved to the server with revision checks, so two people cannot overwrite each other.
Outage data, load shapes, PUE shapes and wind and solar profiles shared across the organization, with every version kept.
Duplicate a scenario, change one assumption and compare them side by side. Ready-made example projects help you start.
Ask questions about a finished run and get answers drawn from that run's own numbers. It is off unless your organization's administrator turns it on.
Supply availability: the share of hours in which the portfolio serves the full facility load. Three nines is 99.9%, about 8.8 hours short in a year. It describes the power supply only. It is not an Uptime Institute Tier rating, which classifies a facility's design.
The standard NERC and IEEE adequacy metrics. LOLH, loss of load hours: expected hours a year with any shortfall. EUE, expected unserved energy: MWh a year not served. LOLEV, loss of load events: separate shortfall events a year. Each is reported with the spread across simulated years.
Shortfalls come from coincidences: a unit trips during maintenance on a still, cloudy week. Nameplate sums and single-year studies miss these. Simulating many years, hour by hour, shows how often they happen and how sure the estimate is.
No. Nine Firm is a resource adequacy model: it balances supply and load every hour. Power flow, short circuit, stability and ride-through studies of the network are performed by the transmission provider or your engineers. Nine Firm tells you whether there is enough supply, and in which hours it is tightest.
The IT load and PUE, each plant's units and size, hourly wind and solar profiles, the grid import and its outage history, and the battery. Outage data has sourced starting points; cost defaults are placeholders to replace with your own terms.
In the browser, with simulations on our servers. Every request is tied to a signed-in person and their organization, and every database query is limited to that organization, so another organization cannot list, open or run your projects. The security overview is written for IT reviewers and says what has not been done yet.
Accounts are open by invitation or access code. Contact us for a demo on a site you are weighing.
Leave your email and we will get in touch to walk you through Nine Firm on a site you are weighing. Accounts are open by invitation; if you already have an access code, use Create account at the top of the page.
