Sign in

We built an AI model to forecast Argentina's oil and gas.

Three models learned from Argentina's public registers, forecasting 36 months ahead for every well, block and operator.

What the model shows you.

Production decline

Learns how every producing lateral declines to forecast the next three years.

Activity pace

Reads frac crews and completion timing to project how many new wells come online.

Seasonal gas demand

Separates the winter swing in gas demand from real decline before forecasting.

+7 more forecasts from the same model

  • Frac jobs ahead
  • Wells coming online
  • Wait from drilling to first production
  • Ultimate recovery per well
  • The month still being declared
  • A low and a high for every month
  • Blocks and operators ranked by expected growth

Forecast starts where the declarations stop.

See global forecast reports

How the forecasts are built

  1. We compile public declarations and private data

    Every well's monthly oil and gas, every frac job and every investment plan Argentina publishes, joined with private data and checked every night. The newest month is never complete, so the forecast starts from the last full one.

  2. We rebuild each well's history

    The same well is filed differently from one file to the next. We tie its rows back to one wellhead, so its drilling, its frac job and its output read as one story.

  3. We forecast the wells already producing

    A shale well gives the most in its first months, then slows. The model learned how Argentina's wells slowed, and forecasts every producing well month by month, 36 months out.

  4. We forecast the new wells

    Frac jobs are filed months after they happen, so the latest months always look too quiet. We correct each operator for its own delay, then forecast the frac jobs and new wells still to come.

  5. We take the winter out of gas

    Argentina burns more gas in winter, so output swings with the season. The model learns with that swing removed and adds it back at the end, so a summer dip is never read as decline.

  6. We test it against the past

    We rerun the model as if it were one, two and three years ago, and compare with what was declared since. How far off it was sets the low and the high around every month.

See the forecasts on your blocks