Finding a methane leak can give an operator a chance to recover saleable gas by better directing maintenance spending. Google’s latest satellite research points towards a wider role for Earth observation businesses in helping customers make those decisions.

In a research announcement published on 1 September 2026, Google describes a deep learning model developed in collaboration with NASA’s Jet Propulsion Laboratory, called MAPL-EMIT. It uses observations from NASA’s EMIT instrument aboard the International Space Station to identify methane plumes and estimate their source locations.

Google reports that the model detected 84% of expert-annotated plumes in its benchmark and identified around 50% more plausible plumes across approximately 1,100 EMIT data scenes. Those additional candidates are not all confirmed emissions. The research demonstrates detection capability but the financial return from deploying it remains to be established.

The economic incentive to act is substantial. The International Energy Agency’s Global Methane Tracker 2026 estimates that more than 35 million tonnes of annual methane emissions from fossil fuel operations could be avoided at no net cost. In these cases, the value of gas captured and sold or used exceeds the capital and operating costs of reducing emissions.

That is a global estimate rather than a promise that every detected leak will pay for its repair. A customer still needs to establish how much gas is escaping, if it can be recovered and what an intervention will cost. An EO service could help prioritise the required investigations across a portfolio of facilities which would allow maintenance teams to focus their time and budgets where the likely benefit is greatest.

GHGSat sells satellite and airborne methane monitoring services to help oil and gas operators identify emissions at individual facilities. The company’s business connects specialist observation capability with customers’ operational and reporting needs. For operators, the economic case rests on reducing gas losses, targeting inspections and supporting emissions reporting. So Google’s research enters a market where commercial suppliers already offer these services but it could create complementary services and raise awareness of the EO-based solutions.

A credible offer could combine satellite observations with asset records, engineering expertise and follow-up measurements. An operator would receive a prioritised investigation list that details evidence supporting each alert and a record of action taken. Revenue could come from ongoing monitoring and verification services, provided suppliers can demonstrate reliable delivery and a useful improvement in the customer’s decisions.

Quality control will be central to that proposition. The dataset documentation reports false positive rates of approximately 3–5% in a reviewed sample of high-confidence plumes, compared with approximately 50–55% for medium-confidence detections. Sending unfiltered candidates to customers could create unnecessary inspections and erode trust.

The same documentation distinguishes methane enhancement, the additional gas measured in the atmospheric column, from an emission rate. Estimating the latter requires further methods and wind data, introducing additional uncertainty. Coverage also depends on when and where EMIT collects observations. A missing plume cannot, on its own, establish that a facility has stopped emitting.

These constraints create room for specialist suppliers to assess confidence, arrange complementary observations and explain what the evidence supports. Buyers would need clear service terms covering observation frequency, validation and the treatment of uncertain results.

Regulation adds another potential source of demand. The European Commission’s current timetable introduces requirements from 1 January 2027 for importers to demonstrate equivalent methane monitoring, reporting and verification for fossil fuels under contracts concluded or renewed after 4 August 2024. Earlier contracts are subject to an obligation to make all reasonable efforts. Methane intensity reporting follows in 2028, with intensity limits applying to contracts concluded or renewed after 5 August 2030.

Satellite evidence could support supplier assessments and emissions investigations within that process. A plume map alone does not establish regulatory compliance. Commercial services will need to connect observations to the relevant facilities, reporting methods and verification requirements. The strongest business case will come from showing how better methane information changes an operational decision and delivers value to the customer.