American independent oil and gas company Anadarko is migrating its exploration and production (E&P) models, applications and platforms to Google Cloud as it ramps up its relationship with the cloud services provider.

Anadarko developers are using GCP to build apps that monitor drilling and completions operations in real time, Google said Tuesday and meanwhile running live data analysis from offshore and onshore operations.

(The $11 billion revenue company is best known for its involvement in the largest gas find in 20 years in Mozambique. Anadarko first discovered gas off the country’s coast in 2010 and has now identified a massive 75 trillion cubic feet of natural gas.)

Google Cloud is carving out something of a niche in the upstream O&G space, with tie-ups in play with Total and Chevron, to name two others. Rival AWS touts a relationship with BP, but this is running SAP applications rather than on the E&P side.

Anadarko Cloud Migration

The shift to Google Cloud is allowing Anadarko to be “more precise with their efforts and making it faster for them to learn what’s happening with ongoing operations”, Google Cloud said in a release Tuesday.

Darryl Willis, Vice President, Energy, Google Cloud blogged:Since early 2018 they’ve been streaming all rig and completion crews data into a GCP (Google Cloud Platform) tenant. They’re now deploying process analytics models and allowing real-time data analysis from onshore and offshore facilities, which lets them predict—and ideally avoid—unplanned shutdowns.”

See also: AWS vs Azure vs Google Cloud: Who Wins on Latency, Performance?

Anadarko is tapping a range of Google Cloud services, he noted: “For example, they’re applying deep learning methods to build high-density seismic interpretations for subsurface characterization. They’re using Google Kubernetes Engine and BigQuery to get near-limitless scalability for their application performing log correction.”

He added: “On the back end, Cloud TPUs and Nvidia GPUs are powering their model training workflows. And the combination of Cloud Pub/Sub, Cloud Dataflow and AutoML is allowing them to deploy predictive maintenance systems faster, and at a lower cost than traditional IT operations.”

Willis also pointed to a joint project with French supermajor Total. This is an AI effort to help geologists, geophysicists, reservoir and geo-information engineers explore and assess oil and gas fields faster and with minimal environmental impact. The aim is to make it possible to interpret subsurface images (notably from seismic studies using Cloud Vision) and automate technical document analysis using Cloud Natural Language.