Big Tech Expands AI Across Climate, Search and Public Services

Nvidia Opens Weather AI

Nvidia this week announced Earth-2, a groundbreaking suite of open, AI-powered weather and climate models that promise to transform how forecasts are made and who can make them. Unveiled at the American Meteorological Society annual meeting in Houston, Earth-2 includes a family of models, libraries and frameworks designed to accelerate forecasting workflows from raw data through to high-resolution predictions.

Earth-2 is the first fully open AI weather software stack, intended to democratize tools historically limited to governments and supercomputing centers. Instead of relying on traditional physics-based numerical weather prediction which often requires massive compute resources, Earth-2 models use AI to generate 15-day medium-range forecasts, short-term storm nowcasts, and data assimilation far more quickly and cheaply.

In addition to speed and cost advantages, Earth-2’s open-source nature lets developers, national meteorological agencies, energy firms, climate researchers and insurers customize and deploy forecasting models on their own infrastructure, preserving data sovereignty and analytical control.

Agencies such as Taiwan’s Central Weather Administration and The Weather Company, as well as enterprises in energy and finance, are piloting Earth-2 in workflows ranging from grid optimization to risk analysis for extreme weather. This reflects a broader shift in climate tech toward AI-assisted predictive analytics for real-world decision-making.

As reported by PYMNTS, Nvidia is furthering its data center push by investing $2 billion in CoreWeave.

Google Brings ‘Personal Intelligence’ to Search

Google’s Search team rolled out a major enhancement with the launch of “Personal Intelligence” inside AI Mode in Search. This new feature allows users (in the U.S., opt-in only) to let Google access private content like Gmail and Google Photos to provide more contextually tailored answers, for example, trip planning suggestions based on email itineraries or recommendations informed by past experiences captured in photos.

Unlike conventional keyword search, AI Mode with Personal Intelligence synthesizes signals from private data to anticipate user needs and deliver proactive insights. Google has emphasized privacy protections, noting that models don’t train on personal content and accessing this data is optional for subscribers.

The update underscores a broader trend: search engines evolving into assistants that act on personal context rather than simply retrieve results. But the shift also raises questions around data use and privacy expectations, especially among users and regulators wary of AI reading across personal inboxes and photo libraries.

Safety and Public Sector AI: Meta and UK

Meta made headlines with a pause on teen access to AI characters across its platforms, citing a need to refine safety and parental controls ahead of rolling out an updated version of the experience, reported by TechCrunch. This move reflects ongoing concerns about the interaction between youth engagement and AI-generated content, and Meta plans built-in controls that steer younger audiences toward age-appropriate topics.

Meanwhile, in the U.K., the government announced a partnership with Meta-backed AI experts to modernize public services using open-source AI. The initiative, set to run throughout 2026, will develop tools for transport systems, public safety and defense infrastructure, ensuring that government retains control and sovereignty over mission-critical AI deployments.

Enterprise and Defense: Salesforce’s Win

Salesforce secured a $5.6 billion contract from the U.S. Army to accelerate military modernization and readiness through advanced digital solutions. The multiyear award places Salesforce at the heart of cross-domain initiatives to bring AI, analytics and cloud-native workflows into defense operations from logistics and command and control to intelligence fusion.

The contract underscores how enterprise software firms are increasingly integral to national security tech stacks, as government agencies leverage commercial innovation to enhance agility, data access and operational decision-making. It also reflects sustained defense spending on AI-driven modernization, where civilian and military technology priorities intersect.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.