Turn on committed use discount sharing, and create a commitment for the combined usage Sharing your committed use discounts across all your projects reduces the overhead of managing discounts on a per-project basis, and maximizes your savings by pooling all your discounts across your projects' resource usage. If you have multiple projects that share the same Cloud Billing account, you can enable committed use discount sharing so all of your projects within that Cloud Billing account share all of your committed use discount contracts. Your sustained use discounts are also pooled at the same time. That is, sustained use discounts are calculated using the total resources across these projects, rather than just the resources within a single project.

Reference link- https://cloud.google.com/compute/docs/instances/signing-up-committed-use-discounts#sharing_committed_use_discounts_across_projects

The above is a real world example of a company named Planet, where they sent around 80+ satellites to take pictures of earth every day, 24 hours. They run around 40,000 preemptible VMs concurrently.
Preemptible instances function like normal instances but have the following limitations:
Compute Engine might stop preemptible instances at any time due to system events. The probability that Compute Engine will stop a preemptible instance for a system event is generally low, but might vary from day to day and from zone to zone depending on current conditions.
Compute Engine always stops preemptible instances after they run for 24 hours. Certain actions reset this 24-hour counter.
Preemptible instances are finite Compute Engine resources, so they might not always be available.
Preemptible instances can't live migrate to a regular VM instance, or be set to automatically restart when there is a maintenance event.
Due to the above limitations, preemptible instances are not covered by any Service Level Agreement (and, for clarity, are excluded from the Compute Engine SLA).
The Google Cloud Free Tier credits for Compute Engine do not apply to preemptible instances.

Reference link- https://cloud.google.com/compute/docs/instances/preemptible

Explanation
Collect all weather data evenly across all cities. Mainly because it seems that the emphasis for data collection for ML is to make sure there are no holes in your data collection.