Source: OpenAI raised its projected cloud spending to ~$750B through 2030, up from a projection of ~$600B earlier this year, reflecting new cloud agreements
THE SO WHAT
A projected ~$750B in cloud spend through 2030 makes AI compute look like a sovereign-scale utility, not a SaaS line item. If your roadmap depends on hyperscaler economics, assume AI capacity will be pre-allocated by a few anchor tenants and lock in long-term commitments or alternative providers sooner than you planned.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIBluehost promises its new AI agents will deliver "a site or store that runs itself"
“A site or store that runs itself” is the new marketing baseline for SMB web platforms—Bluehost leaning into AI agents means hosting is shifting from infrastructure to ongoing operations. If you sell into SMBs, assume expectations for autonomous marketing, merchandising, and support are being set by their hosting provider, not by you.
Applied AIWhy trust remains AI’s workplace challenge
The bottleneck on workplace AI isn’t features, it’s trust—without transparency and accountability, employees route around tools no matter the promised efficiency. If you’re rolling out AI internally, invest as much in explainability, audit trails, and clear escalation paths as you do in model performance.
Applied AIAccelerating the frontiers of scientific discovery: Google’s $40M commitment to the Genesis Mission
Google’s $40M in AI tokens and credits for the Genesis Mission is another data point that frontier model access is becoming a strategic lever in science, not just enterprise IT. If you’re in R&D-heavy sectors, expect funding and partnerships to increasingly hinge on your ability to plug into these subsidized AI research ecosystems.
Applied AIHarry Potter publisher to receive millions in Anthropic copyright settlement
A $1.5B settlement that sends a multimillion-pound check to a single publisher like Bloomsbury is a clear signal: large rights holders can extract real cash from model training disputes. For operators, assume future AI contracts and data deals will carry explicit licensing line items—budget for it and tighten your own IP provenance.