Google's NotebookLM integrates Zillow's cited home‑buying guidance into notebooks

Google’s NotebookLM integrates Zillow guidance into home‑buying notebooks

Zillow announces a partnership with Alphabet’s Google to surface Zillow’s home‑buying guidance inside a featured NotebookLM notebook, bringing citation‑backed, AI‑generated answers into an experimental research tool used for consumer queries. The integration delivers responses explicitly tied to Zillow.com articles and supports multiple exploration modes, including “Audio Overviews” that convert written guides into conversational audio hosted by two AI voices, the companies say.

The notebook is designed to help prospective buyers navigate high‑stakes moments — for example, the difference between preapproval and prequalification, typical steps after an offer is accepted, and budgeting for costs beyond a down payment — by providing fast, clearly cited guidance where many consumers are beginning research. Zillow frames the collaboration as part of a broader effort to extend its data and editorial resources into emergent AI platforms, saying it aims to reduce misinformation and reach users earlier in their decision process.

Google’s NotebookLM, positioned as a personalized research and thinking tool, benefits from the partnership by gaining authoritative, domain‑specific content to ground its outputs, a priority for companies developing consumer‑facing generative systems. Steven Johnson, co‑founder and editorial director of NotebookLM, says the featured notebook “meets home buyers at the start of their journey with answers they can trust,” while Zillow’s Jen Berger highlights the platform reach: “millions of people turn to Zillow at the start of their home journey,” she says. Both firms emphasize citation‑grounded responses and stress that the notebook will evolve as AI use in early‑stage research grows.

CoreWeave scales GPU capacity with $4 billion AI‑specialised data‑centre JV

Separately in the AI infrastructure market, CoreWeave unveils a $4 billion joint‑venture site designed as one of the largest U.S. GPU‑specialised data centres, targeting customers that need dense accelerator clusters for model training and inference. The facility adopts a GPU‑as‑a‑service model to let enterprises and hyperscalers rent high‑performance compute without upfront hardware ownership, reflecting surging demand for GPU capacity across cloud and enterprise customers.

OpenAI tempers compute outlook to roughly $600 billion through 2030

Industry sources report OpenAI is revising its multi‑year infrastructure projection, telling investors it now expects to spend about $600 billion on compute through 2030. The figure — characterized as a recalibration rather than a retreat — underlines both the scale of capital required for frontier model development and the pressure on chipmakers, cloud providers and specialised data‑centre builders as demand and procurement cycles evolve.