Only about five months have passed between the podcast release (21 June 2025) and the current date (30 November 2025), and the claim is explicitly framed as “over the long term”. That horizon is typically several years, so there hasn’t been enough time to decisively confirm or falsify the prediction.

Current evidence actually points the opposite way in the short run:

• Snowflake continues to center its business on consumption‑based pricing and is actively promoting guides and playbooks that advocate usage‑based models, not retreating from them. (snowflake.com)
• Independent writeups of Snowflake’s pricing in 2025 still describe the familiar pay‑for‑what‑you‑use model across compute and storage (credits per warehouse size, TB‑per‑month storage), rather than a shift to a fundamentally different structure. (blog.twingdata.com)
• Specific tweaks, like Snowpipe moving to a fixed credit amount per GB for ingestion, simplify billing but remain per‑usage; they don’t abandon the consumption framework. (docs.snowflake.com)
• Financially, Snowflake’s product revenue is still growing >20% annually with strong net‑revenue‑retention and repeated guidance beats in 2025, which suggests the model is economically viable so far, even if customers are more cost‑conscious and optimize usage. (reuters.com)
• There is no clear evidence of large‑scale customer migration from Snowflake to lower‑cost Postgres/Supabase‑style alternatives; if anything, we see Snowflake acquiring Crunchy Data to offer its own Postgres option and case studies of firms migrating to Snowflake from other platforms. (constellationr.com)

Given the explicitly long‑term nature of the prediction and the relatively short time elapsed, plus the lack of clear evidence either of mass migrations away from Snowflake or of its pricing model collapsing, the fairest assessment as of November 2025 is that the prediction’s truth value is still undetermined.