Filter Power 12 planning through the strength of the available evidence without turning roadmap signals into invented specifications.
Power 12 has not been announced as a product. The planning question is whether the project can wait for stronger evidence or should act on supported Power 11 hardware now.
The confidence label is part of the fact. Remove the label and the record becomes less useful.
No IBM model, price, core count, clock speed, availability date, or sales manual has been published for a Power 12 product.
Planning implication: Use shipping Power 11 documentation for current purchases and keep future assumptions outside hard requirements.
Public reporting attributes a hub-and-spoke chiplet direction for the generation after Power 11 to IBM Power processor leadership.
Planning implication: Preserve design optionality, but do not infer model-level specifications.
IBM Power processor leadership publicly described Spyre continuity beyond the current generation without publishing a Power 12 product configuration.
Planning implication: AI inference remains a credible multi-generation theme, while accelerator form and performance remain unconfirmed.
Industry analysis extrapolates a future window from historical Power and IBM i release patterns.
Planning implication: Use the window for scenarios only. Do not delay a required project as though the date were committed.
Samsung process-node and 2.5D packaging combinations discussed in outside analysis have not been confirmed by IBM as Power 12 specifications.
Planning implication: Track the claim, but do not use it to size, price, or approve hardware.
A useful ledger does more than collect predictions. It records when a claim appeared, who said it, how direct the source is, what the claim does not establish, and which planning decision would actually change if stronger evidence arrives.