AMD stays a long, but the story is crowded - investigate only if you can isolate new supply or customer deployment detail
The Opportunity
The pipeline call is LONG 46% on AMD because the underlying mechanism is still supplier-side demand strength. The constraint is edge decay: Tier-1 pickup and broad reprints mean the market knows the general shape of the story. The remaining opportunity is not “AMD is in AI” - it is whether there are incremental, under-discussed details around packaging allocation, accelerator availability, or named deployments that can shift the next revision cycle.
The Timing
INVESTIGATE is mandated here because this sits in the propagation monitor: direction is intact, but being early is hard. Market regime is Mixed 58 with crosswind 56, so reactions can be sharp to single-data-point updates. Freshness is not scored (Fresh -), which is consistent with a theme that is already circulating. The timing tripwires are customer disclosures (hyperscalers, Oracle-style deployments), supply-chain confirmations on packaging capacity, and any evidence of pushouts.
The Evidence
Hydration is weak in this run, so we are not presenting a full linked evidence rollup. The upstream record indicates Tier-1 domains present (e.g., FT/Reuters/Bloomberg in the 6B bundle) and broad propagation, which is why the edge is classified as decaying. Representative domains include ft.com , reuters.com , and bloomberg.com . To justify action beyond “long beta,” you need one new auditable sub-claim the reprints do not already carry.