• Type: refresh (subtype: narrative-prose + social-capture integration)
• event_id: 2026-06-16-memory-supercycle-phase0a-prose-refresh
• Source: Full-scan Phase 0a pass; ~280 fresh captures from June 16 run. Key memory-relevant captures: @NuttyCLD "Memory is the cleanest barometer of the AI boom" (60 likes, 10 retweets, 2026-06-16); ParadisLabs SNDK vs SNAP comparison (128 likes); Phase 0b precomputed summaries (2026-06-16) for cohort numerics.

Key captures that changed the read:

1. NVIDIA halving a memory tier in next-gen AI servers — reported, not confirmed (@NuttyCLD, 2026-06-16). A report that NVIDIA could halve a key memory tier in its next-generation AI server specifications sent memory stocks "swinging." NuttyCLD characterized the first market read as "demand destruction" and the second as "supply rationing" — and suggested both miss a bigger structural shift. This is logged as a new watch-for, not a thesis change: the critical distinction is tier substitution (some HBM replaced by cheaper DRAM) versus total memory reduction. Tier substitution redistributes demand within the cohort; outright reduction would be the first genuine demand-composition change. The TrendForce mid-June contract-price print becomes the cross-check gate. Filed to Watch For.

2. SNDK social mindshare remains high (ParadisLabs SNDK vs SNAP AI comparison, 128 likes, 23 replies). SNDK is prominent enough in the memory-ticker conversation to be used in a format-content comparison — confirms that the social consensus around SNDK is durable post-strike, consistent with the 2026-05-20 R20 Aschenbrenner positioning note (SNDK as the dominant 13F conviction signal in memory).

Shift summary: The Samsung strike chapter is fully closed; the dispute is background noise. The thesis pivots from "strike window + structural thesis" to "pure structural thesis + first deceleration signal watch + NVIDIA memory-tier architecture watch." TrendForce mid-June print and MU Q3 FY26 quarterly print are the next gates. Posture: HOLD; extended and consensus; do not chase.

• Effect on thesis: HOLD; posture unchanged — thesis structurally intact but cohort extended and consensus. The NVIDIA memory-tier architecture development is the one new signal in this refresh worth tracking as a potential demand-composition watch; it does not shift the thesis direction.
• Watch-for additions: NVIDIA memory-tier architecture question (tier substitution vs. total reduction — disambiguate at TrendForce mid-June print and any NVIDIA H2 2026 server spec disclosures).
• No numeric content added (Phase 0a rule). Data table reflects Phase 0b splice as of 2026-06-16.