The 60-day-extendable framing is the cleanest textual structure of the cycle. Yesterday's 30-day framing was reporting noise — the formal MoU text uses 60 days as the operational window with extension by mutual consent. This restores the original Witkoff timeline (60 days post-ceasefire for substantive nuclear-program negotiation) and the digest's framework probability methodology stays calibrated. The Hormuz-mine-clearance commitment is the substantive operational deliverable Iran brings; the sanctions-waivers-for-oil-sales deliverable is the US side's economic-relief counterweight.

"Tangible verification" is Iranian state TV's framing of the trust-but-verify posture. Iran's position: no operational step without on-the-ground verification of US compliance first (sanctions-waiver implementation, blockade-lift demonstration). This is a sequencing-first negotiating posture — operationally workable, but adds 5-10 days to the implementation timeline once the text is signed. Rubio's 'a word, a sentence' framing is the public-confidence signal that the text-finalization is at the wordsmithing stage, not the substantive-parameter stage. June 2-3 fourth-round talks in Washington are the binding calendar checkpoint.

Failure-mode steelman: if CENTCOM's parallel 'self-defense' strikes (Monday-Tuesday) escalate further OR if Israel's southern Lebanon ground push provokes a major Hezbollah counter-strike on Israel proper before June 3, the MoU collapses regardless of textual-readiness. Both failure paths are operationally available in the next 7-10 days. Polymarket Iran-deal-by-mid-year implied probability remains the cleanest crowd read; the trajectory has lifted materially over the past 96 hours.

Netanyahu's public expansion-commitment is the most-direct collision with the Iran-Doha Lebanon-linkage clause. Iran has explicitly demanded any framework include a halt to fighting on all fronts including Lebanon; Israel has now publicly committed to the opposite trajectory. This is the single hardest unresolved variable in the MoU text and the highest-probability framework-collapse path. The June 2-3 dual-track calendar (Iran-Doha + Israel-Lebanon talks both in Washington) is the cleanest forcing-function week of the cycle.

Hezbollah's 32-operations claim is a high-tempo restraint signal: the operations are tactical (troop targeting, infrastructure damage) rather than strategic (rocket barrages into northern Israeli population centers, drone strikes on civilian infrastructure). The strategic-counter-strike scenario (last night's NEW prediction at 45%) has not yet materialized. If Hezbollah maintains tactical-restraint through the June 2-3 talks, the framework-probability lifts further. If a strategic counter-strike happens in the next 7 days, the MoU collapses.

Polymarket Russia-Ukraine framework-within-30-days probability stays at 28%; framework-within-90-days probability cuts slightly as the Iran-window-extension scenario becomes the modal expectation. European mediation channels (Macron, Meloni) continue with secondary-status. Ukrainian deep-strike campaign on Moscow Oblast continues at the post-May-7 elevated cadence.

The trajectory: Micron's market cap grew from ~$108B in May 2025 to $1.01T in May 2026 — a near-10x in one year. This is the cleanest single-stock embodiment of the HBM-wafer-capacity-reallocation cycle theme the digest has tracked since early May (Simon Willison surfaced the spillover-to-consumer-GPU sub-theme May 22). UBS's $1,625 target implies materially higher HBM demand than current consensus prices in — the question for the cohort is whether other broker upgrades follow OR whether $1,625 reads as an outlier-call.

Sector implications: Micron joining the $1T club at near-pure-HBM-leverage validates the structural thesis that memory-and-accelerator are now a single sub-cohort, not separate. Samsung's HBM business is the binding comparison — its hyperscaler customer mix and capacity-expansion roadmap are now the most-relevant peer-set datapoint. SK Hynix (the third major HBM supplier) faces the most-direct competitive pressure as UBS validates the structural memory-tightness premise.

Distributional shadow: HBM sold out until 2026 = ongoing consumer-GPU and gaming-hardware supply-tightness through the rest of FY27. This is the real-economy spillover the digest has been tracking since May 22. Watch for consumer-electronics-vendor price hikes and gaming-hardware-availability complaints over the next 2-3 quarters.

Pre-open technicals: futures-near-flat after Tuesday's +0.61% record close is a normal consolidation pattern — not a reversal signal. Polymarket's 61% open-higher (down from 91% yesterday) reflects the post-record-close baseline; the 61% level is still meaningfully above 50% coin-flip, signaling continued positive bias. The asymmetric tail-risk this morning is to the upside (if Iran-Doha text-finalization breaks overnight) rather than downside.

Cross-asset confirmation: WTI below $90 directly relieves the energy-shock-premium overhang; 10Y yield path matters more than ever for the AI-cohort multiple-expansion. If rates back up materially on the risk-on tape, the chip-cohort move from the rate-duration side compresses. Bond market is the binding cross-check.

Cycle structural read consolidates: AI-equity multiples remain rate-duration-sensitive rather than earnings-catalyst-sensitive. The chip-cohort leadership (Micron $1T, Nvidia $81.6B record) is the cleanest carry trade; the cycle continues to grind toward the June 17 FOMC + Iran-Doha-resolution joint inflection.

Structural read: the AI-capex cycle has moved through three phases — (1) hyperscaler announcements (2024), (2) accelerator-vendor margin-expansion (early 2025), and now (3) memory-vendor margin-expansion (mid-2026). Micron's $1T milestone validates phase 3 as standalone-investable. The implication for cohort positioning: HBM-vendor long exposure (Micron, SK Hynix, Samsung's HBM unit) is now the cleanest AI-cycle carry trade for the next 2-3 quarters, distinct from accelerator-vendor (Nvidia, AMD) and hyperscaler (Meta, Microsoft, Google) exposure.

Distributional read: HBM-wafer-capacity-reallocation toward AI accelerators continues to ripple into consumer-GPU and gaming-hardware availability. Simon Willison surfaced this sub-theme May 22; today's UBS-target + HBM-sold-out-until-2026 confirms the spillover persists through FY27. Expect consumer-electronics-vendor price hikes and gaming-hardware shortage complaints to surface in earnings prints through Q3-Q4.

Risk: if a memory-pricing-cycle peak emerges (Samsung capacity expansion, SK Hynix Korea fab ramp, Chinese memory makers' yield improvements), the HBM-tightness premise gets challenged. Monitor Korean Hi-Tech earnings (SK Hynix Q2 prints, Samsung Memory division commentary) for the first signs of supply-side easing.

Adoption-metric watchlist: (1) AI Ultra subscriber growth at the new $100/month price point — does the 60% cut from $250 deliver the user-base expansion Google is betting on? (2) Spark active-user retention through the first 7-10 days — is the 24/7 cloud-resident-agent value-prop sticky enough to drive daily-active usage? (3) MCP-integration usage frequency by category (productivity Canva > scheduling OpenTable > payment Instacart) — does the user-trust gradient drop sharply on payment-flow vs non-payment integrations?

Security read: PromptArmor's Microsoft Copilot Cowork demo generalizes to any agentic system with email-compose + external-image-render. Spark's MCP-payment-flow integrations (Instacart, and any future financial-flow MCPs) are now the highest-priority security category. The first prompt-injection-induced unauthorized-transaction incident — if it happens in the next 4-6 weeks — would be the cycle's first major operational test of consumer-agentic-AI deployment trust.

Competitive posture: OpenAI's ChatGPT Pro at $200/month + Anthropic's enterprise pricing both face direct pressure from Google's $100/month cut. The strategic question is whether OpenAI and Anthropic match in the next 4-8 weeks or defend their margin. Mythos (Anthropic's frontier model that has been described as 'surging forward' in the cohort race) is the offsetting positive for Anthropic's pricing power.

The Bay Area + Korea + Taiwan chip-cohort triangle is the structural beneficiary of the HBM-leveraged AI cycle. Bay Area's specific carry: Nvidia (HBM-buyer), Synopsys/Cadence (EDA tools), Applied Materials/Lam Research/KLA (equipment makers selling into both Korean memory fabs and Taiwanese accelerator fabs). The UBS Micron $1,625 target indirectly validates the equipment-vendor long thesis on continued capex through FY27.

SF Climate Week thread (Day 11+): AI-data-center-energy-demand is now the most-discussed sub-theme. The chip-cycle's structural memory-tightness and accelerator-deployment growth both translate into materially higher data-center power consumption. Bay Area civic implications include grid-load planning for new AI-data-center campus projects (PG&E rate-design + transmission-line approval as binding variables).

The Bay Area transit-funding gap (Muni/BART/Caltrain/AC Transit = 80% of Bay Area transit ridership, with BART's $376M structural deficit as the highest-visibility sub-problem) is the binding operational constraint on the SB 79 density model. Without regional/state/federal closure of the operating-deficit gap, the TOD-density premise loses its foundation. SB 63 ballot measure is the regional revenue answer; passage critical.

What changed: BusinessToday explicitly states 'monsoon official incoming misses its May 26 deadline, Kerala onset now expected between June 2-4.' Multiple outlets corroborate the delay narrative. The digest's framing on May 24 ('declared TODAY, earliest since 2009') was based on IMD's then-current reporting; in the 72 hours since, the IMD's operational read has either revised or the earlier declaration was provisional. The clean self-correction signal: the operational monsoon-onset date is now best-tracked as June 2-4, not May 24.

Operational impact: when the monsoon does arrive, IMD says initial intensity may be subdued — no sharp surge, slower northward progression than usual. NW India mainland onset remains roughly June 5-10 in earlier estimates; with the Kerala onset slipping to June 2-4, NW mainland onset slips into the second week of June or later. Rural-demand FY27 thesis trajectory shifts modestly later.

IMD has issued yellow alerts for four Kerala districts (Thiruvananthapuram, Kollam, Alappuzha, Ernakulam) with thunderstorm + heavy-rainfall warnings through May 26-27. Heatwave-to-severe-heatwave conditions continue in Central + NW India for the next 2-3 days. The dual-zone weather picture (active rain south, acute heat north) persists; the timing of southern→NW progression is the binding macro variable.

Macro-stress timeline now extends: the original early-onset trajectory (Kerala May 24 → NW mainland late May / early June) implied 2-3 weeks of acute heat-stress before relief; with Kerala onset delayed to June 2-4 and NW mainland onset slipping further, the acute heat-stress window now extends 4-5 weeks total. Rural agricultural production faces additional water-stress days; tier-2/tier-3 consumption deceleration compounds. UBS's 6.2% FY27 growth cut from 6.7% gets reinforced.

Power-grid implications: NW grid load through late June is now expected to remain elevated; coal-fired plant utilization stays high. Air-conditioning demand sustained at peak-summer levels through the extended heat-stress window. Watch for any grid-stress incidents (load-shedding, fault tripping) over the next 7-10 days — those would be acute political-economy stress signals.

Per Holland & Knight, Duane Morris, Womble Bond Dickinson, and WR Immigration's published practitioner alerts, the legal challenges are anticipated 'potentially on grounds that USCIS has imposed new substantive standards without notice-and-comment rulemaking under the APA.' The memo is effective immediately and applies to ALL pending and future Form I-485 filings — meaning ~700K+ Indian-origin H-1B-to-EB-2/EB-3 AOS-pending population is directly affected the moment any new I-485 is filed or adjudicated.

Venue read: NDCA (Northern District of California) or DC District Court remain the two most-likely venues. NDCA gets the geographic-concentration argument (largest affected H-1B/L-1 population) and a generally pro-immigrant-rights bench; DC District Court gets the nationwide-impact APA framing with proximity to the rulemaking record. Plaintiff coalitions may pursue both venues in parallel for split-track risk management. If a TRO or PI lands in the May 28-June 4 window, that would directly pause the memo before it materially shifts AOS approval rates.

Indian-origin tech worker community impact continues to track. Bhutoria's 'unprecedented crisis' + 'backdoor ban' community-pushback framing remains the clearest community-level signal. WR Immigration employer-advisory framing is the most-cited HR-team operational guide. The prediction (litigation achieves partial pause/PI within 60 days: 55%) remains operationally on schedule.

If the AOS memo litigation succeeds with a nationwide pause, the signature rule may become next-priority for a similar APA-rulemaking challenge — the procedural footprints are nearly identical. Current litigation-sequence path: AOS memo first (May 28-June 4 filing window), signature rule second (if AOS succeeds, next 30-60 days).

Key threads: simulation-as-first-class-infrastructure for product orgs, LLM-prototype-velocity-vs-production-reliability gap, internal-eval design tradeoffs, and Parakhin's framing of how a large product org should architect its AI-development tooling. Concrete examples from Shopify's deployment of internal AI tools across the engineering org.

Why listen: directly relevant to platform-infra engineering leadership — internal-tool design tradeoffs, customer-simulation infrastructure, and the org-design pattern of treating eval/simulation as a first-class engineering function rather than QA afterthought are all transferable to any large-scale platform team building AI features.

Topics: safety-budget engineering, hardware-in-the-loop simulation architecture, edge-vs-cloud inference tradeoffs for safety-critical systems, regulatory frameworks across L3-L5 autonomy levels, autonomy-platform commercial-model evolution from pure-tooling toward integrated-autonomy-stack.

Why listen: physical-AI distributed-systems engineering — the disjoint constraints between cloud-scale platform infrastructure and edge-compute autonomy stacks are an instructive contrast for platform engineering leaders thinking about agentic-AI architecture in the age of cloud-resident agents.

Inputs: (1) MoU text refined to 60-day extendable with Hormuz-mine-clearance + sanctions-waivers + no-nukes commitment as substantive deliverables; (2) Rubio's 'a word, a sentence' framing signals wordsmithing-stage finalization; (3) Iran's 'tangible verification' posture adds 5-10 days to implementation; (4) Israel's expansion-vow on Lebanon is the structural counter-pressure.

Failure-mode steelman: if Netanyahu's expansion-of-Lebanon-operations provokes a major Hezbollah counter-strike in the next 7 days, the Lebanon-linkage clause becomes a framework-blocker. If Iran's 'tangible verification' demand stalls implementation post-text-finalization, the MoU signs late but lacks operational follow-through. Either failure path drags probability below 50%.

Inputs: (1) Memo imposes new substantive standards without notice-and-comment rulemaking — strong APA challenge surface; (2) H-1B/L-1 dual-intent statutory framework provides clean Congress-intent counter-argument; (3) ~700K+ Indian-origin H-1B-to-EB-2/EB-3 AOS-pending population provides standing concentration for class-action framing; (4) NDCA + DC District Court are both venues with track-record of granting preliminary relief on procedural-rulemaking challenges.

Failure-mode steelman: if the first filing lands in a less-sympathetic venue (Texas Eastern, Florida Southern, Arizona) OR if the court rejects the APA-rulemaking framing in favor of the agency's discretion argument, no early relief lands. Probability the litigation eventually achieves partial pause within 60 days: 55%. Probability of TRO/PI in the first 14-21 days specifically: 55%.

Inputs: (1) Futures flat-to-up pre-open (+0.13%) is normal post-record consolidation, not reversal; (2) WTI below $90 sustains the energy-shock-premium relief; (3) Micron $1T + UBS $1,625 target sustain the chip-cohort multiple-expansion path; (4) Polymarket 61% open-higher is the crowd confirmation.

Failure-mode steelman: intraday Iran-headline reversal (worst case: Trump-Iran public spat, CENTCOM strike, Hezbollah counter-strike) reverses the tape; OR 10Y yield backs up sharply on the risk-on tape (rates-duration compression). Probability of fresh-record close: 60% — moderately above the 50% coin-flip baseline, consistent with the Polymarket open-higher 61%.

The substantively positive signals are the Iran-side concrete operational commitments. Hormuz-commercial-traffic-within-one-month is a publicly-stated calendar commitment that markets read as credible (oil -5.55%). This sits alongside the morning's 60-day extendable MoU framing with mine clearance + sanctions waivers + no-nukes commitment. Iran is moving the substance forward.

The substantively negative signals are US-side political-rhetoric hardening + parallel kinetic escalation. Trump's 'fumes' + 'not satisfied' + 'finish the job' framing is the most-direct ultimatum-style language of the cycle. The White House's 'complete fabrication' rebuttal on the Iran-leaked MoU details indicates Iran is overpromising the US-side concessions for domestic-political purposes — a typical end-stage negotiating pattern but one that creates trust-deficit headlines. The new US strikes (Hormuz site, 4 drones, Bandar Abbas ground control) are framed as 'measured, purely defensive' but represent ongoing parallel-pressure mechanics.

Market read: positive. S&P fresh record close on the Hormuz-commitment + oil collapse. Political-rhetoric read: hardened. Both reads are consistent with end-stage negotiating behavior — substantive movement on the parameters with public posturing on the framing. The June 2-3 fourth-round Washington talks remain the binding calendar checkpoint.

The post-Tuesday tempo-reset is operationally significant — high-intensity operations can be sustained for 1-3 days but typically require regrouping windows. If the IDF is pre-positioning for further northward ground push past the Litani in the June 2-3 dual-track week, today's relative pause is the preparation-phase. If the IDF is pausing to absorb international-diplomatic-pressure post-Tuesday's expansion, today is the de-escalation signal.

Hezbollah-restraint signal: the strategic counter-strike (rocket barrage into northern Israeli population centers OR drone strike on civilian infrastructure) has not materialized. The 45% probability from last night's NEW prediction now leans modestly toward the restraint scenario — possibly Tehran signaling restraint to Hezbollah ahead of the June 2-3 Doha+Washington dual-track week. If restraint holds through June 3, the framework probability lifts; if a counter-strike happens in the next 5-7 days, the framework collapses.

European mediation channels (Macron, Meloni, Tusk) continue with secondary-status. Ukrainian deep-strike campaign on Moscow Oblast continues at post-May-7 cadence. Framework-within-30-days probability stays at 25-28%; framework-within-90-days probability cuts further as Iran-window-extension scenario becomes the dominant expectation.

Sector rotation today: Dow leadership (defensives + financials + industrials) vs Tuesday's tech leadership. Today's pattern is broadening-of-the-rally rather than concentrated-tech-leadership — a healthier structural read for the cycle. Energy stocks underperformed on the oil -5.55% leg; defensives, financials, and rate-sensitive industrials all participated.

Cross-asset: WTI -5.55% to $88.68 + 10Y yield path (verification needed) + risk-on-narrow-breadth pattern. The energy-shock-premium relief continues to be the dominant driver. Iran's Hormuz-within-1-month commitment is the credibility anchor markets are pricing.

Predictions methodology hit: the morning's 60% fresh-record-close call HIT cleanly. Polymarket open-higher 61% was the crowd confirmation. Methodology consistency between editorial call and crowd read produced the win. Updated cycle resolved-predictions track: 3/6 directional, 4/6 if framework-probability calls trending in the right direction count partially.

UBS's $1,625 target stays the high-end of the 46-broker consensus distribution. Watch for follow-on broker upgrades over the next 1-2 weeks — a wave of upgrades would validate the UBS outlier-call; a stable distribution with UBS as the lone $1,500+ target would suggest the thesis is structurally-priced.

Distributional shadow: HBM-sold-out-until-2026 confirms consumer-GPU and gaming-hardware supply tightness through FY27. Nvidia consumer-GPU pricing + AMD Radeon pricing + console-hardware (PS5 Pro, Xbox refresh) availability all remain operationally constrained.

Operational read: the 'unexpectedly high LLM costs from staff usage' signal is the cleanest demand-pull data point of the cycle. Enterprise IT teams discovering LLM costs they didn't budget for because employees adopted the tools organically = bottom-up adoption, not top-down mandate. This pattern typically precedes structural-revenue-acceleration in B2B SaaS cohorts.

Cohort implications: if Anthropic prints its first profitable quarter in the next 1-2 quarters, the enterprise-AI-vendor unit-economics premise gets validated structurally — different from OpenAI's consumer-mass-market path. The two paths are operationally distinct but both signaling pmf. Google Spark beta launching this week into this validated-demand backdrop has materially better launch conditions than the digest had been pricing.

SQLite AGENTS.md significance: SQLite is one of the most-deployed databases in the world (every smartphone, browser, embedded system). Its explicit rejection of agentic PRs sets a precedent that other major open-source projects will likely adopt. The asymmetric pattern — accepting AI-assisted bug reports if well-documented but rejecting AI-authored code contributions — is the right operational discipline for projects where code-quality and security are paramount.

Spark beta adoption metrics window continues. Three trust-ceiling variables remain: permission-grant rates, prompt-injection security incidents, regulatory signaling. The Anthropic-OpenAI-pmf framing from Simon's other post today is the macro-context: Spark launches into a structurally-validated demand-pull environment.

The rally-broadening pattern (Dow leadership vs Tuesday's tech leadership) is healthier for the cycle. Bay Area participation: AI-cohort + chip-cohort still positive; defensives (PG&E, Wells Fargo Bay Area exposure) and industrials (Tesla, Lockheed-adjacent suppliers) leading the breadth-broadening. SpaceX roadshow June 8 runs into a tape with both growth-and-value participating.

SB 63 ballot measure regional revenue answer remains in play; passage critical. Without state/federal closure of the BART $376M structural deficit + Muni/Caltrain/AC Transit operating-deficit gaps, the TOD-density premise loses its operational foundation.

Treasure Island TOD: thousands of housing units + public parks + ferry access + new transit connections all proceeding on the long-term plan. West Oakland Mandela Station TOD: residential + retail + office + walkability + public-infrastructure upgrades. Both represent the SB 79 model in early-implementation form.

State-preemption architecture: SB 79 overrides local jurisdiction blocking power on zoning grounds for transit-adjacent density. Bay Area cities historically the strongest opponents (Berkeley, Cupertino, Palo Alto, Mountain View) — the implementation-interpretation battles over the next 12-18 months will determine the practical impact of the state-preemption model.

Rural-demand FY27 thesis trajectory shifts modestly later on the onset-delay. UBS's 6.2% FY27 growth cut (from 6.7%) gets operationally reinforced. The India-US trade deal ("within weeks" per Rubio) remains the most-important offsetting positive variable; if the deal closes in the next 4-6 weeks, India absorbs the energy-shock + monsoon-delay combination with the trade-deal-services-and-tech-export-tailwind partially offsetting.

India-US trade deal calendar: Rubio said "within weeks" — the tariff-line-by-tariff-line negotiation continues on the US side as the bottleneck. Deal scope per prior reporting: digital trade, services market access, defense-tech procurement preferences, adjusted tariff schedules. India side has been signaling readiness; closure in next 4-6 weeks would be a material macro positive offsetting the monsoon-delay + UBS cut combination.

If a filing lands tomorrow (May 28), the TRO/PI request typically follows within 24-48 hours. NDCA or DC District Court continue as the most-likely venues. The litigation-track probability (TRO/PI within 14-21 days from sympathetic venue: 55%) stays operationally on schedule for this morning's prediction calibration.

Indian-origin tech worker community impact: ~700K+ H-1B-to-EB-2/EB-3 AOS-pending population continues to be most-directly affected. Bhutoria's 'unprecedented crisis' framing + WR Immigration employer-advisory framing + AILA litigation-preparation framing all continue to track. The May 28-June 4 filing window is the cleanest binary the digest is tracking across all sections.

If the AOS memo litigation succeeds with a nationwide pause, the signature rule may become next-priority for a similar APA-rulemaking challenge — procedural footprints are nearly identical. Current litigation-sequence path: AOS memo first (May 28-June 4 filing window), signature rule second (if AOS succeeds, next 30-60 days).

Key threads: simulation-as-first-class-infrastructure for product orgs, LLM-prototype-velocity-vs-production-reliability gap, internal-eval design tradeoffs, org-design pattern of treating eval/simulation as a first-class engineering function rather than QA afterthought.

Why listen: directly relevant to platform-infra engineering leadership — internal-tool design tradeoffs, customer-simulation infrastructure, and treating eval/simulation as a first-class engineering function are all transferable to any large-scale platform team building AI features.

Topics: safety-budget engineering, hardware-in-the-loop simulation architecture, edge-vs-cloud inference tradeoffs for safety-critical systems, regulatory frameworks across L3-L5 autonomy levels, autonomy-platform commercial-model evolution.

Why listen: physical-AI distributed-systems engineering — disjoint constraints between cloud-scale platform infrastructure and edge-compute autonomy stacks are an instructive contrast for platform engineering leaders thinking about agentic-AI architecture in the age of cloud-resident agents.

The methodology-confirmation: when editorial call and crowd-read aligned within 1pp (60% vs 61%), the directional outcome was high-conviction. The methodology learning from yesterday's MISS (single-narrative directional calls in active-diplomatic windows are higher-variance) was operationally calibrated today by using the framework-probability lens instead of single-narrative.

Pattern for cycle predictions: when Polymarket and editorial call align, conviction is high. When they diverge (yesterday's 60% editorial-down vs 91% Polymarket-up), the Polymarket signal is the higher-information-density read in active-diplomatic windows. Future cycles will weight Polymarket more heavily during active negotiation periods.

Mixed-signal day reads bearish for the June 2-3 specific calendar. The end-stage negotiating pattern (substance moving + posturing hardening + parallel pressure) is consistent with a framework signing in the next 14 days but NOT necessarily at the June 2-3 round specifically. The framework-within-14-days probability stays ~58%; framework-by-June-3 probability cuts to 48%.

Failure-mode steelman: Trump's 'finish the job' framing + the Bandar Abbas strike could push Iran to harden its 'tangible verification' posture, extending implementation timeline by 5-15 days. Iran's leaked-MoU-details-then-WH-fabrication-rebuttal pattern indicates trust-deficit headlines that complicate text-finalization. If Lebanon escalation resumes tomorrow at Tuesday's pace, framework probability drops below 40%.

Restraint-pattern read: when both sides operationally pause ahead of a calendar checkpoint, the standard interpretation is preparation-for-negotiations rather than preparation-for-escalation. Tehran's possible signaling-restraint-to-Hezbollah ahead of Doha + Washington dual-track is the most-likely operational read.

Failure-mode steelman: if Netanyahu's expansion-vow translates to renewed strike tempo tomorrow (Thursday) or Friday before the June 2-3 talks, Hezbollah's tactical-restraint becomes much harder to hold. The 38% probability stays operationally appropriate; not zero, but materially lower than yesterday's framing.

AILA, ImmDef, APAJC, MALDEF, and major immigration law firm consortia have all signaled active complaint-preparation; the question is venue-coordination not preparation-readiness. NDCA filing earliest tomorrow morning (May 28) PT; DC District Court filing earliest tomorrow morning ET. If the filing lands tomorrow, TRO request likely follows within 24-48 hours, putting a possible early-relief decision into the May 30-June 4 window — exactly overlapping the Iran-Doha June 2-3 dual-track week.

Failure-mode steelman: practitioner coalitions choose to delay filing into the June 2-4 window to time the TRO/PI decision against the Iran-Doha headlines (less competition for judicial attention). Probability of filing in next 48 hours: 65%; probability of filing by June 4 (the full original window): 90%.