CMQ_033predictionSemis/MemoryHBM-demand

Global HBM demand will reach 32 billion Gb by 2026 — HBM is the absolute critical resource constraint in the AI compute supply chain.

Predictor: Morgan Stanley

Prior probability

82.0%

Current probability

75.1%

evolves via intake + LBP

Conviction

5/5

Signal quality

Resolution

pending

Window

2026-01-01 – 2026-12-31

Edges in / out

1 / 0

Tickers exposed

Prediction text

Global HBM demand will reach 32 billion Gb by 2026 — HBM is the absolute critical resource constraint in the AI compute supply chain. | Quarterly HBM shipment data

Key catalyst: Quarterly HBM shipment data

Watch events: SK Hynix / Samsung / Micron HBM shipment disclosures; NVIDIA reserved HBM capacity.

Resolution evidence

Status: pending

SK Hynix + Samsung + Micron combined HBM capacity ramping to meet ~30-35B Gb range through 2026; demand-driven.

Predictor: Morgan Stanley

κ + Brier as of 2026-05-22

Full calibration →

κ (discount)

0.633

Brier

0.0442

excellent

Hits / Misses

1 / 0

of 2 resolved

Hit rate

50.0%

Calibration plot (stated vs observed)

Evidence about this node from Morgan Stanley is multiplied by κ in /api/intake. Lower κ = less weight; floors at 0.10 (effectively silenced) and caps at 1.00 (full weight).

Reference class: memory_supercycle_2y

Linked

All classes →

Named memory-market supercycle (DRAM/HBM) persists/intensifies into stated year (n>=2y forward)

Base rate

—

4/8 historical

Inside weight

—

Outside weight

—

no pull

inside 75.1% → blend 75.1% (Δ 0.0pp)

Tetlock-style outside view: at TRF=1 (just predicted), outside view dominates (w_in=0.3). At TRF=0 (deadline), inside view dominates (w_in=1.0). The blend regularizes overconfident inside views toward the historical base rate.

Probability over time

7 prob_history rows

intake v2milestone miss sweeplbp propagationreference class assignedlegacy v1prior_prob (analyst seed)current = 75.1%

Milestone chain

Pre-event signals (upstream prereqs + window checkpoints) → resolution event → downstream cascades. Status/dates update from linked nodes; re-derive nightly via scripts/ops/derive_milestones.py.

Leading chain: 1 fired ✓ · 2 overdue ⏱ · 2 pending

2026-01-10hitMicron HBM capacity sold out through 2026 confirmed
How: Micron publicly confirms HBM capacity fully sold out for 2026 calendar year
Source: CNBC — AI memory is sold out, causing an unprecedented surge in prices (Jan 2026)conf 99%
2026-03-04overdueQ1 window check-in (25%)
2026-05-06overdueQ2 window check-in (50%)
2026-07-08pendingQ3 window check-in (75%)
2026-07-01 → 2026-08-31pendingQ1-Q2 2026 quarterly HBM shipment data confirms 32B Gb annual run-rate
How: SK Hynix + Samsung + Micron quarterly disclosures imply ≥30B Gb HBM annual run-rate (≥7.5B Gb Q1+Q2 average)
Source: Future Memory Storage — How AI Growth Will Drive HBM Demand Beyond 2025conf 70%
Notes: Direct measurement of headline 32B Gb claim via memory-maker quarterly shipment data.
2026-09-09pendingGlobal HBM demand will reach 32 billion Gb by 2026 — HBM is the absolute critical resource constraint in the AI compute supply chain.
2025-12-26hitAI consumes 20% of global DRAM wafer capacity in 2026
How: TrendForce/IDC reports confirm AI workloads consume ≥20% of global DRAM wafer capacity in 2026 with HBM and GDDR7 leading
Source: TrendForce — AI to Consume 20% of Global DRAM Wafer Capacity in 2026conf 95%
2026-10-01 → 2027-03-31pendingHBM market revenue surges to ~$51 billion in 2026 from ~$3B in 2023
How: Total HBM market revenue ≥$45B in 2026 per industry analyst consensus (TrendForce, IDC, Morgan Stanley)
Source: Morgan Stanley / Tech-Insider — Memory Chip Shortage 2026: HBM Takes 23% of DRAM Wafersconf 85%
Notes: 32B Gb is approximately equivalent to 4TB total HBM stack delivery to AI accelerator market — calibrates to ~$51B revenue range.
2026-12-31pendingHBM supply fulfillment rate falls to ~2% in 2026 — extreme constraint
How: HBM supply fulfillment rate (production capacity, yields, utilization) measured ≤5% relative to demand in 2026
Source: NextPlatform — HBM Supply Curve Gets Steeper, But Still Can't Meet Demandconf 85%
2026-06-01 → 2027-12-31pendingHBM critical-resource positioning drives semiconductor M&A or vertical integration
How: At least one major hyperscaler or chip designer announces HBM-focused M&A, JV, or capex commitment ≥$5B
Source: Industry consensus on HBM as binding constraintconf 55%

What if this resolves?

Clamp this prediction TRUE or FALSE and run a counterfactual Gibbs sample. Surfaces the predictions whose marginals shift most under that assumption.

(live posterior: 75%)

Click a button to clamp this prediction and run a Gibbs sample. Returns the predictions whose marginals shift most. ~30s per run; ideal for stress-testing "if X resolves, what else moves?"

Evidence chain

Every probability update with full Bayesian provenance — chronological, latest first

LBP2026-05-24T02:00:02Z75.1%-2.0pp

Network propagation: 77.1% → 75.1%

4-iter LBP, residual 0.01000 · damping 0.5, w_intrinsic 0.5 · method lbp_v3 · run 806b02f8

intake_event_update2026-05-21T23:15:16Z77.1%+7.9pp

intake:7afeeb9a-f217-4dd2-b910-24ff14bdfc39 bayesian_v2 inside=0.771 blend=0.771 LLR=0.404 κ=0.58 no_blend

Raw metadata

{
  "trf": 0.6127227050582519,
  "kappa": 0.5833,
  "base_rate": null,
  "predictor": "Morgan Stanley",
  "total_llr": 0.6931471805599453,
  "bayesian_v2": true,
  "prior_logit": 0.8117762442493454,
  "bayes_factor": "1.5:1 favoring",
  "blend_reason": "no reference_class linked",
  "inside_prior": 0.6924878819697828,
  "kappa_source": "predictor_table",
  "blend_applied": false,
  "contributions": [
    {
      "llr": 0.6931471805599453,
      "kappa": 0.5833,
      "label": "Demand for HBM at NVIDIA + Stargate level alone supports the 32B Gb forecast.",
      "adjusted_llr": 0.4043127504206161
    }
  ],
  "evidence_kind": "intake_event_update",
  "inside_source": "history_v2",
  "inside_weight": 1,
  "outside_weight": 0,
  "posterior_prob": 0.7713745528923279,
  "evidence_origin": "daily_intake",
  "llm_suggestions": [
    {
      "polarity": "corroborates",
      "status_change": "unchanged",
      "evidence_strength": "moderate",
      "delta_prob_suggestion": 0.05
    }
  ],
  "posterior_logit": 1.216088994669962,
  "predictor_brier": 0.01,
  "evidence_doc_ids": [],
  "inside_posterior": 0.7713745528923279,
  "blended_posterior": 0.7713745528923279,
  "reference_class_id": null,
  "total_adjusted_llr": 0.4043127504206161,
  "predictor_n_resolved": 1
}

LBP2026-05-17T02:00:01Z69.2%+3.5pp

Network propagation: 65.7% → 69.2%

5-iter LBP, residual 0.00689 · damping 0.5, w_intrinsic 0.5 · method lbp_v3 · run e607fa96

metadata_milestone_miss_sweep2026-05-14T22:14:31Z65.7%-5.1pp

metadata_milestone_miss_sweep bayesian_v2 n=1 inside=0.657 blend=0.657 LLR=-0.237 κ=0.58 no_blend

Raw metadata

{
  "trf": 0.6320693458696455,
  "kappa": 0.5833,
  "base_rate": null,
  "predictor": "Morgan Stanley",
  "total_llr": -0.4054651081081644,
  "grace_days": 7,
  "bayesian_v2": true,
  "prior_logit": 0.88791260307383,
  "bayes_factor": "1.3:1 against",
  "blend_reason": "no reference_class linked",
  "inside_prior": 0.7084592193620118,
  "kappa_source": "predictor_table",
  "n_milestones": 1,
  "blend_applied": false,
  "contributions": [
    {
      "llr": -0.4054651081081644,
      "kind": "quartile_checkpoint",
      "kappa": 0.5833,
      "label": "Q2 window check-in (50%)",
      "weight": 0.05,
      "strength": "weak",
      "confidence": null,
      "source_url": null,
      "adjusted_llr": -0.2365077975594923,
      "expected_date": "2026-05-06",
      "measurement_criterion": null
    }
  ],
  "evidence_kind": "metadata_milestone_miss_sweep",
  "inside_source": "history_v2",
  "inside_weight": 0.5575514578912482,
  "outside_weight": 0.44244854210875184,
  "posterior_prob": 0.6573269625234701,
  "posterior_logit": 0.6514048055143378,
  "predictor_brier": 0.01,
  "inside_posterior": 0.6573269625234701,
  "blended_posterior": 0.6573269625234701,
  "reference_class_id": null,
  "total_adjusted_llr": -0.2365077975594923,
  "predictor_n_resolved": 1
}

metadata_milestone_miss_sweep2026-05-02T22:07:21Z70.8%-4.6pp

metadata_milestone_miss_sweep bayesian_v2 n=1 inside=0.708 blend=0.708 LLR=-0.237 κ=0.58 no_blend

Raw metadata

{
  "trf": 0.6650500679109432,
  "kappa": 0.5833,
  "base_rate": null,
  "predictor": "Morgan Stanley",
  "total_llr": -0.4054651081081644,
  "grace_days": 7,
  "bayesian_v2": true,
  "prior_logit": 1.1244204006333227,
  "bayes_factor": "1.3:1 against",
  "blend_reason": "no reference_class linked",
  "inside_prior": 0.7548077345060716,
  "kappa_source": "predictor_table",
  "n_milestones": 1,
  "blend_applied": false,
  "contributions": [
    {
      "llr": -0.4054651081081644,
      "kind": "quartile_checkpoint",
      "kappa": 0.5833,
      "label": "Q1 window check-in (25%)",
      "weight": 0.05,
      "strength": "weak",
      "confidence": null,
      "source_url": null,
      "adjusted_llr": -0.2365077975594923,
      "expected_date": "2026-03-04",
      "measurement_criterion": null
    }
  ],
  "evidence_kind": "metadata_milestone_miss_sweep",
  "inside_source": "history_v2",
  "inside_weight": 0.5344649524623397,
  "outside_weight": 0.4655350475376603,
  "posterior_prob": 0.7084592193620118,
  "posterior_logit": 0.8879126030738305,
  "predictor_brier": 0.01,
  "inside_posterior": 0.7084592193620118,
  "blended_posterior": 0.7084592193620118,
  "reference_class_id": null,
  "total_adjusted_llr": -0.2365077975594923,
  "predictor_n_resolved": 1
}

LBP2026-04-30T16:39:51Z75.5%-2.4pp

Network propagation: 77.9% → 75.5%

5-iter LBP, residual 0.00825 · damping 0.5, w_intrinsic 0.5 · method lbp_v2 · run 0c8a4ea3

LBP2026-04-30T02:18:57Z77.9%-4.1pp

Network propagation: 82.0% → 77.9%

5-iter LBP, residual 0.00825 · damping 0.5, w_intrinsic 0.5 · method lbp_v1 · run 592311ef

Network propagation neighbors

Top edges sorted by latest LBP cross-impact

All propagation →

Top incoming (parents)

Edges that influence THIS node's belief

Kind	Node	Their prob	P(c\|s=T)	P(c\|s=F)	Δ implied
killer	TK02 AI Compute Supply Shock (TSMC/Taiwan Disruption)	12.0%	0.050	0.820	-0.024

Top outgoing (children)

Predictions THIS node influences

No outgoing edges.

Ticker exposure

6 ticker(s) linked

Beneficiaries (5)

MU NVDA AMD ATEYY STX

Prerequisites (1)

Predictions that must hit first

Type	Pred	Title	Domain	Lag
killer	TK02	AI Compute Supply Shock (TSMC/Taiwan Disruption)	—	—

Dependents (0)

Predictions enabled by this

Type	Pred	Title	Domain	Lag
No dependents

Linked documents (4)

Auto-generated by cosine similarity from Polymarket / Manifold / EDGAR / GDELT

Sim	Source	Title	Market prob	Polarity	Reviewed	Published
0.623	gdelt	microsoft q3 results top expectations driven by intelligent cloud strength 1091450.html	—	mentions	pending	2026-04-30
0.546	edgar_8k	Gitlab Inc. (GTLB) (CIK 0001653482)	—	mentions	pending	2026-05-11
0.546	edgar_8k	Gitlab Inc. (GTLB) (CIK 0001653482)	—	mentions	pending	2026-05-19
0.546	edgar_8k	Gitlab Inc. (GTLB) (CIK 0001653482)	—	mentions	pending	2026-06-02

Raw metadata

From Thesis_Timeline_v1.0_FINAL workbook

{
  "nia": false,
  "qty": "32B Gb",
  "mode": "FORECAST",
  "role": "Cited-Firm",
  "context": "MS 2026 Asia semi report identifies HBM as the defining chokepoint across entire AI factory stack.",
  "to_year": 2026,
  "cited_by": "Synthesis report",
  "conv_cues": "specific quantitative; MS research",
  "direction": "NUMERIC_TARGET",
  "from_year": 2026,
  "timeframe": "by 2026",
  "conv_level": "HIGH",
  "milestones": [
    {
      "kind": "llm_pre_event",
      "label": "Micron HBM capacity sold out through 2026 confirmed",
      "source": "CNBC — AI memory is sold out, causing an unprecedented surge in prices (Jan 2026)",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -5,
      "source_id": null,
      "confidence": 0.99,
      "source_url": "https://www.cnbc.com/2026/01/10/micron-ai-memory-shortage-hbm-nvidia-samsung.html",
      "expected_date": "2026-01-10",
      "observed_date": "2026-01-10",
      "research_origin": "deep_research",
      "measurement_criterion": "Micron publicly confirms HBM capacity fully sold out for 2026 calendar year"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q1 window check-in (25%)",
      "status": "overdue",
      "weight": 0.05,
      "ordinal": -4,
      "source_id": null,
      "expected_date": "2026-03-04",
      "observed_date": null,
      "miss_emitted_at": "2026-05-02T22:07:21.384228+00:00",
      "miss_emitted_by": "metadata_milestone_sweep"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q2 window check-in (50%)",
      "status": "overdue",
      "weight": 0.05,
      "ordinal": -3,
      "source_id": null,
      "expected_date": "2026-05-06",
      "observed_date": null,
      "miss_emitted_at": "2026-05-14T22:14:31.900138+00:00",
      "miss_emitted_by": "metadata_milestone_sweep"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q3 window check-in (75%)",
      "status": "pending",
      "weight": 0.05,
      "ordinal": -2,
      "source_id": null,
      "expected_date": "2026-07-08",
      "observed_date": null
    },
    {
      "kind": "llm_pre_event",
      "label": "Q1-Q2 2026 quarterly HBM shipment data confirms 32B Gb annual run-rate",
      "notes": "Direct measurement of headline 32B Gb claim via memory-maker quarterly shipment data.",
      "source": "Future Memory Storage — How AI Growth Will Drive HBM Demand Beyond 2025",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -1,
      "source_id": null,
      "confidence": 0.7,
      "source_url": "https://files.futurememorystorage.com/proceedings/2025/20250805_BMKT-102-1_Ellie-Wang.pdf",
      "expected_date": "2026-07-31",
      "research_origin": "deep_research",
      "expected_date_range": {
        "to": "2026-08-31",
        "from": "2026-07-01"
      },
      "measurement_criterion": "SK Hynix + Samsung + Micron quarterly disclosures imply ≥30B Gb HBM annual run-rate (≥7.5B Gb Q1+Q2 average)"
    },
    {
      "kind": "event",
      "label": "Global HBM demand will reach 32 billion Gb by 2026 — HBM is the absolute critical resource constraint in the AI compute supply chain.",
      "status": "pending",
      "weight": 1,
      "ordinal": 0,
      "source_id": "CMQ_033",
      "expected_date": "2026-09-09",
      "observed_date": null
    },
    {
      "kind": "llm_pre_event",
      "label": "AI consumes 20% of global DRAM wafer capacity in 2026",
      "source": "TrendForce — AI to Consume 20% of Global DRAM Wafer Capacity in 2026",
      "status": "hit",
      "weight": 0.4,
      "ordinal": 1,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://www.trendforce.com/news/2025/12/26/news-ai-reportedly-to-consume-20-of-global-dram-wafer-capacity-in-2026-hbm-gddr7-lead-demand/",
      "expected_date": "2026-12-26",
      "observed_date": "2025-12-26",
      "research_origin": "deep_research",
      "measurement_criterion": "TrendForce/IDC reports confirm AI workloads consume ≥20% of global DRAM wafer c
... (truncated)