What if a global treaty halts frontier AI training?
A frontier-training moratorium pulls the AI-capex bid: NVDA leads lower as the order-book bellwether, dragging HBM (Micron), custom silicon (AVGO) and the broad semi complex. Direct rhyme is the Jan-2025 DeepSeek shock and Oct-2024 ASML bookings miss, which repriced chip demand 10-15% in days. Skeptic's note: a treaty is slow and leaky, so the equity hit is a sentiment derating, not a fundamentals collapse; the real risk is multiple compression on the capex names, with inference demand a partial offset.
how we built this number — every step
The class rate is measured from our dated, sourced event library (decade-normalized Poisson — the full table is public at base_rates.json). The variant’s share within its class is the analyst’s editorial call, published so you can audit it. A wider range means thinner precedent. Full recipe: methodology · scored at Reality Check.
The butterfly cascade
How this trigger trickles across markets, left → right — the root shock, its first‑order moves, then the ripple effects. Drag any node; tap a market for its real price history.
Resolution timeline — how this probability is moving
Our model's odds (gold) over time vs the crowd's (Polymarket, blue), from the past toward the 1–3 years horizon. Each dot is a real macro event that nudged the probability — green pushed it up, red pushed it down. Tap a dot for the source. The gold path is an illustrative reconstruction anchored to today's estimate — real dated events, not a live re-estimate history.
What it would mean
If this plays out, it is a mixed shock. A global treaty/moratorium halts frontier AI model training. The trigger decomposes into signed root‑shocks — AI capex ▼ — which propagate through our causal graph to the markets below.
If it happens — the markets it would move
Biggest moves first. Projected moves are cascade-model priors; hist A–B% = what comparable past events actually did (measured abnormal returns), and model prior · unmeasured marks markets with no analogue backing yet. Tap any market for its price history.
| Market | Class | Projected move | |
|---|---|---|---|
| 1 | Nvidia NVDAon Hyperliquid 📈 chart | Equity | ▼ -1.9% hist -1.16–-0.48% · other way +4.19% (n=12) |
| 2 | Broadcom AVGOon Hyperliquid 📈 chart | Equity | ▼ -1.4% hist -1.3–+0.4% · other way +5.73% (n=12) |
| 3 | Micron MUon Hyperliquid 📈 chart | Equity | ▼ -1.3% hist -4.1–+0.37% · other way -1.63% (n=12) |
| 4 | Semiconductors SMHon Hyperliquid 📈 chart | Equity | ▼ -0.9% hist -0.63–-0.04% · other way +1.37% (n=12) |
| 5 | AMD AMDon Hyperliquid 📈 chart | Equity | ▼ -0.6% hist -1.19–+0.11% · other way -3.31% (n=12) |
| 6 | TSMC TSMon Hyperliquid 📈 chart | Equity | ▼ -0.6% hist -0.58–+0.19% · other way +1.86% (n=12) |
| 7 | Marvell MRVLon Hyperliquid 📈 chart | Equity | ▼ -0.6% hist -0.79–+0.87% · other way +0.78% (n=12) |
| 8 | ASML ASMLon Hyperliquid 📈 chart | Equity | ▼ -0.5% hist -2.61–+0.95% · other way -0.21% (n=12) |
| 9 | Qualcomm QCOMon Hyperliquid 📈 chart | Equity | ▼ -0.5% hist -2.57–+0.78% · other way -2.95% (n=12) |
| 10 | Intel INTCon Hyperliquid 📈 chart | Equity | ▼ -0.4% hist -5.17–+1.18% · other way -2.72% (n=12) |
Probable recommendation
Historical precedent — what analogous events actually did
Across 40 analogous events (overlap‑weighted), as abnormal returns — market beta stripped, so it's the event's own effect, not the market backdrop. Shown at 20 days (persistent) and 5 days (immediate); ↺ fades = the two horizons disagree. Confidence = consistency × sample × significance.
| Asset | History says | Abnormal (20d · 5d) | Hit | n | Confidence | vs cascade |
|---|---|---|---|---|---|---|
| MU MU | SHORT | -3.0% · 5d -3.2% | 71% | 40 | 0.32 | ✓ matches cascade |
| INTC INTC | SHORT | -4.4% · 5d -2.7% | 65% | 40 | 0.30 | ✓ matches cascade |
| NVDA NVDA | SHORT | -0.2% · 5d -3.1% | 63% | 40 | 0.19 | ✓ matches cascade |
| Gold XAU | LONG | +0.5% · 5d -0.2% ↺ fades | 61% | 40 | 0.19 | · |
| AMD AMD | SHORT | -0.8% · 5d -2.2% | 61% | 40 | 0.17 | ✓ matches cascade |
| Bitcoin BTC | SHORT | -4.4% · 5d -3.1% | 59% | 40 | 0.14 | · |
| QCOM QCOM | SHORT | -2.2% · 5d -2.7% | 57% | 40 | 0.11 | ✓ matches cascade |
| US dollar DXY | SHORT | -0.0% · 5d +0.0% ↺ fades | 55% | 40 | 0.09 | · |
| SMH SMH | LONG | +0.3% · 5d -0.8% ↺ fades | 53% | 40 | 0.06 | ⚠ differs |
| MRVL MRVL | LONG | +1.2% · 5d -2.0% ↺ fades | 53% | 40 | 0.05 | ⚠ differs |
| 10y yield DGS10 | SHORT | -1bp · 5d +2bp ↺ fades | 53% | 40 | 0.05 | · |
| High-yield credit HYG | LONG | +0.2% · 5d +0.2% | 51% | 40 | 0.02 | · |
| AVGO AVGO | LONG | +1.1% · 5d -1.9% ↺ fades | 45% | 40 | 0.00 | ⚠ differs |
| TSM TSM | LONG | +0.5% · 5d -0.3% ↺ fades | 45% | 40 | 0.00 | ⚠ differs |
Why this probability
Binding global AI training moratorium is novel; race dynamics and US-China rivalry make treaty implausible short-term. A base‑rate‑anchored prior, continuously scored against what actually happens — not a forecast.