The Silk - Core Forecast

AMD: Short AMD or reduce long exposure. Per-trade probability: 77% over 4-day holding period [n=1686]. Over 30 days (~6 independent 4-day windows), repeated mean-reversion trades have compounding edge.

NVDA: Short NVDA or hedge long positions. Asymmetric risk: -12% vs +4% expected range over 30 days.

ALB (Lithium proxy): Short ALB or take profits on lithium longs. Skewed to downside: -20% vs +5% expected range.

USD/CNY: Short USD/CNY. Moderate conviction — WATCH level with momentum continuation bias.

AI/semiconductor mean reversion from CRITICAL extremes: NVDA (+3.38σ), AMD (+3.08σ), and TSMC (+1.95σ) above 30-day means after 20-60% 30d rallies. Calibrated mean reversion rate within 6 days is 77% [n=1686]. Over 30 days, multiple independent 4-day reversion trades likely. Expect 5-15% pullback in these names from current levels.

Lithium/battery metals breakout sustainability: Lithium +29.6% 30d with ALB at +3.79σ (highest CRITICAL signal). While mean reversion base rate is 77% [n=1686] for the near-term, the underlying demand catalyst (grid storage buildout, EV adoption) provides fundamental support. Near-term pullback likely but 30-day net direction may remain positive if demand signals persist. LJ3 bear case: supply response from Chinese producers could cap upside.

A procurement timing platform for battery manufacturers and grid storage companies that uses statistical mean-reversion signals on lithium (and related commodity) prices to generate automated buy/hold/defer recommendations for raw material purchasing. Like a harmonic resolution in music—where a sustained dissonant chord creates predictable tension that resolves to the tonic—the platform identifies when extreme price deviations (like the current +3.79σ breakout) are approaching their cadence point, enabling procurement teams to lock in forward contracts at optimal timing rather than reacting to spot prices. Revenue model: subscription SaaS + percentage of documented savings on timed purchases.

A real-time brokerage layer that sits between AI application developers and GPU cloud providers, automatically routing inference jobs to the cheapest available compute by exploiting the 'contrast' between spot pricing across providers (AWS, Azure, CoreWeave, Lambda, etc.) during the NVIDIA-driven capacity expansion. Like a painting that derives meaning from the juxtaposition of light and dark areas, the system derives margin from the price differentials (negative space between posted prices and clearing prices) that naturally emerge when supply is rapidly expanding unevenly across providers. Revenue model: takes a thin spread on every routed inference call.

A real-time monitoring SaaS platform that automatically detects AMD GPU compute capacity surges and price deviations across cloud providers, colocation facilities, and secondary hardware markets — then triggers automated procurement, rebalancing, or hedging actions for enterprise AI/ML teams. Like gravitational lensing reveals hidden mass in cosmology, the platform maps the 'dark matter' of distributed GPU availability (underutilized capacity, gray-market inventory, OEM pipeline commitments) that becomes visible only when a breakout like AMD's signals a demand-supply phase transition. Value is captured through subscription fees plus a percentage of cost savings from optimized GPU procurement timing.