The Silk - Core Forecast

AMD: Reduce AMD long exposure or hedge via put spreads. The 4-day mean reversion probability is 77% [n=1686], but the 30-day path involves multiple independent 4-day windows where edge compounds through repeated trials. Avoid outright shorts given poor short signal calibration (40% [n=93]).

NVDA: Trim NVDA longs or sell covered calls. Mean reversion probability 72% (adjusted from 77% base rate). Risk/reward favors patience over chasing at these levels.

Lithium (ALB proxy): Maintain lithium/ALB exposure with tight stops below $190. Momentum continuation probability 58% [n=128] over 4-day windows. Multiple independent trials over 30 days compound the edge if trend persists.

Crude Oil: Tactical long crude oil for mean reversion bounce. 60-65% probability of recovery to $95-100 over 30 days [uncalibrated]. Risk management critical given strong downtrend.

Gold: Hold gold as portfolio hedge. 55% probability of maintaining above $4,650 over 30 days [uncalibrated]. Asymmetric upside if geopolitical risk escalates beyond current stable regime.

AI/Semiconductor Mean Reversion at Statistical Extremes: AMD at +4.79σ and NVDA at +3.20σ above 30-day means represent the most extreme readings in the current signal set. Base rate for mean reversion within 6 days at 2σ+ is 77% [n=1686]. AMD's +92.4% 30d move is unsustainable at current velocity. Partial retracement of 15-25% in AMD and 8-12% in NVDA over 30 days is the base case. This would drag QQQ and SPY given concentration.

A real-time network map product that visualizes the cascade exposure of mid-cap and small-cap companies whose revenue, supply chain, or product roadmaps are structurally coupled to AMD's ecosystem (motherboard OEMs, firmware vendors, cooling solution providers, cloud instance resellers, etc.). Modeled on contour maps with 'elevation' representing concentration risk and 'watershed boundaries' showing which firms drain into AMD vs. NVIDIA vs. Intel dependency basins, the tool sells circuit-breaker alerts to portfolio managers and procurement teams when a node crosses a cascade-failure threshold — analogous to a topographic flood model where a breakout at the peak predicts which downstream valleys get inundated first.

A real-time monitoring and optimization SaaS platform that visualizes GPU cluster utilization across enterprise AI workloads using art-inspired compositional principles — negative space shows idle capacity, contrast highlights bottlenecks, and balance metrics guide workload scheduling. As NVIDIA's breakout drives massive GPU procurement, enterprises are spinning up clusters faster than they can efficiently operate them, creating a predictable gap between hardware acquisition and operational maturity. This tool is the obvious next step: a Datadog-style observability layer purpose-built for GPU-dense AI infrastructure, with a visual language that makes complex multi-node utilization patterns immediately legible to non-specialist decision-makers.

A screening and capital-allocation SaaS for smalland mid-cap investors that identifies 'new niche' companies likely to absorb demand, talent, and supplier capacity spilling over from a sustained Russell 2000 breakout. It combines post-breakout factor rotation signals with ecological niche-filling models to rank overlooked suppliers, distributors, and service firms, then monetizes via subscription fees, model licensing, and premium portfolio-integration APIs.