Four independent scanners run against a proprietary supply chain knowledge graph to detect bottlenecks at every stage of their lifecycle — from architecture lock-ins and cross-sector demand stacking to capacity gap math and earnings language shifts — months to years before they surface in analyst reports.
Sector analysts model demand within their coverage. No single analyst aggregates demand across AI, defense, energy, automotive, and quantum simultaneously. When multiple sectors stack demand on the same input material, the gap between siloed forecasts and actual aggregate demand creates a window where supply chain-aware investors have an edge.
Semiconductor fabs take 2-4 years to build. Rare earth mines take 7-10 years to permit. But architecture decisions that lock in future demand happen in a single product announcement. The lag between demand commitment and supply response is where bottlenecks form — and where the math becomes visible to anyone tracking both sides.
The most asymmetric opportunities emerge when a small or mid-cap company controls a critical node in a supply chain that feeds into hundreds of billions in downstream value. These companies are often overlooked by large-cap analysts until the bottleneck becomes obvious — by then, the stock has already re-rated.
Each scanner operates on a completely different data source and methodology. The architecture scanner reads product announcements and derives bill-of-materials dependencies. The cross-demand scanner runs graph computations across sector demand models. The capacity gap scorer compares public CAGR data. The earnings tracker parses transcripts. When two or more of these independent methods flag the same material or component, it is unlikely to be noise — it is a structural signal that the bottleneck is advancing through its lifecycle.
Each scanner produces an independent risk score based on the severity of the bottleneck it detects. The combined score (0–100) shown on each card is a weighted blend of all scanners that flagged that item, normalized so the highest-risk item scores 100. Earlier-stage detections carry more weight — an architecture lock-in (Stage 1, 1.5x) carries more weight than an earnings language shift (Stage 4, 0.8x) because it represents a longer lead time and a larger information asymmetry. Items flagged by multiple scanners naturally score higher than single-scanner detections, which is why multi-signal items rise to the top of the watchlist.
| Name ▼ | Ticker ▼ | Stage ▼ | Scanners ▼ | Score ▼ | Origin ▼ | Sev ▼ | Thesis ▼ |
|---|
The prediction engine runs on top of ForcedAlpha's supply chain intelligence graph — a continuously updated map of 325 nodes and 598 edges spanning semiconductors, defense, energy, critical minerals, and robotics. Every node represents a material, component, company, or process. Every edge represents a dependency.
This graph structure is what makes cross-demand detection possible. When the architecture scanner identifies a new dependency (for example, co-packaged optics requiring CW DFB lasers), it checks the graph for existing nodes, maps upstream material requirements, and scores supply constraints based on geographic concentration, supplier count, and qualification timelines. Without the graph, each scanner would operate in isolation. With it, they share a common map of how supply chains actually connect.
The prediction engine combines public data — earnings transcripts, product announcements, capacity expansion filings, and industry CAGR estimates — with proprietary graph-based computation that maps how supply chains actually connect across sectors. The graph itself, the dependency mappings, the cross-sector demand aggregation, and the scoring methodology are what give the system its edge.
When a predicted bottleneck later appears in our Tier 1 alert system (the highest-conviction tier, requiring multi-source convergence and graph severity confirmation), it validates the prediction engine. Multiple items currently on the prediction watchlist have related tickers that were independently promoted to Tier 1 — the prediction engine flagged the supply constraint before the broader convergence data confirmed it. This is the feedback loop: predict early, validate later, calibrate continuously.
See structural supply chain risks before they appear in analyst reports. The earlier you detect a bottleneck, the more asymmetric the opportunity.
Start Free Trial