Procurement decisions, quantified in seconds
CausalIQ is a decision intelligence platform for AI switch silicon procurement. Real-time inventory tracking, Monte Carlo risk simulation, Nash equilibrium bidding, and a natural language interface — all in one API.
REQUEST
{
"query": "Should we pay the TH6 premium?"
}
RESPONSE · 310ms
"answer": "Yes. TH6 on-hand (450) is below
safety stock (600). Paying the
12% premium yields 96% prob
of margin > 20%."
}
Silicon procurement runs on gut feel and spreadsheets
AI infrastructure demand is volatile, lead times are 26–52 weeks, and foundry supply is concentrated. One bad procurement call can mean $50M in stranded inventory — or a missed quarter.
No probabilistic view of risk
Point forecasts don't capture the fat tails. A single demand spike or DRAM cost shock wipes out margin — and no one saw a distribution.
Foundry allocation is a game
AWS, Azure, and Meta are bidding for the same wafers. Without knowing competitor strategies, you're playing poker blind.
Analysis takes days, not seconds
By the time procurement models a scenario in Excel, the allocation window has closed and a competitor has locked up supply.
Three steps from data to decision
Connect your supply chain data
Push inventory positions and real-time cluster telemetry via REST API. The system tracks on-hand units, safety stock, in-transit orders, and reorder points per silicon generation.
// 400">"text-violet-400">POST /api/v1/telemetry/ingest
{
400">"switch_id": 400">"sjc-spine-07",
400">"silicon_gen": 400">"TH6",
400">"ecn_marks_per_sec": 1842,
400">"rdma_util_pct": 73.4,
400">"port_util_pct": 88.1
}Run stochastic risk scenarios
Monte Carlo engine simulates gross margin outcomes across thousands of demand and cost scenarios. Nash equilibrium solver computes optimal bid prices against competitor strategies.
// 400">"text-violet-400">POST /api/v1/simulation/monte-carlo
// Response
{
400">"mean_gross_margin_pct": 27.2,
400">"p5": 20.4,
400">"p95": 34.1,
400">"probability_above_20pct": 0.96,
400">"var_95_pct": 20.4
}Act on quantified recommendations
Natural language query engine synthesizes inventory state, simulation outputs, and Nash bids into a single, specific procurement recommendation — in seconds.
// 400">"text-violet-400">POST /api/v1/query/nl
{ 400">"query": "Should we pay the TH6 spot premium?" }
// Response
{
400">"answer": "Yes. TH6 on-hand (450 units) is below
safety stock (600). With 96% probability of margin
>20%, paying the 12% premium is optimal. Recommend
aggressive bid at 400">$21,840 to beat Azure allocation."
}Probabilities, not predictions
Rather than a single forecast that will be wrong, CausalIQ runs 1,000+ stochastic simulations and returns the full margin distribution. Know the 5th percentile, the median, and the upside — before you commit capital.
Monte Carlo — Gross Margin Distribution (n=1,000)
Everything procurement needs
Live Silicon Position Tracking
Monitor on-hand units, in-transit stock, and safety-stock coverage across every silicon generation and warehouse in real time.
< 1s latencyStochastic Margin Simulation
Run thousands of simulations across DRAM cost shocks, demand variability, and supply constraints. Get a full probability distribution — not a point estimate.
1,000+ scenarios/secAsk in Plain English
AI answers procurement questions against live inventory, Nash equilibrium bids, and shortfall data — no SQL, no dashboards.
AI poweredCompetitive Bid Optimization
Model the foundry allocation market as a multi-player game. Compute the optimal bid price against AWS, Azure, and Meta before you submit.
Multi-player game theoryAutomatic Risk Escalation
The system flags HIGH and CRITICAL shortfalls the moment on-hand drops below safety stock. No manual threshold tuning required.
AUTO-escalatedDemand Signals from the Network
Ingest real-time ECN marks, RDMA counters, and cluster utilization from AI switches. Demand forecasts update from actual infrastructure data, not just POs.
SQS-backed async pipelineRunning right now at api.netcausal.ai
No sandbox, no mock data. Every endpoint below hits the live system — backed by Aurora PostgreSQL, ECS Fargate, and AI.
Real-time supply risk across all silicon generations
400">"text-violet-400">curl https://api.netcausal.ai/api/v1/inventory/shortfall-alerts{
400">"alerts": [
{
400">"silicon_gen": 400">"TH7",
400">"on_hand_units": 80,
400">"safety_stock_units": 200,
400">"shortfall_units": 120,
400">"severity": 400">"CRITICAL"
},
{
400">"silicon_gen": 400">"TH6",
400">"shortfall_units": 150,
400">"severity": 400">"HIGH"
}
],
400">"count": 2
}Ask a procurement question in plain English
400">"text-violet-400">curl -X 400">"text-violet-400">POST https://api.netcausal.ai/api/v1/query/nl \
-H "400">"text-violet-400">Content-Type: application/json" \
-d '{400">"query": "What is the TH6 stockout risk?"}'{
400">"answer": "TH6 has 450 units on-hand vs 600
safety stock — a 150-unit shortfall. With
800 on-order, cover arrives in ~23 days.
Recommend expediting 200 units immediately."
}Quantify margin risk before committing
400">"text-violet-400">curl -X 400">"text-violet-400">POST https://api.netcausal.ai/api/v1/simulation/monte-carlo \
-H "400">"text-violet-400">Content-Type: application/json" \
-d '{400">"silicon_gen":400">"TH6",400">"pay_premium":true,400">"n_simulations":1000}'{
400">"mean_gross_margin_pct": 27.2,
400">"p5": 20.4,
400">"p95": 34.1,
400">"probability_above_20pct": 0.96
}Fully deployed on AWS — not a prototype
Try it in 30 seconds
The API is live. No sign-up, no API key — hit the endpoints below directly or explore the interactive docs.
# Check live supply shortfalls right now
400">"text-violet-400">curl https://api.netcausal.ai/api/v1/inventory/shortfall-alerts
# Ask a procurement question
400">"text-violet-400">curl -X 400">"text-violet-400">POST https://api.netcausal.ai/api/v1/query/nl \
-H "400">"text-violet-400">Content-Type: application/json" \
-d '{400">"query": "What is the TH7 stockout risk this quarter?"}'