New ProductLive API

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.

96%
margin probability accuracy
310ms
NL query response time
3
silicon generations tracked live
api.netcausal.ai — live
connected

REQUEST

POST /api/v1/query/nl
{
"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%."

}
TH6 shortfall:150 units
Margin prob:96%
Optimal bid:$21,840
The Problem

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.

How It Works

Three steps from data to decision

01
Ingest

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.

REST APIAurora PostgreSQLSQS
telemetry.json
// 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
}
02
Simulate

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.

Monte CarloNash EquilibriumAI Engine
simulation.json
// 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
}
03
Decide

Act on quantified recommendations

Natural language query engine synthesizes inventory state, simulation outputs, and Nash bids into a single, specific procurement recommendation — in seconds.

Natural LanguagePrescriptiveExplainable
query.json
// 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."
}
Monte Carlo Engine

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.

Mean gross margin27.2%
Worst-case (P5)20.4%
P(margin > 20%)96%

Monte Carlo — Gross Margin Distribution (n=1,000)

4%
<15%
7%
15–20%
18%
20–25%
32%
25–30%
26%
30–35%
13%
>35%
Mean margin27.2%
P(margin > 20%)96%
Capabilities

Everything procurement needs

📡
Real-Time Inventory

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 latency
🎲
Monte Carlo Engine

Stochastic 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/sec
🧠
NL Query Engine

Ask in Plain English

AI answers procurement questions against live inventory, Nash equilibrium bids, and shortfall data — no SQL, no dashboards.

AI powered
⚖️
Nash Equilibrium

Competitive 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 theory
⚠️
Shortfall Alerts

Automatic Risk Escalation

The system flags HIGH and CRITICAL shortfalls the moment on-hand drops below safety stock. No manual threshold tuning required.

AUTO-escalated
🔌
Telemetry Ingestion

Demand 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 pipeline
Live API

Running 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.

Shortfall Alerts

Real-time supply risk across all silicon generations

request.sh
400">"text-violet-400">curl https://api.netcausal.ai/api/v1/inventory/shortfall-alerts
response.json
{
  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
}
NL Query

Ask a procurement question in plain English

request.sh
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?"}'
response.json
{
  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."
}
Monte Carlo

Quantify margin risk before committing

request.sh
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}'
response.json
{
  400">"mean_gross_margin_pct": 27.2,
  400">"p5": 20.4,
  400">"p95": 34.1,
  400">"probability_above_20pct": 0.96
}
Production Stack

Fully deployed on AWS — not a prototype

ECS FargateAurora PostgreSQLElastiCache RedisSQSAWS ALBCloudflareFastAPISQLAlchemyAI/NLP EngineNext.jsVercel
No auth required for demo

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.

quickstart.sh
# 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?"}'