# Feedstock Blending Optimizer

*/Opportunities/Feedstock_Blending_Optimizer*

## Opportunity Overview

**Wedge**: The initial beachhead isolates independent renewable diesel and sustainable aviation fuel plants in North America facing acute margin compression from variable organic feedstocks. Winning this niche delivers immediate, measurable yield improvements by maximizing cheaper, low-grade inputs without violating final product specifications. Expansion proceeds sequentially to global biodiesel facilities, followed by integration into traditional petroleum refineries blending complex crude slates.
**Timing**: The deployment of inline near-infrared sensors across tank farms now produces sub-second physical property data instead of batch-delayed lab results. Concurrently, non-linear neural solvers process these multi-variable constraints fast enough to execute real-time closed-loop control, replacing offline daily schedules.
**Why This I C P**: Renewable diesel and sustainable aviation fuel producers face extreme feedstock variability, processing differing grades of tallow, soy, and used cooking oil. This continuous physical volatility destroys their margins under static planning, forcing them to adopt real-time optimization faster than traditional crude refiners with highly stable inputs.
**Size Of Prize**: Globally, roughly 3,800 complex petroleum refineries and advanced biofuel plants spend an average of $400k annually on legacy planning software and manual blend optimization consulting. Multiplying these 3,800 addressable facilities by the $400k software and service replacement cost yields an annual recurring prize of approximately $1.5B.
**Gap Narrative**: Refineries and biofuel plants rely on static, linear programming models to determine feedstock blends, leaving margin on the table when raw material qualities fluctuate intra-day. Plant operators require dynamic, non-linear optimization that adjusts blend ratios in real-time based on incoming inline sensor data to hit product specifications at the absolute lowest input cost. Existing linear tools fail to reconcile live spot market pricing with continuous physical variations in the pipes.
**Defensibility**: The product builds defensibility through extreme workflow lock-in at the plant control layer. Once the software continuously writes blend setpoints to the distributed control system, removing it requires a reversion to manual, margin-destroying operator control. The non-linear models also compound in accuracy as they ingest specific plant operating data over time, creating a predictive capability that generic off-the-shelf solvers cannot replicate.
**Why This Thesis**: Deploying as a Software layer that sits atop the existing Distributed Control System directly matches the security and latency demands of plant engineers. It acts as an always-on supervisory controller, writing setpoints directly to valve controllers rather than generating offline reports that operators must manually input.

## Opportunity Linked Thesis

**Thesis**: [Software](/Theses/Software)

## Opportunity Linked I C P

**Icp**: [Petrochemical Refinery](/CompanyTypes/Petrochemical_Refinery)

## Opportunity Market Sizing

_Illustrative — target and order-of-magnitude estimate figures, not an achieved track record (this Thing is concept-stage)._

**S A M**: ~$200M-300M for North American and European advanced process facilities
**S O M**: ~$10M-25M
**T A M**: ~3,000 global refining and petrochemical facilities × ~$200k-300k/yr ≈ ~$600M-900M
**Growth Rate**: ~8-12%/yr, driven by feedstock price volatility and the operational complexity of introducing renewable blending components
**Paid Comparable Spend**: ~$150k-300k/yr per facility on legacy linear programming software licenses and manual blending engineer analysis

## Opportunity Incumbents

- [Aspen PIMS](/Products/Aspen_PIMS) — Tool
- [AVEVA Supply Chain](/Products/AVEVA_Supply_Chain) — Tool
- [Honeywell Forge Blending](/Products/Honeywell_Forge_Blending) — Tool
- [Excel Solver Models](/Products/Excel_Solver_Models) — Spreadsheet
- [Engineering Consultants](/Products/Engineering_Consultants) — Service
- [In-House Python Scripts](/Products/In-House_Python_Scripts) — DIY

## Opportunity Win Conditions

**Kill Thresholds**:
- Time to deploy and map facility constraints > 30 days
- Engineer override rate > 50% after week 3 of live deployment
- Demonstrated margin improvement < $0.05 per barrel in 30 days
- Pilot conversion rate to $150k ARR contracts < 20% after 90 days
**Leading Metrics**:
- Days to first executed blend recommendation
- Percentage of automated blend ratios approved without manual engineer override
- Calculated margin improvement per blended barrel in USD
- Historian data ingestion and sync latency in seconds
**What Proves Right**: Refinery process engineers adopt the optimizer to generate daily run plans, consistently executing its blend ratios instead of relying on legacy Excel solver models. The software identifies feasible blends using lower-cost or alternative feedstocks while strictly maintaining product specifications like octane and sulfur limits. Facilities deploy the system and measure direct reduction in raw material costs per barrel within the first 30 days.
**What Proves Wrong**: Deployment stalls because mapping the specific tank telemetry and piping constraints of a facility requires months of custom engineering rather than days of configuration. Blending engineers consistently override the optimizer recommendations due to unmodeled physical constraints or process safety risks. The cost and timeline to integrate with legacy historians exceed the projected revenue of the software contract.

## Opportunity Build Profile

**Hardest Part**: Accurately modeling the non-linear physical and chemical interaction effects of varying feedstocks to guarantee output specifications without over-consuming expensive premium inputs.
**Min Viable Scope**: Focus strictly on day-ahead, offline recipe generation for a single product category using static lab results. Exclude closed-loop control system integration, real-time adjustments, and multi-facility procurement routing.
**Cold Start Problem**: The system requires deep historical lab assay and yield data to train baseline models before suggesting safe operational changes. Break this by ingesting 12 months of historical batch data from a single plant design partner to run backtests that prove margin improvements offline.
**Time To First Value**: 3–4 weeks of historical data ingestion and shadow-mode validation to generate the first trusted recipe
**Data Moat Available**: true
**Technical Difficulty**: High

## Neighborhood

### Where the gap lives

- [Chemical refineries](/Customers/Chemical_refineries) — latent gap · Customers

### Incumbent in

- [In-House Python Scripts](/Products/In-House_Python_Scripts) — incumbent in · Products
- [Excel Solver Models](/Products/Excel_Solver_Models) — incumbent in · Products
- [Honeywell Forge Blending](/Products/Honeywell_Forge_Blending) — incumbent in · Products
- [AVEVA Supply Chain](/Products/AVEVA_Supply_Chain) — incumbent in · Products
- [Aspen PIMS](/Products/Aspen_PIMS) — incumbent in · Products
- [Engineering Consultants](/Products/Engineering_Consultants) — incumbent in · Products

### Applies thesis

- [Petrochemical Refinery](/CompanyTypes/Petrochemical_Refinery) — applies thesis · CompanyTypes

### Embodies

- [Software](/Theses/Software) — embodies · Theses

### Similar Opportunities

- [Predictive Marine Fuel Mixing](/Opportunities/Predictive_Marine_Fuel_Mixing) — similar · Opportunities
- [Distillation Yield Engine](/Opportunities/Distillation_Yield_Engine) — similar · Opportunities
- [Refinery Dynamic Deadband Adjustment](/Opportunities/Refinery_Dynamic_Deadband_Adjustment) — similar · Opportunities
- [AI Yield Optimization for Oleochemical Plants](/Opportunities/AI_Yield_Optimization_for_Oleochemical_Plants) — similar · Opportunities
- [Feedstock Sourcing API](/Opportunities/Feedstock_Sourcing_API) — similar · Opportunities
- [Refiner Energy Optimizer](/Opportunities/Refiner_Energy_Optimizer) — similar · Opportunities
- [Autonomous Loop Tuning](/Opportunities/Autonomous_Loop_Tuning) — similar · Opportunities
- [Algorithmic Refiner Energy Optimization](/Opportunities/Algorithmic_Refiner_Energy_Optimization) — similar · Opportunities
- [Combustion Optimization API](/Opportunities/Combustion_Optimization_API) — similar · Opportunities
- [Refiner Tuning Agent](/Opportunities/Refiner_Tuning_Agent) — similar · Opportunities
- [Algorithmic Mineral Blending](/Opportunities/Algorithmic_Mineral_Blending) — similar · Opportunities
- [Yield Setpoint API](/Opportunities/Yield_Setpoint_API) — similar · Opportunities
- [AI Scale-Up for Chemical Producers](/Opportunities/AI_Scale-Up_for_Chemical_Producers) — similar · Opportunities
- [Fuel Deployment Dispatcher](/Opportunities/Fuel_Deployment_Dispatcher) — similar · Opportunities
- [Cracking Unit Diagnostics](/Opportunities/Cracking_Unit_Diagnostics) — similar · Opportunities
- [Real-Time Kiln Optimization](/Opportunities/Real-Time_Kiln_Optimization) — similar · Opportunities
- [Dynamic Recipe Tuning for Biomanufacturing](/Opportunities/Dynamic_Recipe_Tuning_for_Biomanufacturing) — similar · Opportunities
- [Generative Shift Handovers for Refineries](/Opportunities/Generative_Shift_Handovers_for_Refineries) — similar · Opportunities
- [Thermal Energy Telemetry for Mills](/Opportunities/Thermal_Energy_Telemetry_for_Mills) — similar · Opportunities
- [Hog Fuel Blending Agent](/Opportunities/Hog_Fuel_Blending_Agent) — similar · Opportunities