Industries / Food, farming and forestry / Food and beverage processing
Rendering
Rendering is the industry that collects slaughter by-products, dead farm animals, used cooking oil and unsold food and cooks them into fats and protein meals for animal feed, pet food, soap, chemicals and fuel.
1882
Ira C. Darling & Co. opens in Chicago as a branch of L. B. Darling & Company, a Rhode Island maker of bone fertilizer to which the meatpacker Gustavus Swift had turned for rendering; Darling Ingredients dates its founding to that year.
33
Milestones achieved across 6 domains.
9
To watch for: announced, not yet confirmed outside.
12
Open deliverables. Nobody has shipped them.
In one minute
Where it started
Rendering, cooking the fat out of animal scraps, supplied tallow for candles and soap for centuries. The modern trade grew up beside the Chicago meatpackers of the late 1800s, and its machines came in steps: Valerius Anderson's screw press in 1900, dry cookers from Germany in the 1920s and continuous cookers in the 1960s. After BSE (mad cow disease) spread through British cattle in meat and bone meal, laws from 1988 onward set what may be fed to which animals and how plants must cook.
Where it stands
By the North American Renderers Association's own count, US renderers turn about 56 billion pounds (25 million tonnes) of raw material a year into 22 billion pounds (10 million tonnes) of fats and proteins, and members of EFPRA, the European renderers' association, take in about 18 million tonnes a year. Fuel has become the biggest new buyer of fat: by USDA's count, animal fats and greases made up 37% of the feedstock for US biodiesel and renewable diesel in 2023. Smell and wastewater still take plants to court, and existing plants in the EU must meet the sector's first binding odor limits by December 2027.
What is open
Much of the work still rests on judgment and on tests that miss what matters. Operators decide by experience when a cook is done, the official pepsin test can pass a meal that animals digest poorly, nobody has followed plastic through a plant, and no public test shows which laboratory checks can tell used cooking oil from fresh palm oil. Public records are scattered: nobody tallies renderers' water permit breaches by pollutant, tracks how EU permits write in the new odor limits, or has tested software on sorting by-products into the EU's risk categories. Most of these jobs need a laboratory, public data and AI models that read documents, not new machines.
Domains
AchievedWatch forOpen
Domain
Progress
Last achieved
Watch for
Open
Collection
Collection is the work of picking up what renderers process: offal, fat and bones from slaughterhouses and butchers, animals that died on farms (fallen stock), used cooking oil and grease-trap waste from restaurants, and unsold food from shops and factories.
5 · 1 · 2
Last achievedFormer foods enter feed law2017
Watch forProving where oil was collected
2
openCooking and separation
Cooking and separation is the plant work of grinding raw by-products, cooking the water out of them, and splitting what is left into liquid fat and a solid that is pressed, dried and milled into meal.
6 · 2 · 2
Last achievedRendering's energy share measured2021
Watch forCookers run by learned control
2
openFats and oils
This is the work of grading and selling the fat that rendering yields, and the used cooking oil that collectors bring in, as tallow, greases and poultry fat for feed, pet food, soap, chemicals and fuel.
4 · 1 · 2
Last achievedOver a third from fats2023
Watch forQuick scan for cashew oil
2
openProtein meals
Protein meals are the dry, ground solids left after the fat is pressed out, such as meat and bone meal, poultry meal, feather meal and blood meal, sold for pet food, farm and fish feed, and fertilizer.
6 · 1 · 2
Last achievedPlastic counted in poultry meal2024
Watch forBritain moves to allow meal
2
openOdor and emissions
Odor and emissions work is keeping the smell, air pollution and wastewater of a rendering plant from reaching its neighbors and the rivers nearby.
6 · 2 · 2
Last achievedTighter water limits withdrawn2025
Watch forSouth St. Paul odor plan
2
openDisease control
Disease control is the set of laws and plant practices that stop infections in dead animals and food waste from returning to livestock and people through feed.
6 · 2 · 2
Last achievedChina bans swill for pigs2018
Watch forCanada eases its feed ban
2
open
Advances to build with
4 research notes

Claude Fable 5.2
Three working devices
Anthropic says its AI model Claude Fable 5.2 designed three prototype devices that worked without electrical modifications, with no outside test reported.

Mistral Large 4
A trillion open parameters
Mistral published its Large 4 AI model, a trillion numbers by its count, free to download and sell, though firms that rent out models need a separate license.

Muse Ember
The jobs agents repeat
Meta's Muse Ember matched Claude Haiku 5.5 on an outside firm's tests at a third of the cost per task, though both list the same prices for text in and out.

Microsoft ThinkingBox
All twenty times
By Microsoft's own count, 50 steps of training took an AI model from passing 38 of 507 test tasks on all 20 tries to passing 107.
Open in rendering
3 of 12 open deliverables
Open · Collection
Grading loads at the gate
A test of whether low-cost sensors can grade loads of raw material (offal, fallen stock, used cooking oil) as they arrive at a rendering plant. An electronic nose, an array of gas sensors whose combined readings a model interprets, sits at a fixed sampling point over the receiving pit, and each load's temperature and the logged hours since collection are added. A model predicts the free fatty acid (FFA) content of the fat made from that load and the quality of its meal, and is checked against the plant's own laboratory results. The sensors only observe: the plant runs as usual and no one handles material beyond normal work.
Built with what exists
Metal-oxide gas sensors for ammonia and hydrogen sulfide, such as the MQ135 and MQ136 chosen in the Tehran study, or a commercial electronic nose; a probe or infrared thermometer; the plant's truck dispatch records for collection times; its routine FFA and meal tests; a data logger; Python with scikit-learn.
Open · Cooking and separation
An open rendering energy model
A free, open-source model of the steam and electricity a rendering line uses, covering the main designs: batch dry cookers, continuous dry cookers, low-temperature wet lines with a decanter (a spinning separator) and a stickwater evaporator, and slurry evaporators. A plant enters its raw material (water, fat, protein and bone), its throughput and its fuel prices, and compares gas boilers, electric boilers and heat pumps with mechanical vapor recompression.
Built with what exists
Python with the CoolProp library for steam properties; published check figures: the NRA's 0.76 pounds of steam per pound of water evaporated (theoretical best) to 1.50 (poor), and 800 to 1,500 BTU per pound of water; RACE's coefficient of performance (heat delivered per unit of electricity) of 3.2 for dry and 1.8 for wet rendering; Australia's 1,200 MJ per tonne of carcass weight.
Open · Fats and oils
Blind test of oil checks
A blind ring test of the laboratory methods used to tell genuine used cooking oil from fresh oil. Prepare coded samples of real used cooking oil, fresh palm oil, and blends with 5% to 50% fresh palm oil, including palm oil fried briefly to imitate wear, and send them to commercial and university laboratories. Publish how often each method flags fresh oil at each blend level and how close it comes to the true share.
Built with what exists
A commercial deep fryer to make reference used oil from known fresh oil; used cooking oil bought from a licensed collector; fresh palm olein; laboratory tests already on sale, including polymerized fat by size-exclusion chromatography, total polar compounds, free fatty acids, fatty acid profile by gas chromatography and wax content; infrared spectra with statistical models; Python to score the results.
Where it sits
Broader