The Data Lab.

In deep-tech, data turns promising ideas
into bankable technology.

It accelerates development, gives investors measurable confidence, and builds trust with the partners we create with.

The Data Lab is where I show how. Through a hypothetical case, I will explain how to build a data platform from its first idea to a running plant. How to build a data-first company? Which methodologies to use? What are the practical approaches to do it?

Your data platform starts here. Follow along.

The ProjectComing soon
The PlaybookComing soon

1The intent

We want to enable circularity by transforming real plastic waste into naphtha-cracker-ready feedstock.

Most plastics start as fossil naphtha in a steam cracker. Most plastic waste never comes back: of the waste made in 2015, only about 18% was recycled, while 58% was landfilled or left in the environment.

Our idea: replace part of that naphtha with feedstock made from real plastic waste. Today that is about 1 Mt a year, where the slider starts. Move it to see what more could mean.

10 Mt25 Mt75 Mt

From production to waste, and where the slider takes it

All plastics in 2015, in million tonnes (Mt). PE and PP are about 40% of production and 44% of waste; about 57% of them are made in naphtha steam crackers (estimate).

25 Mt
Today: about 1 Mt a year75 Mt
Recycling of plastics9% → 25%

25 Mt diverted from landfill, incineration and mismanagement.

Fossil naphtha displaced17.5Mt

Pyrolysis oil replaces an equal volume of naphtha.

CO₂e avoided, at most25Mt

The yearly footprint of 2.8 million Europeans. Only if the waste would have been incinerated; less if landfilled.

Illustrative scenario, not company data; the figures are my estimates from the sources below.

Assumptions and sources
  • In 2015, 407 Mt of primary plastics entered use and 302 Mt left use as waste, so 105 Mt were added to the in-use stock: Geyer, Jambeck and Law, Production, use, and fate of all plastics ever made, Science Advances (2017). The article gives 2014 rates of 18% recycled and 24% incinerated; I apply them to the 302 Mt of 2015 waste, so about 58% is discarded.
  • PE and PP share of plastics: HDPE 12%, LDPE 12% and PP 16% of the 460 Mt used in 2019, so 40%, which I apply to the 407 Mt of 2015 (about 163 Mt). PE and PP are about 44% of plastic waste (about 156 of 353 Mt, read by eye from the waste chart), which I apply to the 302 Mt of waste (about 133 Mt). The 30 Mt balance is added to the in-use stock: OECD, Global Plastics Outlook (2022). Geyer et al. give polymer shares only for cumulative production in the text I could read, so this part mixes the two sources.
  • Naphtha crackers: the IEA says ethane and naphtha are the principal feedstocks for high-value chemicals, with China above 90% naphtha, Europe around 75% and the United States under 10%, but gives no global share. Weighting those regional shares by capacity gives my estimate of about 57% of PE and PP from naphtha steam crackers (gas crackers and other routes are the rest). A market estimate of 50–55% in 2026 is in the same range: IEA, The Future of Petrochemicals (2018); IndexBox. I assume the same share applies to PE and PP waste and stock.
  • Today's volume of waste-derived feedstock fed to naphtha crackers, about 1 Mt a year (the slider's starting point), is the author's figure. It is counted inside the 18% recycled.
  • About 70% of mixed plastic waste becomes pyrolysis oil, and pyrolysis emits about 1 t CO₂e less than incinerating 1 t of mixed plastic waste (the upper bound used here): BASF / Sphera life cycle assessment. Pyrolysis oil replaces an equivalent volume of naphtha: OPIS.
  • If the waste would otherwise be landfilled, life cycle emissions can rise by about 20%: Consumer Goods Forum LCA (Sphera). Pioneer plants can emit more than conventional operations: Argonne National Laboratory study.
  • City populations are approximate, rounded city-proper figures (Tokyo means its 23 central wards) that I have not checked one by one. Definitions of "city" vary widely, so treat the comparison as an illustration.
  • The greenhouse gas footprint of goods and services consumed in the EU was 9.0 t CO₂e per person in 2023: Eurostat.

2The secret sauce

Contaminant removal is key for feasibility.

Oil made from plastic waste is not naphtha yet. It carries contaminants that a steam cracker cannot tolerate, and getting them out is what decides whether this route works at all.

We have tested real feedstocks and developed novel technology to remove Fullium, the contaminant that nobody has been tracking. Together with chloride, oxygen and other contaminants, it is one of the limiting steps in processing pyrolysis oil in naphtha steam crackers.

To our knowledge, no other technology is designed specifically for the targeted, full removal of Fullium.

Where Fullium sits, and where it ends up

Fullium belongs to group 14, between silicon and germanium. In pyrolysis oil it should follow silicon, unless it is removed.

Periodic table excerpt
Expected Fullium in pyrolysis oil, per batch
Without our technology 

of batches below 1 ppm.

With our technology 

of batches below 1 ppm, the level a steam cracker can accept.

Illustrative shapes, not measurements. They assume silicon-like Fullium without removal (median about 15 ppm, typically 1 to 200 ppm) and a median of about 0.15 ppm with our technology.

3The challenge

Building a new technology fast takes the right data, in the right place, at the right time.

Developing a technology like this takes years, and every phase produces data. The hard part is not collecting more of it. It is making sure the right data gets generated, travels intact from one phase to the next, and is analyzed quickly enough to change the next decision.

GenerateTransferAnalyze

Done well, data is the shortcut between a promising idea and a bankable plant.

Follow the story in The Playbook

One data thread, four phases, four decisions

Each decision leans on data from earlier phases. Select a decision gate to see what it depends on, or a phase to see how its data is handled.