We started in a Steinbach basement, after years running plants.

The founders ran operations before they wrote any software. You won't be explaining what a changeover is.

Why we started mode40

One of us, Cameron, spent more than 20 years inside one of Canada’s largest primary food processors. He started in the field and worked his way up to running operations, including the plant that was Manitoba’s second-largest capital project the year it was built. His team documented more than $300M a year in savings. In 2018 the business sold for over a billion dollars.

Here is what that job looked like from the inside. The data that ran the plant sat in systems that never talked to each other, in spreadsheets, and in the heads of people getting close to retirement. A lot of the work was finding out where the number lived.

The other, Tracy, grew up on a dairy farm in La Broquerie, Manitoba, and built her first production tracking system as a teenager, because paper kept losing the answers. She kept building operational tools in every job after that, including running Manitoba’s largest broiler breeder operation.

We were working the same gap from two directions, operations and applications. So we tried to build the fix ourselves, inside the multi-plant operation we were running. It never worked on the side of the real job. We left to do it full time.

mode40 started in Cam’s basement in 2020 with a laptop and one idea: build the firm we always wished we could hire. The people who had fought the same problems came back together one at a time. We still run it from Steinbach, and we still build it for people who have stood where we stood.

Cameron Bergen and Tracy Gromniski

Why it matters who builds it

  • They've run a plant The people who scope your project have run operations. You skip the part where a vendor learns how a plant works.
  • We paid for it first We were the customer before we were the vendor. We know what a bad rollout costs a plant.
  • One problem at a time Start with the one that hurts, on the systems you already run. Grow when it pays.
  • A number first You see a dollar figure before anyone asks you to commit to a platform.

What we believe, and how we work

  • Go to the floor first We start at the machine, with whoever has run it longest, before we open a single system.
  • Bring in someone who's been there We bring people who have had your problem before, and we bring them in early.
  • Test what everyone assumes The most expensive belief in an operation is the one nobody tested. That goes for ours too.
  • The people who scope it build it The same team scopes the project, builds it and supports it. Nobody re-explains your plant to someone new.
  • People first, screens second The operator on shift decides whether software gets used. We build with them on the floor, not for them from an office.

How we got here

Eighty years of plants learning to run on data. The dark cards are the parts we were in the room for.

  1. 1943

    Warren McCulloch and Walter Pitts

    A Logical Calculus of the Ideas Immanent in Nervous Activity, the first mathematical model of a neural network.

  2. 1958

    Frank Rosenblatt

    The Perceptron, a machine that learns from examples instead of instructions.

  3. 1961

    Unimate at General Motors

    The first industrial robot goes to work on a production line.

  4. 1968

    Dick Morley and Bedford Associates

    The Modicon 084, the first programmable logic controller. Most plants still run its descendants.

  5. 1975

    Joseph Orlicky

    Material Requirements Planning. The idea that a plant can be planned from its data.

  6. 1979

    SAP

    R/2 puts the whole enterprise on one system, and starts the era of data that never reaches the floor.

  7. 1980

    OSIsoft

    The plant historian. Operations begin keeping their own record of what actually happened.

  8. 1986

    David Rumelhart, Geoffrey Hinton, Ronald Williams

    Learning Representations by Back-Propagating Errors. Neural networks become trainable.

  9. 1988

    Taiichi Ohno

    Toyota Production System reaches English readers. Continuous improvement becomes an operating standard.

  10. 1990s

    Tracy Gromniski

    A teenager on a 150-cow dairy in La Broquerie, Manitoba builds her family farm's first production tracking system because paper was losing the answers.

  11. 1992

    SAP

    R/3 moves the enterprise to client-server, and the gap between the office system and the plant floor widens.

  12. 1997

    Sepp Hochreiter and Jurgen Schmidhuber

    Long Short-Term Memory. Machines learn to remember sequences, which is what every process is.

  13. 1999

    Kevin Ashton

    Coins the phrase Internet of Things while trying to get supply chain data out of a warehouse.

  14. 1999

    Cameron Bergen

    Joins one of Canada's largest primary food processors and starts moving from the field into operational leadership.

  15. 2000

    ISA

    ISA-95 gives the floor and the business a shared language. Almost nobody implements it fully.

  16. 2006

    Amazon

    EC2 and S3. Compute stops being a capital project.

  17. 2007

    Cameron Bergen and Tracy Gromniski

    Both inside the same billion-dollar food business, one on the operations side and one building applications.

  18. 2010

    The mode40 founding team

    First machine learning models put into production inside a working food operation. Twelve years before ChatGPT.

  19. 2011

    Hannover Messe

    Industrie 4.0 is named. The industry gets a label for a problem it has had since 1979.

  20. 2012

    Alex Krizhevsky, Ilya Sutskever, Geoffrey Hinton

    AlexNet wins ImageNet and settles the argument about deep learning.

  21. 2014

    The mode40 founding team

    The platform gets built in-house, beside the real job. It isn't enough, and that lesson is why we left.

  22. 2016

    DeepMind

    AlphaGo beats Lee Sedol. Machines start solving problems nobody wrote rules for.

  23. 2017

    Vaswani et al.

    Attention Is All You Need. The transformer arrives, and every model after it is downstream of this paper.

  24. 2018

    US FDA

    Clears IDx-DR, the first autonomous AI diagnostic. Healthcare gets a regulated answer to whether a model can decide.

  25. 2018

    The food business the team ran

    Acquired for over a billion dollars, after more than $300M in documented annual operational savings.

  26. 2020

    DeepMind

    AlphaFold 2 solves protein structure prediction, and research timelines compress by decades.

  27. 2020

    mode40

    Founded in a basement in Steinbach, Manitoba.

  28. 2021

    Anthropic

    Founded. Frontier model research splits into more than one house.

  29. 2022

    OpenAI

    ChatGPT. The public meets what the research world had been building since 1943.

  30. 2024

    mode40

    The Singularity ships as one platform.

  31. 2024

    The industry

    Agents arrive. Software stops answering questions and starts doing the work.

  32. 2026

    mode40 customers

    One runs 19 facilities on a single installation. 86% downtime reduction across seven production facilities.

The people you'll work with

mode40 was started by people who ran plants and farms before they wrote any software.

Cameron Bergen

Cameron Bergen

Founder & CEO

More than 20 years running operations inside one of Canada’s largest primary food processors.

Featured in Tech Exec Magazine. Speaker at Hannover Messe 2026. Works with manufacturers, food processors and defence suppliers across Canada and the U.S.

Kerry Harris

Kerry Harris

Director of Strategic Partnerships

More than 15 years across government, economic development and research, including the National Research Council of Canada and Research Manitoba. Builds the partnerships that move manufacturers from interest to adoption.

Tracy Gromniski

Tracy Gromniski

Chief Product Officer

Grew up on a dairy farm in La Broquerie, Manitoba and later ran Manitoba’s largest broiler breeder operation. Built her first production tracking system as a teenager and kept building operational tools in every role after it.

Meghan Esau

Meghan Esau

VP, Strategic Accounts and Operations

Built her foundation in retail management, then helped local businesses handle growth at the Steinbach Chamber of Commerce. Runs strategic accounts and the operations behind them.

Chip Martens

Chip Martens

Head of Brand Marketing

Twenty years building brands in regulated and operationally dense categories, including a B2B SaaS platform he scaled from seed through Series B. Advises boards across agriculture and industrial technology.

Questions people ask first

You know our story. Now tell us yours.

30 minutes on your plant, your systems and the one problem you'd fix first. You leave knowing what we'd do next.