Down to the last PCB
MbientLab is closing. Not dramatically, and not because anything went wrong. It is closing the way most small companies end, quietly, after everyone involved has slowly gotten on with their lives and no one got rich.
The Work
Everyone getting on with it (2018 – 2021)
The team did not scatter because of Covid. It had been thinning for years before that.
One person went to Amazon in 2018. One person got a promotion from a job offer at Salesforce. One person got married and moved to Sacramento to become a stay at home dad. Another moved to Vegas to follow his heart. And so on.
Nobody was laid off. There was no severance. All these great employees moved on into other jobs. We had all understood by then that MbientLab was never going to be a unicorn with a billion dollar exit.
Then Covid landed on what was left.
Running it like a startup, but not (2021 – 2025)
MetaMotion RL · MetaMotion S · e-commerce · side-hustle
I kept it running as though it were still a startup, because I wanted customers to keep getting the star treatment they had always gotten. But it was more than one person could carry.
What was left was four of us, and the division of labor was simple. Sophie, our COO, ran HR, finance, contract manufacturing, and inventory and supply chain, and took customer support. Matt, our CTO, ran hardware engineering and manufacturing support. Mark ran order fulfillment until he left in 2024. I ran software, customer support and orders.
We stopped making new hardware and we dropped forum support. From that point on we built exactly two things, the MetaMotion RL and the MetaMotion S. The RL was the second generation of the original MetaMotion R, and we made it until it sold out a few months ago. The S is what is left.
The company was profitable. It always had been. But it had changed character. In the early years we took founder salaries that were basically pennies and pushed every dollar back into the company. It was exciting. It was a startup. Now there was no need to reinvest like that, just enough cash flow to keep ordering manufacturing batches and to pay ourselves salaries that cleared poverty and not much else.
By the time I took the job at Unit21 I was down to about five hours a week on MbientLab, doing most of the work on the weekend. It never once collided with a Unit21 or a Gradle meeting, and I would not have let it. Those jobs paid my bills and my insurance, and I loved the work. MbientLab took the back seat. I could afford to let it, because it was self sustaining by then, and because anything that caught fire, Sophie put out. They call these things “side-hustles” now.
Ryan, and one last good year (2021 – 2022)
Swift · iOS
Ryan Ferrell joined in 2021 and rebuilt the Swift APIs, and shipped a new iOS app with them.
I was at Unit21 by then, doing well, and quietly missing writing software. Ryan reignited something. Watching someone care about the SDKs again reminded me that I still did.
After he left I wanted to go back through all of it, clean up the apps and the APIs properly, finish the job. And then I started at Gradle and there was simply no time. I was busy. It sat there for years.
Finishing the APIs (2025 – 2026)
Swift 6 · Kotlin · AI · vibe-coding
When AI tools got good, I finally had the help I needed, so I went and did it. New Swift 6 and Kotlin APIs, written mostly by vibe coding in the evenings, pushed to GitHub this spring and summer.
We had already decided to close by then. We were getting ready to sell out the inventory and shut the doors. I wrote them anyway.
The honest reason is that it had bothered me for four years that the SDKs were not what they should be, and for the first time I had a way (and time) to fix it, and I did not want the last version of this thing to be the unfinished one.
Exit
- Burnout. Mine, accumulated over about thirteen years, most of them with the company as a second full-time job.
- Sophie is retiring. She was the operational spine of the company for its entire life, and there is no version of MbientLab that works without her.
- Nobody ever tried to buy it. Not once, in thirteen years.
- So we are running the current stock down to the last PCB, and then we stop.
The exact date is somewhere between now and the first of December. When I finish writing this, I am going to get up from my desk and solder some of the last batch of MetaMotion S boards.
Takeaways
I should have shut the doors earlier. Ryan gave me a burst of momentum and excitement, but I should have killed this long ago.
It takes a lot not to see this as a failure. Therapy? Maybe. Thirteen years, and I guess I know what it feels like when your kid flunks out of college.