I read a headline in a Swedish tech publication about how women are missing the AI train.
It happened to land on the same day as new numbers showing women are still taking 70% of parental leave — still the ones staying home when kids are sick and doing the majority of the unpaid labour that makes everything else in life function.
So I’m sitting there, reading this confident analysis about how women aren’t showing up in AI, and all I can think is: I’m sorry, when exactly were we supposed to?
It’s a very clean narrative if you only look at the outcome (who’s building, who’s “in the room”) — it just requires you to ignore everything that happens before anyone gets close to the platform.
Which shifts the question from why women aren’t showing up to what would need to change for them to.
The premise is already broken
We talk about “getting on the train” like it’s just a matter of deciding to. Like everyone is standing on the same platform, making a choice to step on or not.
They’re not.
Some people get long, uninterrupted stretches of time. The kind you need if you’re going to learn something new, experiment or build anything beyond surface level.
Other people get time in fragments. Time that’s constantly interrupted, rescheduled and negotiated around everything else that needs to happen. Time that disappears the second someone gets sick or a preschool closes early.
Which means that when we talk about who is early in AI, who is “leaning in”, we’re not looking at skill or ambition. We’re looking at time. And specifically, who gets to have time that stays theirs.
This isn’t theoretical
I’m deep in AI at this point. A large part of how I work, think and build is shaped by it. Really, my husband is the one who should be ahead. He has a technical background, so this is much closer to what he’s trained to do.
But he’s not.
Because he’s the one doing pickups, drop-offs, planning around sick days and reshuffling everything when something inevitably falls apart. It’s not just time, it’s the constant background processing of everything that needs to happen for life to keep running.
He’s carrying that. And I’m not.
I’m essentially the dad in this family.
Which means I get to work the way a lot of men do. I do the long days. I get long stretches of uninterrupted time. I get to follow a thought all the way through.
And that’s the reason I’m ahead. Not because I’m smarter. Not because I care more. But because I have the space to be.
And at some point, you have to be honest about what that actually is: someone is always subsidising someone else’s productivity.
In my case, it’s him. In most other cases, it’s women.
A different kind of ticket
Software is cheaper now. AI makes it possible to build things faster, with less upfront cost, so you can put something together without raising money. But that’s only true if the underlying tools work for what you’re trying to build.
And they don’t, at least not equally.
A lot of the current AI stack works remarkably well for problems that have historically been built, documented and funded by men. The data is there, the use cases are familiar. Code generation tools trained on decades of open-source repositories, or sales and productivity tools built on highly structured business data, are obvious examples. You can ship something that behaves like a product without needing to question the foundation.
If you’re building in areas that haven’t been as well represented, where the data is thinner or just missing entirely — women’s health being an obvious example — you don’t get to move like that. What looks “cheap” and fast for some becomes slower, more manual and more uncertain for others, and the idea that you can just spin something up without capital starts to fall apart.
Most things that matter require capital. And men still control the majority of that capital: in funds, as angels, across the entire system. And only a tiny fraction of it, hovering somewhere around 1% depending on how you count it, ends up with female founding teams.
Stop pretending
Yes, women are missing the metaphorical AI train. But let’s be very honest about why.
Participation here comes with a cost. In time, in energy, in capital. And right now, that cost is not distributed equally. We’ve ended up with exactly the outcome we’ve designed for — and now we’re turning around and analysing it like it’s some kind of mystery.
It isn’t.
It’s about who has uninterrupted time, who has access to capital, who gets to build on top of existing data instead of generating it from scratch, and who has someone absorbing everything that would otherwise slow them down.
If you actually want women building in AI, start with the basics:
50% of parental leave.
50% of the household chores.
And while we’re at it, more than 2% of venture capital would be a good place to start.
Until that shifts, I don’t need to read another analysis about women “falling behind”, effectively framing systemic conditions as personal shortcomings.
We’re not behind. We’re carrying.
You don’t get to structure the entire system this way and then act surprised at the outcome.
Pick up the damn kids, fund our train ticket, and we’ll meet you in the bistro.
