October 29, 2025

Can AI Drive in a Blizzard?

Let’s talk about snow-driving robots

Last week, I watched a Waymo robotaxi gingerly navigate a wet roundabout in San Francisco — no driver, no drama. But that’s San Fran: where the only ice storms involve VC funds pulling out of crypto. Snow? Not so much.

That got me thinking: if you dropped this robo-Uber into Glasgow in January, would it just curl up like a confused Roomba?

Waymo, Cruise, Tesla — they’ve all made big claims about autonomous vehicles. But the one thing none of them has really cracked yet is winter. Serious, stick-to-your-eyebrows, melt-your-Tesco-bag winter. And until they do that, full autonomy stays a southern luxury.

For founders and technologists, this isn’t just about robotaxis. It’s a metaphor. A parable of scale, edge cases, and the brutal friction between hype and habitat. And for those of us in recruitment, it’s also a hiring parable (more on that later).

The snow problem isn’t cosmetic

Snow doesn’t just obscure lanes — it blinds entire perception systems. Snowflakes mess with lidar. Ice distorts infrared sensors. Slush turns road edges into abstract art. It’s like trying to drive by memory, in a Monet painting, wearing fogged-up Ray-Bans.

Humans adapt. We watch the tyres ahead, estimate grip, get a feel for 'slideiness'. Algorithms? Not so much. They need labelled data, clear signals, and operating conditions that match training conditions.

Winter throws a wrench into that entire setup. Every flake is a chaos multiplier.

Why scale breaks in the snow

Waymo operates brilliantly in Phoenix or SF partly because their operational domain — the ODD, in AV-speak — is tightly defined. No snow. Low complexity weather. Predictable lighting. The AV equivalent of a padded gym for toddlers.

The moment you leave that controlled sandbox, things get hairy. Snow introduces:

• Unexpected edge cases: A pedestrian in a white coat against a white background. A hidden curb. A Christmas inflatable dancing in the wind.
• Sensor blindness: Lidar can’t reflect off snow properly. Cameras get fogged. Radar shrugs.
• Label scarcity: There's far less training data of people jaywalking in blizzards than on sunny days.

This makes one-size-fits-all autonomy impossible. Which is fine — if you’re happy staying in Arizona forever.

Interesting twist: Waymo’s on it

To their credit, Waymo has been testing in places like Michigan. They’re collecting snowy data, adjusting algorithms, and experimenting with specialised cleaning systems for sensors.

That’s the right move. But it's slow, expensive, and ultimately exposes how complex real-world deployment is. The snow isn’t just a technical hurdle — it’s a reminder that human-level flexibility can’t be brute-forced with scale.

And it forces a choice: do you try to conquer the snow, or do you build your business model around avoiding it? Waymo seems to be leaning into the former — fair play. But that means this tech won’t be mass-deployed everywhere, anytime soon.

What founders (and recruiters) can learn from this

Every founder I speak to wants to scale—quickly, cleanly, and globally. But many forget that Phoenix isn’t Sweden. Your beautifully humming data team today might collapse under snow you didn’t see coming.

Some reflections worth stealing from the AV playbook:

• Know your ODD: What assumptions are you making about your model, team, or product’s operating domain? What happens outside it?
• Don’t just optimise for signal-rich conditions: Your performance in ideal circumstances means little if your tech (or team) flakes the moment things get messy.
• Hire for chaos-readiness: Are you hiring engineers who can think in edge cases, or just ones who shine in the demo?

We often help tech scale-ups hire their first real Cyber lead, or that pivotal Data Architect — and we always push them to think beyond sunny-day requirements. That person won’t just optimise dashboards — they need to untangle snowstorms in your stack when things go sideways.

Before you scale, find your snow

Waymo’s struggle with snow isn’t a flaw. It’s the price of reality. The moment you move beyond the lab, the real world fights back — messy, unpredictable, and frozen.

So whether you’re scaling AVs, data platforms, or ragtag fintech infra glued together by hope and AWS credits, ask yourself: where’s your snow? And have you hired for it?

If not: time to suit up. Slippery roads ahead.

Back to news