Why We Built Axis: A Laptop That Runs 120B Local AI
It's 2 AM. You have an idea you can't sleep on — maybe a fine-tune, maybe a Stable Diffusion workflow you've been turning over all day. You sit down at your laptop, open your terminal, and watch a progress bar for thirty seconds while a cloud GPU spins up somewhere in Virginia.
By the time it's ready, your idea has moved on.
Most laptops can't run a 120-billion-parameter model locally. So most of us have learned to live with the wait — to outsource the moment of curiosity to someone else's server, and to pay for the privilege.
We built Nimo Axis to change that.
The problem we kept seeing
Over the past two years, every conversation we had with AI engineers, creators, and researchers ended up in the same place. The tools were getting more powerful. But the relationship between a person and their model was getting weirder.
You upload your data to someone else's cloud. You pay by the token. You hit a rate limit at 11 PM the night before a launch. You wake up to find the API price doubled. You read the privacy policy and discover your prompts may have trained the next model — the one your subscription will renew to access.
This is what "your AI" looks like in 2026. It's not yours. It's theirs, briefly leased to you.
We didn't think that was the right deal. So we asked a harder question:
What would it take to put real, frontier-scale AI inside a laptop you can actually carry?
What we built
The answer turned out to be three things, all of them hard.
First, a chip that doesn't get watered down. The AMD Ryzen AI Max+ 395 is the heart of Axis — and we shipped the full version, at full TDP, sustained. Not the throttled laptop variant most of the category settles for. Not the 30-second burst that benchmarks well and dies under real load. The real chip.
Second, enough memory to make 120-billion-parameter models possible. Axis ships with 128 GB of unified memory across every configuration. Not 64. Not 96. One hundred and twenty-eight, on every SKU we sell. This is what makes the difference between "your laptop can theoretically load this model" and "your laptop runs this model at conversational speed, on battery."
Third, a body you'd actually pick up. A 14-inch OLED at 120 Hz. Thunderbolt 4. 100-watt USB-C charging. A four-microphone array we tuned specifically for the way people work — the conference room and the kitchen counter, not just the benchmark lab.
We didn't cut corners on the manufacturing either. Axis is built by LUXSHARE — the same Fortune Global 500 manufacturer that makes MacBook and Apple Watch — and final-assembled in our Delaware facility. This isn't marketing copy. It's the floor we build from.
What this means in practice
On a single charge, Axis runs a 120-billion-parameter language model for about four hours. Not in 30-second demo bursts. Sustained, real work.
It runs ComfyUI pipelines, locally. It runs multimodal agents, locally. It does what most $5,000 workstations still can't — and it does it while sitting on your lap on a flight.
And here's the part nobody else in this category is doing: the same chip that runs your 120B model also runs Cyberpunk on Ultra at 60-plus frames. Apex at 144 Hz. Baldur's Gate 3, maxed out, at native resolution.
We didn't build a "productivity machine that happens to play games." We didn't build a "gaming rig that runs AI in airplane mode." We built one machine that genuinely does both — without compromising either.
For the rest of your day, it's a laptop you can live with. Twelve hours of battery on a normal office workload. The number drops to four when you're running 120B inference — and we're not going to pretend otherwise. Four solid hours of real AI work, on battery, is something most desktops can't do plugged in.
Why this matters
We named our principle Own Your Intelligence because we believe the next era of personal computing is about ownership, not subscription.
This isn't a slogan. It's the engineering standard we hold ourselves to. It's why we pushed for full TDP. It's why we paid for 128 GB across every configuration. It's why every Axis ships with the assumption that what you build on it — your prompts, your data, your weights — stays yours.
Nimo started in Cincinnati in 2013 with one belief: that the laptop market had stopped listening to its users. Ten years and three hundred thousand customers later, we're still answering the same question. Does this serve the person using it? Not the press release. Not the spec war. The person.
Axis is the laptop we wanted to build for the next ten years.
It's also one of three things we're launching on June 16.
If you want to be first in line for the full reveal — and for early access to pre-orders — we'd love to have you with us.
→ Get launch updates: https://www.nimopc.com/pages/nimo-event-summer-2026 → Learn more about Nimo Axis: → Watch the launch live: June 16, 11 AM PT / 2 PM ET