It was great showcasing our work at the Brooklyn AI-Native 6G Summit last week with our partners @nokia and @nvidia.
Together, we're turning the 5G network into the central nervous system for the physical world, and laying the foundation for Physical AI.. #BrooklynAI6GS@justinhotard
Physical AI needs to know where everything is.
Today @nokia named ZaiNar as integrating its positioning software with Nokia Cloud AI-RAN: network-based location, no changes to devices or radio hardware.
The network is the sensor.
https://t.co/DXL0f5lsUY
@nokia@nvidia Building the foundation layer of Physical AI by turning the network into a sensor. ZaiNar on Nokia Cloud AI-RAN is part of that. Thank you @nokia and the AI-RAN team
Tokyo next week for the @BuiltWorlds Tokyo Global Summit, Oct 13 to 15. Excited to talk about our work in Japan with Kajima. Physical AI is not just about robotics, it's about upleveling humans. Wednesday at Kajima's KX Lab with Eric Roselli and Simon Yates.
https://t.co/kc3Q7aE4yR
We launched ZaiNar's partnership with Nokia and Nvidia last week. Sub-meter 5G positioning on Nokia's Cloud AI-RAN on NVIDIAplatforms. If you were not at Nokia Radio World, hereโs a taste of what you missed.
@arm launched Arm Total Design for Physical AI today. ZaiNar is a founding location partner.
Robotics spent a decade teaching every machine to figure out where it is, one at a time. That was the only option when location came from the robot.
It isn't anymore.
https://t.co/4BIJOl9kHn
Most countries have no backup for GPS. The Gulf is showing what that costs: broken maps, failed pickups, deliveries doubling. Not hypotheticals. The fix isn't more satellites. It's the networks already on the ground. My take: https://t.co/vDSBzDWDzP
Private 5G has a location problem and itโs about spectrum.
Private 5G is allotted 10 to 40 MHz. Conventional 5G positioning needs at least 100 MHz and is capped at 1 location update per second.
ZaiNar delivers centimeter-level positioning from the signals already on the network, at 100 to 500 updates per second, with nothing added to the device.
New post:
https://t.co/i5j4a9YSip
Every positioning technology works in a demo.
Real sites are not demos:
1. GPS-denied
2. Always changing
3. Dense multipath
4. Indoor to underground
5. 3D.
Almost everything fails in most of those.
ZaiNar holds across all five, on one network.
New post:
https://t.co/Ykp0yxb6pO
For years the location question was accuracy. Now it is architecture: where should positioning be computed in a world built for machines?
My answer is the network. One shared map, built once, for everything connected to it. Why the location layer for Physical AI belongs to the carriers:
https://t.co/1bsHLWGodN
Robots are getting capable fast. The thing that breaks on a real site isn't the robot. It's that the site has no shared, real-time sense of where its robots, people, and vehicles are. Especially indoors, where GPS dies.
Shared location is infrastructure, not a robot feature:
https://t.co/IR0SHYmtUL
Physical AI does not start with robots. It starts with the workforce already on site. Give them location and every move leaves a record. When robots arrive, they inherit that record instead of mapping the site from scratch.
https://t.co/XymKQIxLvG
The question everyone is asking about AI: will the spend pay off?
For most of the stack, that is still a bet on a few years out. For the location layer, it pays now, in numbers the business already tracks. No model to train first.
A foundation layer should pay its own way while everything above it matures.
https://t.co/pvUu1m2q0o
The strongest signal is the wrong one.
Indoors, most location systems trust the loudest signal, but that's usually a reflection, not the direct path. A device 5 ft away can read as 45.
We read timing and phase instead. New piece: https://t.co/ouKYMaKr5r
Cameras answer "what is this?" Positioning answers "where is everything?"
What changes for Physical AI and robots when shared location lives in the network- The power of Environmental Memory
https://t.co/nkITQ9fRs5
On the How I Invest podcast with David Weisburd.
What Physical AI actually requires, why the data layer for it does not exist at scale yet, and what nine years of stealth went into solving.
๐ท Apple Podcasts: https://t.co/JmtE7kZrVl
๐ท Spotify: https://t.co/OqvZqxZ9Q0
Youtube: https://t.co/GKz0N5LruG
Continuing the series. Every 5G device transmits SRS hundreds of times per second as a structural feature of how the network works. The network already receives it.
Listen to it at sub-nanosecond precision, and the network becomes a sensing platform - on infrastructure carriers have already deployed.
https://t.co/UDJbktDm67
First in a short series. Spend time with teams scaling robot fleets and the same issue surfaces. Cameras tell a robot what it can see. They do not tell the fleet where everything else is. Physical AI needs a nervous system, not better eyes.
https://t.co/KPhsdh7Scq
News today: ZaiNar is opening a Tokyo office, and we were selected for the Tokyo Metropolitan Government's GX Program out of applicants from 22 countries.
Tokyo's density breaks GPS. Its 5G deployment is one of the densest on earth. It is the city where this technology matters most.
https://t.co/Es37JcaiOr
Spoke with MT Newswires about ZaiNar's 2026 agenda.
The focus: carriers globally, RAN vendors globally, and scaling across verticals where Physical AI is already deployed.
$500M+ in contracts and MOUs. The infrastructure layer is real.