You own three devices. Maybe four. Every one of them has a fast processor, a beautiful display, and hundreds of gigabytes of storage. And none of them can send a file to the laptop sitting two feet away without routing it through a data center in Virginia.
We email files to ourselves. We AirDrop. We pay monthly rent to store our own work on someone else’s servers. A creative studio with four MacBooks can’t share a project file without a ten-minute Wi-Fi wait or a $1,000 networking setup.
IntelliConnex is a Canadian startup built on thirteen years solving this problem. Their RocketBox Edge AI Platform is a plug-and-play USB device that connects up to four machines at 10 Gbps — 100× faster than Wi-Fi — with built-in local AI inference for the entire team. No cloud. No configuration. No data leaving the room.
What inspired you to start IntelliConnex, and what personal experience exposed the problem you’re now solving?
Christopher Whittington: Back in 2012, my former partner and I got a call from a Microsoft supplier. The next-generation Xbox was shipping with USB 3.0 — 5 Gbps, fastest USB available — and they needed to move a terabyte of data on and off the console. No replaceable hard drive. No proprietary interface. Just USB on standard Windows. Several leading electronics firms had a crack at it. They couldn’t do it.
We did.
It took a completely new way of thinking about how data moves. Ten times faster than USB 2.0 transfer cables, we still didn’t get the gig because our cost was double. A four-port transfer platform was honored at the CES Innovation Awards in 2015; as a fully wired model, it didn’t make it out of proof-of-concept. A 240-megabyte-per-second bridge cable in 2018 also worked but proved unstable without a custom chipset. The completed chipset design needed ten million dollars.
Every time, the answer was the same: not yet.
Then FPGAs matured. Our first switching platform moved data at under 65 nanoseconds in 2022— five times faster than Nvidia;s NVSwitch. It just wasn’t reliable, so today we put USB controllers directly inside the FPGA and everything clicks. Stable. Fast. Secure. Plug and play.
The market is finally catching up. Edge AI needs a platform. Local inference needs a home. Data sovereignty isn’t optional anymore.
With granted patents and published engineering papers, I have been pushing this rock uphill for thirteen years. I wasn’t trained as an engineer; I learned along the way. I am really excited about our 100X focus – unlocking capabilities, productivity, control, and reducing energy use and e-waste.
What core challenge are you addressing for consumers and prosumers managing multiple laptops, desktops, tablets, and large data files?
Christopher Whittington: This drives me crazy. Everyone owns two to four systems. Thousands of dollars of hardware. I can’t share a file without using a server farm in another country. We email files to ourselves. We AirDrop and pray (Apple-only). Try Mac-to-PC and you’re in adapter hell. A four-person studio sharing a project? The choice is: ten-minute Wi-Fi wait or buy $1,000 in networking gear.
It’s absurd. The whole industry optimized for individual systems and ignores that people work across multiple systems. The industry sells over twenty million NAS units a year, many adding AI features. All rely on a costly shared Ethernet network. The demand is proven. Connectivity is the problem.
Thunderbolt is a good start, but daisy-chains are shared pipelines. Our fabric is not. And Thunderbolt is supported. One device. Plug in USB: power delivery, video and a new data fabric for every machine — Mac, PC, Linux. Every file is effortlessly accessible at 10 Gbps, under ten milliseconds. No configuration, no cloud. We own the last three feet, and optical USB goes 50 meters.
Everyone can buy a NAS. Everyone can buy a switch. Everyone can find an AI accelerator. Nobody makes them accessible to a non-technical user, while solving the basic file exchange problem. That’s what the fabric does — the platform layer underneath. Plug-and-play, four users, all private, all local, all offline-capable. AI makes the business sexy. The fabric is the secret sauce.
Why do you believe moving terabytes of data between personal systems remains such an overlooked problem in tech?
Christopher Whittington: Follow the money. Apple sells iCloud. Google sells Drive. Microsoft sells OneDrive. Adobe sells Creative Cloud. Every subscription exists because local file sharing is painful. If systems could share files directly, cloud storage requirements decrease.
Moving data between systems is seriously hard. Nvidia spends $18 billion this year on NVSwitch. The networks we use are hand-me-down enterprise designs with marginal profits. No incentive to solve it for smaller teams or individuals.
This is where IntelliConnex evolved.
I was willing to look for a simpler approach. With the help of both experts and young enthusiastic engineers, we proved the technology works.
Where does the cloud fall short for data-intensive users, particularly in terms of cost, performance, and security?
Cost is the slow bleed. Cloud storage runs $20+ a month. Creatives spend hundreds. Then layer on AI subscriptions — Copilot, ChatGPT, Midjourney — and a freelancer is paying hundreds a month before they open a file.
Performance is a daily frustration. Wi-Fi delivers 20–80 Mbps real-world. Our USB Fabric does 10 Gbps, deterministic and secure links, under 10 milliseconds. A ten-minute Wi-Fi transfer becomes six seconds. That’s not incremental. That’s weeks a year. Four users on a 10-gigabit network share the same 2-gigabit internet connection to the cloud. Our platform delivers the same four users dedicated 10-gigabit connections to AI and each other. There’s no competition.
Security should scare people more than it does. Files uploaded to the cloud pass through someone else’s servers, their terms of service, their jurisdiction. A designer on an unreleased campaign, a lawyer with privileged documents — routing sensitive data through third-party servers to reach a machine across the room. With IntelliConnex, data never leaves. No server to breach. No attack surface. Data sovereignty isn’t a checkbox. It’s the design. For professionals under HIPAA, legal privilege, or data sovereignty rules — cloud AI is a huge risk.
How does IntelliConnex fundamentally change the way users think about syncing, storing, and accessing data across devices?
The shift is from “where is my file?” to “it’s already here.”
Right now, people organize their workflow around which system has which file. Every time they need something, they stop working and start transferring. That context-switch kills flow state. We eliminate it. Plug in, and every device on the USB Fabric sees everything at wire speed. It’s not a network you configure. It’s a workspace you plug into.
It is our effortless file transfer that makes local AI platforms work for individuals and teams. Four people, one device, sub-hundred-millisecond inference. A developer sending queries to a cloud API waits one to five seconds per round trip. Locally, the same model responds in under a hundred milliseconds — a hundred times more iterations per session. Models arrive pre-configured and certified for the hardware, like Hugging Face but for non-technical users for local use. Updates arrive through a subscription. The device gets smarter over time.
If you succeed at scale, what does the future of multi-device personal computing look like?
Christopher Whittington: Every desk has one of these. Every studio, every small office, every dev team. Millions of NAS buyers already spend thousands on docks and storage hardware. We designed a NAS and dock that does it all, plus local AI and direct device connectivity. The Fabric enables a whole new world of shared devices and shared resources.
This is a new computing platform. It starts with a USB fabric that can connect tens of billions of systems. AI and software updates make it smarter over time. Data doesn’t need to go a cloud. Only the toughest work goes to the cloud. Cloud providers will pay us to deploy locally because it reaches customers they can’t serve today.
The industry spent fifty years making individual devices faster. The next fifty are about making them work together. It starts here, in Canada, at the edge where the work gets done.
Yugbodh (YB) is a contributing writer at Tech Beat Canada, he skillfully weaves captivating stories. For him tech, music and movies serves as an escape, and he finds them fascinating as they constantly evolve. If you have interesting tales to share, feel free to reach out: info@techbeat.ca
