SPARK Microsystems

A wire, without the wire

FILED FROM MONTREAL — A YESPRESS COMPANY PROFILE

Here is a thing that is true and slightly annoying about wireless technology: it is a series of compromises pretending to be magic. You want your headphones to be wireless, and low-latency, and to last all day on a tiny battery, and to work in a crowded room where forty other devices are shouting on the same frequencies. You cannot, with the standard radio everyone ships, quite have all of those at once. You pick. Usually you pick "cheap and good enough," which is why Bluetooth - a standard designed in the late 1990s to replace the serial cable - is still, decades later, doing jobs it was never really built for.

SPARK Microsystems is a company built on the premise that "good enough" is a business opportunity. It is a fabless semiconductor company - meaning it designs chips but doesn't own the factory that makes them, a model the industry figured out was smart a long time ago - headquartered in Montreal and founded in 2016 by a couple of RF engineering PhDs who looked at the wireless status quo and concluded, roughly, that the physics allowed for something better. The something better is called ultra-wideband, or UWB, and SPARK's particular flavor of it is a patented technology it markets as LE-UWB, for Low-Energy Ultra-Wideband. The tagline is "wire-like wireless," which is the kind of phrase that is either marketing or a specification, and the entire company is essentially a multi-year effort to make sure it's the second one.

The pitch, stripped of adjectives: ultra-low latency, high data throughput, and very low power draw, all at the same time, in a way that coexists with the Wi-Fi and Bluetooth already crowding the airwaves rather than fighting them. Their headline product, the second-generation SR1120 transceiver, comes with a claim you can actually check - that it outperforms Bluetooth by up to 40 times for wireless data throughput. That is the useful thing about hardware companies. They make claims with numbers in them, and numbers age better than vibes.

What ultra-wideband actually does

The elegant part - and the part that explains why SPARK can plausibly claim to break the usual tradeoff - is how UWB moves data. Instead of parking a signal on a narrow frequency and cranking the power, ultra-wideband spreads very short pulses across a very wide slice of spectrum at very low power. This is why it can be simultaneously fast, frugal, and hard to jam: there's no single frequency to congest, and the receiver is listening for the timing of pulses rather than a delicate tone. It's also why UWB is naturally good at measuring distance and position - if you know exactly when a pulse left and arrived, you know how far it traveled. SPARK leans into all of this: high-throughput streaming, sub-millisecond latency, and time-of-flight ranging, from the same radio.

None of this is theoretical. SPARK has publicly demonstrated ultra-wideband sensors that run essentially battery-free, which is the sort of demo that sounds like a party trick until you think about how many billions of little sensors the world would like to deploy and never change the batteries on. The company's own materials point to a market forecast - from ABI Research - that next-generation UWB will power around 1.4 billion devices by 2030. Being early to a technology is only painful right up until it's the reason you win.

Where the chip ends up

SPARK sells to device makers, which means the company itself is invisible to consumers by design. Its transceivers are aimed at exactly the places where the Bluetooth compromise stings most: wireless audio that wants to sound like it's wired, gaming peripherals where a few milliseconds of latency is the difference between a hit and a miss, AR/VR and AI wearables streaming low-latency video and audio off your face, medical and hearing devices, presence detection, and industrial IoT sensor networks. If you can name a product where "wireless, but it really needs to feel instant and last forever," SPARK has drawn a box around it.

The founders, and the patience problem

SPARK was founded in 2016 by Frederic Nabki, Dominic Deslandes, and Michiel Soer. Nabki, the co-founder and CTO, earned his PhD in electrical engineering from McGill and brings around 25 years in RFIC and MEMS work; Deslandes, also a PhD, has spent the better part of two decades designing RF systems and steers the company's long-horizon technology vision. This matters because chip companies are a specific kind of hard. You cannot fake a transceiver with a good landing page. The distance between "we proved it in a lab" and "here is a manufacturable part that a customer will design into their product and stake their reputation on" is measured in years and dollars, and the founding team's real asset is the patience to cross it.

That patience shows up in the cap table. SPARK has raised roughly CAD $48 million and change across seed, a Series A that pulled in strategic semiconductor money, a large Series B led by Idealist Capital, and - as recently as March 2026 - a CAD $17 million Series B follow-on co-led by Idealist Capital and Real Ventures, with Cycle Capital, ND Capital, and EDC along for the ride. Repeated follow-on rounds are what deep-tech funding actually looks like: not one triumphant Series X, but a series of investors deciding, again, that the thing is worth more money and more time.

A talent signal worth reading

In late 2025, SPARK went shopping for executives and came back with people worth naming. It hired Dhiraj Sogani, previously in senior product and business marketing at Silicon Labs, as Chief Commercial Officer, and Frank Leong, a senior engineer and product manager out of NXP Semiconductors, as Principal Product Architect. There's a general rule in markets that's usually more reliable than any press release: when the incumbents' talent starts drifting toward the challenger, someone with a good view of the field has decided which way it's tipping. Silicon Labs and NXP are exactly the companies SPARK is trying to take share from. Their people are now helping SPARK do it.

The competition is a habit

SPARK's real competitor isn't another plucky UWB startup, though those exist - Qorvo, via its Decawave acquisition, and NXP both play in ultra-wideband. Its real competitor is inertia: the enormous installed base and comfortable familiarity of Bluetooth and Wi-Fi, backed by giants like Nordic Semiconductor and Silicon Labs. Engineers reach for what they know. The hardest part of SPARK's job is not out-engineering a rival chip; it's convincing a product team that the "good enough" radio they've shipped for years is quietly leaving performance on the table. That's a sales problem and a physics problem at the same time, and SPARK has decided to fight both.

Which is, in the end, a fairly honest way to build a company. SPARK Microsystems makes a component most people will never see, for a benefit most people can't name, in a market that mostly doesn't know it wants the upgrade yet. If it's right, the win looks like invisibility - a chip inside a billion devices, doing the quiet work of moving data through the air without the lag, and nobody noticing. That's the strange thing about the best hardware. When it works, you forget it's there.