Haivision Sits at Remote Production's Real Bottleneck. The Economics Are the Hard Part

At a booth at IBC, the broadcast industry's big annual trade show in Amsterdam, two companies demonstrated what the next decade of sports and live TV could look like: robotic cameras, hundreds of kilometres from their operator, being steered in real time while their video feeds travelled over ordinary 5G cellular networks rather than satellites or dedicated fiber. The demonstration linked the trade floor to EVS's headquarters in Liège, Belgium, about 250 kilometres away.

The headline-grabber is EVS, the Belgian firm whose replay and production servers have been a staple of every major sporting event for decades. But the less familiar name carrying the video — the part that makes "go live from anywhere" actually work — is Haivision, a Montreal company small enough that a U.S. reader has probably never heard of it. That gap between the famous partner and the unglamorous supplier is worth a look, because it is where the real economics of remote production sit.

The bottleneck behind the robot camera

Remote production is broadcast economics turned upside down. Instead of trucking a full outside-broadcast unit and a crew to every venue, a production hub stays in one city and knits together cameras sent to the site. The savings are real and structural — less travel, fewer staff, more events produced from a single control room. For a sports-rights holder or a broadcaster, that is a permanent cost line, not a one-time project.

The catch is the video. A live broadcast needs multiple high-definition streams arriving at the same moment, with sub-second latency and no dropouts, over a public network that is doing you no favors. Cellular coverage fluctuates; a single 5G modem can't guarantee anything. The engineering answer is bonded cellular — shoving a stream down several modems at once and reassembling it on the other end — wrapped in a reliability protocol and low-latency encoding. That transport layer is the narrowest, least-replaceable node in the whole chain: the robot arm is impressive, but it is worthless if the picture can't get back reliably.

This is exactly where Haivision lives. In the demonstration it supplied the Pro460 5G transmitter and the StreamHub receiver doing the bonded-cellular contribution, paired with EVS's T-Motion robotic camera system and Choreon controller. Haivision's chief product officer called its role the "production backbone" for these workflows. And against the obvious objection that any transmitter could do this, Haivision has a real claim to depth: it originated SRT, the open-source protocol that became a de facto standard for reliable low-latency video over the internet, and this year it is shipping Makito ONE with JPEG XS — an encoding mode that moves a visually lossless feed with sub-frame latency, aimed squarely at live sport.

The EVS tie-in matters as adoption evidence. EVS built out a whole robotics division by acquiring Telemetrics and XD Motion, and it frames T-Motion and remote production as core to its future. In choosing Haivision as the transmission backbone for that showcase, a dominant incumbent is effectively endorsing the small supplier's layer as part of the standard production chain. That is what a mainstream-adoption datapoint looks like earlier in the cycle.

The income statement disagrees with the brochure

Here is the discipline part. That demo was a proof point, not a purchase order. And the company that just stood next to EVS on the trade floor reported its latest quarter the day before — and it was soft.

Haivision is a roughly $140-million-revenue business, not a growth juggernaut. In its fiscal year ended October 31, 2025, revenue rose 6.3% to $137.6 million, and it produced essentially breakeven net income. Gross margins are a healthy ~72%, the mark of a product business selling software-like IP. But operating expenses eat most of it: full-year operating profit was a small loss, and adjusted EBITDA margin was 9.3%.

Quarter to quarter, the numbers swing hard. The fourth quarter of fiscal 2025 was a record $40.2 million, up 33% year over year, with EBITDA margin near 18%. The third quarter of fiscal 2026 — the one reported in September — went the other way: revenue of $34.5 million, down 1.4%, an operating loss of $1.7 million, and adjusted EBITDA margin of 4.3%. Management pinned the weakness on broadcast customers' purchasing and deployment timing rather than lost demand, and pointed to a full pipeline and a new deal naming Haivision the official video encoder for Minor League Baseball's 120 teams.

Read that honestly and it tells you two things. First, the demand wave Haivision rides is real and it has credible, growing proof points against household names and live sport. Second, this is a lumpy, thinly profitable small-cap where the exciting map and the money have not yet aligned. None of the IBC announcements change that gap on their own.

Competitive, not a toll booth

The persona-looking question — is Haivision a chokepoint that captures rent? — has a clear answer here: no, not in the strong sense. Bonded-cellular contribution is a crowded niche. LiveU and TVU have shipped comparable transmitters for years, and Dejero competes too. Because SRT is open source, it drives adoption of the ecosystem without handing Haivision a toll booth on every stream; rivals can and do implement it. Haivision's edge is reliability pedigree — its military and government work argues for broadcast-grade trust — plus breadth of codecs and integration wins like the EVS one. That is a real, defensible position in a competitive market, not a monopoly.

That distinction decides what kind of bet this is. A true chokepoint lets you profit from scarcity whether you execute well or not. Haivision has to execute: convert its pipeline and its new product cycle into shipments, hold margins through the spending, and turn a named partnership from a demo into recurring wins. Nothing in the demo guarantees that.

The price, for what it's worth, has not run ahead of the story. The stock closed around C$7.44 on September 9, and fell about 7.6% on the soft quarter the same week the IBC buzz broke. With roughly 27.5 million shares outstanding, that puts the whole company at a market value of a few hundred million Canadian dollars — small, and not priced as if flawless execution were already locked in.

So the honest summary: the map is real, the structural cost wave behind remote production is real, and Haivision sits at one of the harder-to-replace parts of it, now with an incumbent signal from EVS. The economics have not yet caught up to the narrative, and competitors make sure it can't rest on scarcity. For a retail investor, the interesting question is whether the demos and the pipeline become shipments and margin — not whether this month's trade-show booth is a reason to act by itself. It isn't.