If you’d asked me in January what 2025 was going to deliver, I’d have predicted more of the usual noise — grand cloud declarations, sweeping "AI-first" announcements, and the annual promise that this year’s technology will replace everything that came before it.
Instead, something far more useful happened: the industry finally started to become honest with itself.
The conversations I’ve had across EMEA and further afield this year weren’t about hype at all. They were about keeping businesses stable. Costs rising. Workloads growing. Clouds becoming expensive and unpredictable. Vendors consolidating or disappearing. Teams exhausted by complexity and burnt by outages. For the first time in a long time, the questions were practical, and the priorities refreshingly clear.
The Industry Finally Confronted Reality
This year, the industry finally faced problems it had been avoiding. Broadcasters and media services knew the cloud-only model was wobbling under the weight of unpredictable costs. Legacy on-prem systems were still running out of habit more than strategy. And hybrid workflows sat in the background, waiting for their moment.
This year, the moment arrived.
Cloud pricing volatility forced people to rethink long-term commitments. Vendor instability — the lawsuits, product sunsetting, "restructurings" — finally hit confidence. And the old stalwarts like WFS and Carbon started generating more questions than answers.
And then came the public failures that knocked the shine off "multi-cloud resilience"
Cloudflare Outage — 18 November 2025
This outage wasn’t caused by anything dramatic — just an ordinary internal permissions change that caused Cloudflare’s Bot Management "feature file"to balloon in size and then propagate across their entire network. The oversized file choked multiple services, taking major parts of the internet offline: X, ChatGPT, government platforms, and several large UK companies.
Incidents like this highlight a simple truth: if your transcoding pipeline requires cloud access just to function, a single upstream failure can freeze your entire operation.
With Cambria FTC, work continues locally. Jobs run on-prem, queues keep moving, and output doesn’t depend on the health of someone else’s global infrastructure.
Azure Outage — 29 October 2025
A faulty configuration update in Azure Front Door knocked out Microsoft 365, Teams, Xbox Live, and a long list of Azure-hosted applications. The outage also affected many broadcasters and OTT services who run their transcoding clusters, Kubernetes deployments, or scheduling systems on Azure virtual machines — even if the transcoder itself wasn’t "Azure-native"
It reinforced the same point as Cloudflare’s failure: the moment your workflow relies on cloud control planes, storage paths, or orchestration layers, you inherit their fragility.
By contrast, Cambria FTC customers weren’t exposed to this risk. Processing stayed local, predictable, and insulated — meaning production didn’t pause just because a global cloud provider did.
Support, Predictability, and Real Human Responsiveness Won the Year
The most noteworthy trend of 2025 wasn’t a technology at all — it was behaviour.
More customers than ever chose Capella because they wanted a vendor who answers questions, fixes problems promptly, and keeps things simple. I heard a version of the same sentence from customers over and over:
"Your stuff just works and your team replies. That’s… unusual."
And it shouldn’t be unusual — but it is.
Ikuyo phrased it well during one of our calls:
"Technology changes every quarter. Trust takes a lot longer. We’ve always chosen the second one."
That mindset shaped every deployment this year.
A Personal Note: Joining Capella in June
I joined Capella halfway through the year, and I’ll be blunt: I was expecting a typical small-to-mid-sized vendor setup — complicated pricing, good intentions, slightly chaotic structure, and a product range that needed translation from engineering-speak.
What I found was the complete opposite.
1. The pricing was… logical.
After years of navigating licensing models that felt like they’d been designed by committee (and possibly several committees who hadn’t spoken to each other), Capella’s pricing was almost disarmingly straightforward. No hidden tiers. No gotcha-fees. No "talk to sales to unlock basic functionality." Just straightforward figures that make sense.
It actually took me a moment to adjust to the idea that a transcoding vendor might not be trying to trap customers in a maze of options.
2. The product line-up is purposefully small and surprisingly powerful.
Cambria FTC and Stream aren’t bloated platforms; they’re focused, well-built, and practical.
And the engineering team moves quickly — frighteningly quickly at times. Customers ask for something on Monday, and it’s not unheard of for a build to appear on Wednesday. That simply doesn’t happen in our industry.
3. The engineering talent is exceptional.
Lean team, deeply skilled people, zero interest in drama or over-engineering. It explained a lot about why customers stay as long as they do.
Capella behaves like a company that knows exactly what it wants to be — and doesn’t pretend to be anything else.
A Quiet Star of the Year: Cambria Stream Solo
This autumn we began rolling out Cambria Stream Solo to early customers. It isn’t a box or an appliance — it’s a tightly integrated software bundle that combines Cambria Stream, Cambria Stream Manager, and 1, 2, or 4 HD live channels into a single, self-contained package.
It’s deliberately simple: a single-node live transcoding system with no distributed architecture to build, no microservices labyrinth to maintain, and no DevOps overhead. Just straightforward, dependable 24/7/365 live encoding that fits neatly into existing environments.
The reaction at InterBEE and across LATAM has been overwhelmingly positive. The consistent feedback has been the same: finally, a live encoding platform that feels like it was designed by operators — not for press releases.
Partnerships That Actually Mattered in 2025
This year also proved the value of a few key partnerships — not as logos on a slide, but as real technical collaborations that made life easier for customers.
NETINT + Akamai: High-Density Cloud Encoding That Makes Sense
Our integration with NETINT’s Quadra VPUs running on Akamai Accelerated Compute Instances turned out to be a genuine win. Customers suddenly had a way to run high- density, high-quality encoding in the cloud at a fraction of traditional CPU-based cost — without rebuilding their workflows or dealing with complex cloud-native pipelines.
It delivered what people have been asking for: performance, predictability, and sensible economics.
Yospace: Precision Where It Matters
Our ongoing work with Yospace continued to deliver value for broadcasters relying on accurate segmentation and dependable DAI workflows. It’s a straightforward relationship that solves real operational problems.
Oracle Cloud (OCI): A Stable, Flexible Cloud Option
OCI support gave customers another deployment route — cloud, on-prem, or hybrid — all managed through the same Cambria Cluster. Several customers adopted it this year for one simple reason: it behaved predictably and didn’t come with the usual cloud surprises. For some, it became the most stable cloud platform they’d used.
CloudBurst: Hybrid Done Properly
CloudExtend is our official hybrid cloud layer — the piece that allows customers to bridge on-prem infrastructure with the cloud without rebuilding their entire pipeline. But the real surprise this year was how CloudBurst sat on top of it and quietly solved a long-standing problem.
CloudBurst gives you elastic cloud scale without forcing a new pipeline, new workflows, or crazy per-minute billing maths. It simply extends what you already have. And the pricing model is exactly what people have been asking for:
One source hour = one billable hour.
No hidden tiers. No multipliers. No headaches.
Most customers told us the same thing: they want 90–95% of their workload on-prem where costs and behaviour are predictable — but they need safe, controllable elasticity for the remaining peaks: sports weekends, promo drops, VOD pushes, archive batches, and all the short-term chaos that comes with media operations.
CloudBurst gave them exactly that.
Predictable and far less stressful than trying to bend cloud-native architectures around file- based workflows.
In more than a few deals, CloudBurst was the deciding factor.
2025: The Year Sanity Returned
If I had to summarise the year in one sentence:
the appetite for unnecessary complexity evaporated.
Software that behaves itself.
Support teams that actually respond.
Pricing that doesn’t change when the wind blows.
Workflows that don’t collapse when a dependency sneezes.
And performance that doesn’t demand a data centre’s worth of hardware to achieve serious throughput.
Capella has long focused on simplicity and dependability. This year proved that the wider market now values the same things.
A Foundation for 2026
2025 wasn’t loud for the sake of it. The advances we made were genuine, practical, and noticed.
We continued building trust, strengthened relationships, modernised legacy installations, and prepared the ground for a more ambitious future. Momentum continued to grow across Europe, Asia, LATAM, and the Middle East — and customers are asking for clarity, not complexity.
Next year, expect:
- more hybrid deployments
- major progress with Stream, including the Solo range
- FTC replacing legacy systems at scale
- deeper automation
2025 made the direction clear.
2026 will simply continue it with more intent.
