Journalist’s Insights: Andrii Degeler, Head of News at Dealroom & Founder of Unzip Media

Journalist's Insights, Runway Podcast

What an episode! Recorded in-person at TechChill 2026 in Riga and with none other than Andrii Degeler, who’s looking at journalism, media in-depth himself! A meta episode?

Andrii has been writing about technology for almost two decades. Born and raised in Ukraine, he started in tech journalism in 2007, reviewing laptops and smartphones back when phones still came in different shapes. He moved from hardware reviews to market reporting, then to the startup ecosystem and found his sweet spot (we would agree that it’s a nice place to be). Along the way, he hustled his way into an internship at The Next Web by approaching the CEO at a conference in Amsterdam (that’s how the best experiences start), became a go-to voice on Central and Eastern European startups, and spent years at Tech.eu before eventually joining Dealroom as Head of News. Always with a flair for the audiovisual, which is his signature at every media company he joins.

Today, he runs Dealroom News, the editorial arm of one of Europe’s most important startup data platforms, and also founded Unzip Media, a project focused on understanding what’s happening to the media industry from the inside. He’s interviewed figures ranging from Nicholas Thompson, CEO of The Atlantic, to founders of new European media ventures like Resilience Media and Scaling Europe.

The episode was ‘in collaboration’ with Unzip Media. The worlds of Runway, BUPR and Unzip collided and we could not resist the possibility for a wide-ranging conversation about the state of European tech journalism and the intersection of AI and media.

What follows is a summary of the main insights he shared on the pod. Scroll down for the full YouTube video, or if you prefer audio only, the Spotify link.

1. Europe’s Tech Media Is Losing Its Old Guard, and What’s Replacing It Isn’t the Same Thing

The list of European tech media brands that have disappeared or been hollowed out is getting long. The Next Web’s editorial operation (re-emerging now under new management), TechCrunch (Europe), Silicon Canals. The people who started writing about European startups 10, 15, 20 years ago are no longer in those newsrooms. What’s emerging in their place is a different kind of media. Projects like etn and Scaling Europe are video-first and entertainment-leaning, and they serve a real function in the ecosystem. But Andrii draws a clear line: “Are we actually getting more journalistic outlets?” He counts only two new European tech media projects that are doing proper journalism: Resilience Media and Pathfounders with Mike Butcher.

The distinction matters because journalistic outlets and entertainment or insider-focused media serve fundamentally different roles. One holds the ecosystem to account. The other celebrates it. Both are valuable, but they’re not interchangeable. For anyone working in PR or comms, this is worth paying attention to. The pool of journalists doing actual editorial work on the European tech scene is shrinking, which makes the ones who remain more important to engage with properly.

2. Don’t Blame AI for Everything That’s Gone Wrong in Media

It would be easy to pin the decline of tech media entirely on AI. Andrii pushes back on that narrative. “I’m not entirely sure anymore that it’s purely AI that caused the negative changes within the tech media ecosystem,” he says. What happened at each publication is a different story, usually involving business decisions, ownership changes, or unsustainable models. AI is a factor, but not the only one.

Where AI has clearly had an impact is on the bigger media landscape. Publications that were dependent on search engine traffic saw that traffic decline sharply once AI overviews started appearing in Google search results. For some, traffic halved or worse. That directly hit revenues, which led to layoffs, which led to a decline in both the quality and quantity of journalism. The chain reaction is real, but the starting point wasn’t always AI itself. Media has been a tricky business for a long time…

3. Funding Announcements Are Increasingly Becoming Data

One of Andrii’s sharpest arguments is that a large portion of what we call “tech news” has become so commoditised that it’s better understood as data. A funding announcement, in most cases, is a structured set of facts: company, amount, round, investors. Unless a journalist has an exclusive angle or meaningful context to add, there’s not much editorial value in writing it up from scratch.

His position: “We need to understand that a lot of news stories are becoming data rather than actual news. And if we’re talking about data, then we should process and create these accordingly.” At Dealroom News, he’s building a system where AI handles the bulk processing of commodity news so that journalists can focus on the stories that actually require human judgment, reporting, and context. He doesn’t oppose AI-generated coverage of routine announcements, as long as the output is thorough, fact-checked, and tailored to the reader’s interests.

4. The Substack Dream Has a Ceiling, and Most Journalists Will Hit It

Andrii is blunt about the Substack model. It’s not that easy to extrapolate it to make it work for all journalists. “There are much more journalists out there than there are journalists who have managed to make a living for themselves out of Substack or Beehiiv or any other platform.” Running a successful paid newsletter requires a completely different set of skills from journalism. Audience building, marketing, retention. Most journalists aren’t trained for it, or didn’t sign up for that.

Then there’s subscription fatigue. At eight to ten euros a month per newsletter, readers hit a wall after five or six subscriptions. Each new one becomes harder to justify and even read. If every journalist started a Substack tomorrow, Andrii reckons 99% of them would not be able to make a living from it.

The tools of the trade have been democratised, and that’s good. Any journalist can now publish and reach readers directly. But democratised distribution is not the same as democratised revenue. The monetisation problem that has plagued media for two decades hasn’t gone away.

5. The Facebook Playbook Should Be a Warning for the AI Era

Andrii draws a direct line from the Facebook era to the current AI moment. Media companies were lured into pivoting to social-first and video-first content by platforms that promised reach and audience. “We were duped by Facebook and Zuckerberg in particular into making social first, video first things, and it never panned out.” The platforms closed the tap, and media was left scrambling.

The same pattern is playing out with AI. Search traffic is being displaced by AI overviews. Journalists’ work is being used to train models. And the response from some in the industry is to lean into the new platform dynamics without asking hard enough questions about what happens when the rules change again. Andrii’s read is that media needs to learn from the Facebook era rather than repeating it with different technology.

6. The Dystopian Scenario: Journalists Hired Directly by LLM Companies

The most provocative idea Andrii raises is a scenario where LLM companies bypass media outlets entirely. “There is nothing stopping an LLM company from literally cutting out the middleman, which would be the media outlet, and hiring directly the journalists.” In this model, journalists would report and write, but their work would never be published in any recognisable form. It would be fed directly into a language model, which would then use those facts to generate answers for users.

He notes that this is already happening in adjacent fields. Companies building proprietary language models hire PhD-level scientists and science journalists to write training material that “will never see the light of day in the form it was written.” Extending that model to news journalism is a logical, if unsettling, next step.

His counterpoint: media organisations that want to survive need to embrace AI on their own terms. He points to the Financial Times’ “Ask FT” feature as an example of a publication integrating AI into its own platform rather than surrendering its content to someone else’s.

7. Brand Journalism Might Be the Unlikely Survival Path

Andrii’s own career is a case study in an emerging model. He believes a good option is building a real journalistic operation inside a brand. At Dealroom, he has access to one of the most comprehensive startup data sets in Europe. He’s using it to create Dealroom News, a publication with editorial ambitions that go beyond content marketing.

How does integrity work when you’re attached to a brand? Andrii’s answer is practical: “It’s always the matter of having agreed before you start on what it’s actually going to be like.” You set the rules upfront, you agree on how to handle disagreements, and you’re transparent with the audience. In Dealroom’s case, the alignment is natural. The platform exists to serve the ecosystem, and good journalism about the ecosystem serves that mission.

It’s not a model that works for every brand. But for data-rich platforms that sit at the centre of a specific industry, it’s an increasingly credible option. And for journalists who want to do real editorial work without fighting the existential revenue battles of independent media, it offers something worth considering.

Main Takeaways

Student of PR, comms, or tech journalism? If you’re not able to listen to the whole conversation, here’s a good primer on the things Andrii shared:

Europe is gaining new tech media projects, but most are entertainment or insider-focused. The number of outlets doing actual journalism is shrinking.
AI is not the sole cause of media’s decline. Each publication’s story involves specific business decisions and structural problems that predate AI.
Commodity news like funding announcements is becoming data. AI can process the bulk of it, freeing journalists for reporting that requires human judgment.
The Substack model has real limits. Subscription fatigue and the gap between publishing tools and sustainable revenue mean most journalists will not build a living on it alone.
Media fell for the Facebook playbook once. The lesson for the AI era is to build on your own platform, not someone else’s.
LLM companies could bypass media entirely by hiring journalists to produce training content rather than published articles. This is already happening in adjacent fields.
The FT’s “Ask FT” feature represents one way publications can integrate AI on their own terms rather than surrendering content to third-party platforms.
Brand journalism can work if you set editorial rules upfront and the brand’s mission aligns with good journalism. Dealroom News is a live example.
AI is useful in the newsroom for editing, research, data visualisation, and drafting supplementary content like YouTube descriptions and chapters.
The audiovisual shift in media is real, but making good video is significantly harder than making good audio. The secret to making it work is just starting.
Feedback loops in video and podcasting are incredibly long. It’s a game of showing up, being consistent, and not expecting results to appear quickly.

🎧 This article is based on a conversation from The Runway Podcast featuring Andrii Degeler, recorded at TechChill 2026 in Riga, in collaboration with Unzip Media.

Watch the full episode for the complete chat:

Optionally, listen on Spotify: