Early in the week, late one quiet night, I spotted a mother bear and her cub wandering Vail Village.

By the next morning it had become the standard greeting among 40 of the most senior chief executives in the news media business. Not “how was your flight.” Not “how are subscriptions.” Just: “Did you see a bear, too?”

Some had. Some had not and were faintly aggrieved about it. Some were, I suspect, embellishing. The bear was sighted, re-sighted, disputed, and by the end of three days had acquired a status somewhere between mascot and running joke. We declared the 2026 INMA CEO Roundtable the “Year of the Bear,” and the name stuck faster than anything on the programme.

I want to start there, because it tells you more about what happened in Vail than any agenda would.

The lead to the story

The INMA CEO Roundtable is our fourth annual retreat — invitation-only, limited to the chief executives of INMA’s corporate members. This story is a peek behind the curtain.

Forty CEOs from six continents spent three days in Vail last week, and across nine sessions every INMA faculty member arrived at the same answer to the same question: What does a news company actually sell in an AI economy? Trust. Relationships. Judgment. Provenance. Presence.

The advertising research said it. The audience research said it. The newsroom research said it. The platform analysis said it. Forty CEOs agreed.

And then the room ran into the wall. Our researcher-in-residence, Greg Piechota, asked two questions. Who knows who their super users are? Few hands went up. Who can put a currency figure on what those super users are worth? Even fewer.

So here is the sentence I ended up saying in the Roundtable’s final hour and the one I will be repeating for the next year:

We agree on what we are selling. We have yet to prove what it is worth.

Our advertising lead, Gabe Dorosz, put the same idea in the frame that stuck to everything else in the room: The news industry has spent a year calling trust its moat, and the task now is to turn that moat into a toll. Platforms get paid because they prove what they deliver. Unmeasured trust is a moat with no toll on it.

I opened the Roundtable arguing that the market has moved from an abundance of content to a scarcity of verified trust and offered a leadership frame for the next two to four years: defend, automate, invest, kill. Defend the human core. Automate the plumbing. Invest where scarcity compounds. Kill what is no longer essential.

What was interesting was watching my own frame get taken apart and put back together by the faculty who followed:

• Our product and technology lead, Jodie Hopperton, turned it into “stop trying to do more AI, redesign the work instead.”

• Our newsroom innovation lead, Sonali Verma, turned it into artifacts becoming cheap while journalistic processes become valuable.

• Greg, talking about the competitive advantage in the AI era, landed on context, not the model, as the real moat.

• Our young audiences lead, Kerstin Hasse, turned it into habitat rather than bridge.

• Our platforms lead, Robert Whitehead, turned it into a negotiating posture: Stop positioning as a proud news brand and start positioning as sovereign national intelligence.

By the third morning, six different sessions in six different vocabularies had converged on one instruction: Build products that travel with the customer.

The kill column, for the record, was the hardest for the room to fill in. It always is.

The report is not the point

I have now written the official Roundtable report, which runs to 31 pages and went to the 40 CEOs who were in the room. It captures the essence of the room without the names in it. It may be the most useful document INMA produces all year.

Across the nine sessions we ran 25 interactive exercises — role-plays, structured debates, written exercises, live polls, whiteboard votes, table-by-table scenarios — because nobody flies to Colorado to be presented at.

It also cannot capture what actually happened in Vail.

Because the most valuable material of the week was off the slides entirely.

A chief executive read aloud an e-mail that had landed on his phone mid-session, from his own head of AI, that began: “It is generally unclear who controls this developer’s time.” He told the room this was exactly the question he had written down at the start of the Roundtable. What followed was the most practically useful hour of the week, and none of it was planned.

Two publishers staged a live role-play as a 31-year-old creative and a 21-year-old student, and let the rest of the room pitch them a subscription. The 21-year-old’s reply — “I understand you somehow belong to an endangered species, but the problem is you address my grandparents with your content. They’re in their early seventies. I’m 21.” — did more damage to the room’s assumptions than any research slide could have.

A Nordic CEO said the most strategically uncomfortable thing anybody said all week: that we are still trying to fix distribution while the architecture beneath the Internet changes, which means we are building for something that is going to be obsolete, and we still have to bridge the gap.

You do not get those sentences at a conference. You get them from people who trust the room.

How the room gets built

And the room gets built by everything that is not the room.

We hiked. There is a moment I look for every year, and this year it came on a trail above the village, somewhere between the gondola and the treeline. I looked back down the switchback and saw a chief executive from Scandinavia walking with a chief executive from Latin America, deep in an argument about paywalls, gesturing with both hands, entirely oblivious to the aspens.

Neither of them was performing. Two people who run news companies on opposite sides of the world, discovering they have the same problem.

We bowled — and I will say only that the competitive instincts that built these companies do not switch off at the lane.

We held receptions where the conversation ran long past the point where anyone was still checking a phone. We split into small-group dinners across Vail Village, six or eight to a table, because that is the size at which a CEO stops giving the board answer and starts giving the real one.

People wandered the village together, rode the gondola together, climbed a mountain together, ate breakfast together, kept talking at midnight on a terrace at 8,000 feet. And, evidently, went looking for a bear.

There were even informal morning runs, too, for a group whose enthusiasm at that altitude I admired more than I joined.

None of that is decoration on the programme. It is the programme.

Chatham House rules only work if people believe in them, and people only believe in them if they know each other. The reason a chief executive will read out an internal e-mail from his head of AI, or admit that he cannot name his super users, or confess that his company is “a little bit of bedlam right now,” is that on Tuesday morning he walked up a mountain with the person sitting across from him.

There is no other gathering like this in the news industry. No vendors. No sponsors on stage. No audience. No press. No performance. Forty peers, three days, and an agenda that they themselves set.

Which is really just INMA doing what INMA does. Personal connections. Direct connections. A constant feedback loop running from the room to the research and back into the next room, capturing the direct voice of media leaders — unmediated, unpolished, occasionally contradictory, and far more useful for it.

We exist to help our members get things done. There is no politics in that and no communiqué at the end of it.

That loop is also why what happened next did not surprise me as much as it should have.

Before the first session, every CEO wrote down the one topic they wanted clarity on by the end. Sorted and read back to the room, the list was: AI product and technology, distribution in an AI world, legal and licensing, young audiences, monetisation. A near-exact match with the programme we had built.

Four years ago, this same group overrode our agenda entirely and spent the Roundtable on newsrooms — a subject that appeared nowhere in the programme.

That gap closing is itself a finding. The news industry’s agenda has narrowed and sharpened. Everyone is working the same problem now.

What we carry out

Two things are worth carrying back into the business.

Trust is our moat, and a moat is a cost centre until someone builds a toll on it. That work is measurement, data, and proof, and it needs funding and an owner with a name.

And whatever we build has to travel: the journalism, the brand, and — if our product & tech and young audiences leads get their way — the subscription itself.

The clock the room set for itself is short. In the next six to 24 months, publishers are either paid suppliers or free raw material.

Asked in the closing session what the news industry most needs to invest in, the answers came back: Our product. Our capabilities. Our relationships. And from INMA President Siv Juvik Tveitnes of Schibsted, the last word and the one the room settled on: our people, our organisation, and our leadership.

Which is why the 2027 INMA CEO Roundtable, August 24-26, in Kilkenny, Ireland, will be built on exactly that theme.

The room chose it. The room always does.

As for the bear: I have thought about it more than I expected to. Forty chief executives, carrying the hardest set of problems our industry has faced in 30 years, and what bound them together was a shared sighting nobody could quite verify and everybody wanted to talk about.

Which, when you think about it, is roughly what we do for a living.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .