The Business Models Keeping the News Afloat Amid Uncertainty

From nonprofit conversions to creator-led newsrooms, NPP looks at the news business models that are mostly working (for now), what could work, and what didn't work.

This newsletter is free and will stay free. If it’s useful to you, a paid subscription is the best way to keep it going. —Luis G.

Last month, the Venetoulis Institute for Local Journalism — the nonprofit parent of the Baltimore Banner — announced it would buy the Pittsburgh Post-Gazette from Block Communications, preventing the 240-year-old paper’s scheduled closure on May 3. The deal gives the Venetoulis Institute a second market, a newsroom of 100 journalists, and 60,000 Post-Gazette subscribers to add to the Banner’s existing 80,000. What it gives the Post-Gazette is less clear. The Baltimore Banner, the model this deal is built on, has not broken even.

Block Communications, which had operated the Post-Gazette since 1927, acknowledged losses exceeding $350 million over 20 years. The Post-Gazette was built on print advertising, a revenue structure that stopped working. The nonprofit, subscription-driven model replacing it isn’t new either.

The Baltimore Banner, the Texas Tribune, the Salt Lake Tribune, and ProPublica are all running the same playbook. And the Post-Gazette acquisition is more proof of concept than breakthrough.

What remains to be seen is whether models like these and others the news industry is currently testing can scale and, more importantly, whether they can be replicated.

The industry got here through compounding failures. Craigslist cost local newspapers an estimated $5 billion in classified advertising revenue between 2000 and 2007. The 2008 recession accelerated cuts newsrooms never recovered from.

Then came a decade of hedge fund consolidation — Alden Global Capital, Tribune Publishing, MediaNews Group — in which the remaining assets were harvested rather than rebuilt. The pivot to digital produced banner ads that didn’t pay and a bet on Facebook video that evaporated when the algorithm changed in 2018, dropping referral traffic to publishers by nearly 40%.

Now Google’s AI Overviews are doing the same to search — organic Google search traffic to news publishers dropped 33% globally in the year ending November 2025, according to Reuters Institute data. Each round left fewer resources and less time to find something that works, which is why the testing ground is crowded now.

Scaling news is hard for reasons that don’t change with the business model. Journalism is labor-intensive. A story requires a reporter with relationships, institutional knowledge, and enough time to get it right. Local news compounds the problem because what works in Baltimore doesn’t transfer to Pittsburgh without rebuilding from scratch — staff, sources, community trust, subscriber base. There’s no franchise model for accountability reporting; the product can’t be templated.

Which brings me to this edition of this newsletter. The news organizations in this edition are running experiments with real financial consequences if they fail. Whether any of them can be copied is the question the industry doesn’t yet have an answer to.

---

Too many models, not enough proof

The news industry is running a simultaneous, largely uncoordinated experiment across every possible revenue structure: subscriptions, philanthropy, events, branded content, sports betting partnerships, AI licensing deals, creator-publisher hybrids, reader revenue gamification, paywalled podcasts. The majority of news organizations are now relying on three or four different revenue streams at once. The list keeps growing but the financials haven’t caught up yet.

Some models have documented outcomes worth taking seriously. (More on that below.) Others are working under specific organizational conditions that don’t transfer cleanly — and the journalism press tends to move from case study to case study without asking whether the conditions are replicable.

What’s new about these models is mostly the arrangement, not the ingredients. The revenue streams — subscriptions, advertising, philanthropy, grants — have been around for a number of years. What’s changed is which ones a newsroom bets on, and in what order.

In some instances, the problem amounts to a proliferation problem. A recent analysis of nonprofit newsrooms found that one of the biggest structural barriers to growth is the absence of staff focused on business development. Newsrooms need those operators to scale but can’t fund them. You can replicate the editorial model but you can’t replicate the sustainability without the people whose job it is to build it.

And I think that holds true even beyond nonprofit newsrooms. The gap between a working model and a replicable one runs through that exact shortage. Testing more models without solving the operator problem doesn’t compound progress, rather it multiplies the experiment count without producing any replicable results.

---

The wreckage of scale-first, platform-dependent media

The clearest model failure on record is the VC-backed digital media cluster — BuzzFeed, Vice, Vox, Huffington Post, etc. — that spent a decade raising millions on the promise that massive social reach would translate into advertising revenue. And, well, it didn’t.

Disney invested more than $400 million in Vice and later wrote the whole thing off. NBCUniversal put comparable sums into BuzzFeed. BuzzFeed went public via SPAC in December 2021 at $10 per share and was soon trading at around $1 — a precipitous 90% loss of value. Vice filed for bankruptcy. BuzzFeed is still standing, but just barely; it recently sold a controlling stake to media entrepreneur Byron Allen this month for $120 million.

BuzzFeed’s entire bet was built on Facebook referral traffic, not on readers who chose to come back on their own terms. When Facebook deprioritized news, the model collapsed.

And the platform dependency problem didn’t stop there.

Google search traffic to publishers fell by a third globally between November 2024 and November 2025. Facebook referrals to news properties are down 67% over the past two years. Twitter/X referrals are down 50%.

Publishers who built audience infrastructure around any of these platforms now face the same structural exposure that sank BuzzFeed. Whether they’ll admit it and change course is an entirely different question.

---

The nonprofit model works, but here’s what’s really holding it together

The nonprofit model is the most tested alternative, and the data backs it up. According to the 2025 INN Index, the nearly 400 digital-first nonprofit newsrooms in INN’s membership generated a combined $650–$700 million in revenue in 2024, a 14% increase from 2023. The median member outlet took in $532,000 — up from $477,000 the year before.

The Texas Tribune is still the most-cited case. Founded in 2009 with a single $1 million investment, it grew into a newsroom running on a roughly $13 million annual budget — built across more than 13,000 individual members, over 1,100 corporate sponsors, foundation grants, and an annual festival that drew 9,000 attendees and 1.1 million online session views in 2022 alone. The revenue diversification is real, and it holds up at scale.

What’s harder to replicate is how they got there. The Tribune was built around Austin’s political beat, founded by a well-connected venture capitalist, and run for 13 years by a CEO who had spent 18 years before that at Texas Monthly. Those conditions aren’t a model — they’re a specific set of circumstances that realistically could not replicate in every city in America.

For most nonprofit newsrooms, the revenue diversification the Tribune built over 13 years is still a goal, not a reality. The field runs on foundation grants, and INN’s own numbers show how dependent it remains on that single source.

Foundation funding still accounts for 49% of member revenue — down from 57% in 2017, but still the single largest source. Individual donations add another 32%, which then leaves 18% from earned revenue — events, sponsorships, advertising. The target in nonprofit news is roughly equal thirds across all three. Across INN’s membership, the field isn’t close.

The Pittsburgh Post-Gazette rescue is not a model. Within days of taking over, the Venetoulis Institute cut roughly half the newsroom including, according to the union, 80% of the journalists who had gone on strike. Stewart Bainum and his wife committed an additional $30 million over five years to fund the turnaround — on top of the $50 million already committed to the Banner, which still hasn’t broken even.

That’s one person’s philanthropic capacity deployed across two markets. The Venetoulis Institute’s move keeps a 240-year-old paper alive in a single city. It doesn’t tell us what happens in the next 100 cities without a Bainum equivalent willing to write the check.

---

The creator-led newsrooms come small by design

The most visible alternative to the institutional model is the journalist-owned outlet — small newsrooms built by reporters who left larger organizations, funded almost entirely by readers who chose to pay.

Defector Media reached $4.6 million in annual revenue in 2024 with roughly 40,000 paying subscribers and has been profitable since its first year. The site launched in September 2020, nearly a year after its staff resigned from Deadspin in protest when its private equity owner told them to stick to sports.

That resignation was public and dramatic, and it generated the kind of goodwill that converted directly into founding subscribers. Defector launched with 36,000 paying readers before publishing a single post, ended its first year with $3.2 million in revenue, and has operated without outside investors since. Worker ownership kept the editorial decisions with the journalists and the cost structure low. Subscriptions — not advertising — were the business from the start.

404 Media followed the same logic at a smaller scale. Four former Motherboard journalists launched the site in August 2023, three months after Vice filed for bankruptcy, each putting in $1,000 to cover startup costs.

The four had spent years building credibility covering tech and internet culture at Motherboard where the audience already existed, and the Vice bankruptcy gave readers a reason to follow them. The site was profitable within six months.

Overhead at 404 Media has been minimal. It started with a Stripe account and the Ghost web-hosting platform. When the founders discovered their stories were being scraped and republished by AI content mills, they put their content behind a registration wall — not just as a defensive move, but as a subscriber acquisition mechanism. Readers had to identify themselves to keep reading, and some of them converted.

Both newsrooms succeeded for the same underlying reason, which is that neither started from zero. They had audiences, reputations, and a defined subject matter before they launched. The model worked financially because the pre-conditions were already in place.

The subscriber numbers, though, don’t make an obvious case for replication.

Defector’s subscriber count has floated between 40,000 and 42,500 for two years, peaking each summer before retreating after the site’s September renewal cycle. By 2024, the pattern hadn’t changed, and new subscriber acquisition was underperforming projections most months.

404 Media hasn’t disclosed subscriber numbers publicly. Its founders have declined to share data in every interview on record.

These models work at the scale they’re built for. What they don’t answer is what happens when you try to run the same structure in a market that needs 50 journalists instead of four.

---

The B2B-premium model has the right audience, but the wrong math

A different bet is the B2B subscription model — high-priced, professionally targeted, designed to be expensed rather than paid out of pocket.

Axios Pro launched in January 2022 at $599 per newsletter per year and generated $2 million in its first year with more than 3,000 subscribers. Co-founder Roy Schwartz had said he wanted Pro to eventually represent 50% of Axios’s total revenue. In 2022, against total revenue of just under $100 million, it only contributed about 2%. Cox Enterprises acquired Axios later that year for $525 million.

The clearest existing proof of concept is Politico Pro, which launched in 2010 as a side project and grew into a 200-person operation accounting for roughly half of Politico’s total annual revenue — with a 90% renewal rate and 30% year-over-year growth over three years. A five-person subscription ran $8,000; the company’s VP later said they had underpriced themselves at launch. Axel Springer acquired Politico in 2021 for a reported $1 billion.

The Information runs on similar logic at a smaller scale — about 60 employees, no advertising, individual subscriptions at $399 per year, with a Pro tier at $999 that includes proprietary datasets and org charts. Founder Jessica Lessin built it for professionals willing to pay for reporting they couldn’t get anywhere else.

What those three have in common is that the content is priced as a professional necessity. The ceiling on that model is the ceiling on the audience willing to expense it.

Business professionals with expensed subscriptions churn less than individual readers — that’s the real appeal of the B2B model. But the audience is narrow by design, the price point has to compete against Bloomberg, and any corporate budget cuts can hit these products first.

A business professional’s expensed subscription is one of the easier line items to cut when a company tightens its budget — a risk flagged at Axios Pro’s launch by industry analysts. Another player in this field, Morning Brew, found out the hard way in 2022 because despite running a B2B subscription business, it lost enough of its ad revenue base to cut its staff twice that year.

B2B subscriptions improve retention. They don’t insulate a business from everything else.

---

What’s actually working and what to look out for

Events are the clearest documented revenue bright spot across multiple models. Semafor — the digital news outlet that launched in 2022 — reports that live journalism constitutes about 50% of its revenue. Its World Economy Summit, held in 2025, drew nearly 3,000 attendees including more than half of Fortune’s 500 Global CEOs.

Forbes, after transitioning to a hybrid event model, saw a 60% increase in event revenue with sponsorships accounting for 35% of its advertising revenue. And the Texas Tribune’s annual festival has been an anchor since the outlet’s earliest years. What these examples provide is proof that events create direct audience relationships, real-time brand value, and sponsor revenue that doesn’t depend on platform algorithms. And for that reason, other newsrooms have followed suit and experimented with live journalism events.

Looking beyond the live events model, the creator-publisher hybrid is the newest structural bet of the current cycle. The Washington Post’s Ripple project — which offers writers, including independent Substackers, a revenue-sharing arrangement to publish on the Post’s platform — has emerged as one of the more consequential institutional experiments underway. The arrangement is an admission that individual journalism brands have audience value the Post can’t build internally.

But the risk is there, and the Post learned this lesson when Dave Jorgenson — the Post’s “TikTok journalist” — spent eight years building his audience inside the Post’s brand, then left to launch Local News International independently. The Post’s YouTube channel views dropped sharply after he left. The institution had been renting an audience it hadn’t built, and the creator left with it.

Newsrooms investing in creator-style talent face the same problem — help journalists build their personal brand and risk losing them, or don’t, and lose relevance.

---

Great, but who fills the gaps elsewhere in America?

Every model in this edition starts with the same first step — finding people willing and able to pay for the news product. That’s not an argument against subscriptions. It’s a reminder that a working business model and a solved public interest problem are not the same thing.

Take into account that 213 U.S. counties now have zero local news sources, and another 1,524 have only one remaining outlet. Taken together, roughly 50 million Americans have limited or no access to local news. On top of that, less than 10% of digital-only news startups are located in rural counties, and new outlets are appearing in areas that are more affluent and better-educated — nearly the opposite demographic profile of the communities most affected by closures.

Defector works for 40,000 sports fans. 404 Media works for four journalists and a technology-literate readership. Axios Pro works for professionals with expense accounts. The Texas Tribune works for people who follow Texas politics closely enough to attend a festival. These are real businesses serving real audiences, yet none of them are substitutes for the paper that covered your county commission before the hedge fund arrived.

The models being tested right now are mostly legible to the people already paying attention — to journalism funders, media reporters, and readers of this newsletter. They are less visible to the communities that lost a paper in 2019 and haven’t had one since. The people most harmed by the collapse of local news are the least likely to be reached by any of the experiments currently generating the most enthusiasm.

In my view, that’s the question the industry keeps putting off. Not which model works, but which model works for whom — and at what point does a market get written off as unserved by design. The proof of concept era has produced some durable newsrooms. It hasn’t produced an answer to that question, and most of the models don’t try.