So Close It Hurts

How Appia Bio collapsed right at the clinic door.

Author

Nicolás Cerdeira
August 28, 2025

Hey — It’s Nico.

Welcome to another Failory edition. This issue takes 5 minutes to read.

If you only have one, here are the 5 most important things:

This Week In Startups

🔗 Resources

Your next job will require AI skills.

The workflows turning one engineer into ten.

The physics of sales and 11 signs you’ve got it backwards.

📰 News

Nvidia reports record sales as the AI boom continues.

Anthropic launches a Claude AI agent that lives in Chrome.

Google Translate takes on Duolingo with new language learning tools.

💸 Fundraising

M0 raises $40M Series B as VCs pile into stablecoins.

Stealth AI startup Aurasell raised $30M in 28 hours to take on Salesforce.

German strike drones startup Stark has raised $62m.

Fail(St)ory

The cure that never left the lab

A few days ago, Appia Bio shut down.

The announcement from the CEO was blunt and sad: they had run out of money right before their first human trial. Five years of work ended not with bad science, but with an empty bank account.

What Was Appia Bio:

Appia started in 2020 with a bold idea: make cancer cell therapy something you could pull off the shelf.

Current treatments are personalized, built for each patient, and almost impossible to scale. Appia wanted to flip that. They developed the ACUA platform to engineer immune cells in bulk, freeze them, and have them ready to go. It was a way to bring cell therapy to more patients at lower cost.

Appia wasn’t a scrappy garage project. They had $52 million in Series A funding, a Nobel laureate on the board, and soon after launch, a partnership with Kite Pharma worth up to $875 million. For a young biotech, it was as strong a start as you could hope for.

But in May 2024, the deal ended. No public explanation. Partnerships fall apart all the time in biotech, but when they do, it leaves a scar. Investors notice. Fundraising gets harder.

By mid-2025, Appia was close to filing an IND—the FDA approval you need to start human trials. That’s the moment where years of lab work finally touch patients.

But trials are expensive, and without new capital, they couldn’t take the step. They had reached the cusp of the clinic and then ran out of runway.

The Numbers:

  • 📅 Founded in 2020

  • 💰 $52M raised in Series A

  • 🤝 $875M Kite Pharma deal 

  • 🚪 Shut down August 2025, just before clinical trials

Reasons for Failure: 

  • Cash vanished at the worst possible moment. Appia pushed hard to file its IND in mid-2025—an essential step to actually treat patients. That’s when the bills explode. They got within arm’s reach of trials, and then the money tanked.

  • The Kite Pharma deal ended, and credibility followed. That partnership was a major signal that Appia’s tech mattered. When it expired in 2024, it wasn’t just revenue that disappeared—it was investor confidence. Losing big-pharma validation makes future funding rounds feel much steeper.  

  • Biotech funding froze mid-year. Early-2025 started with optimism, but by Q2 venture capital for biotech had plunged from $7 billion to $4.8 billion—among the lowest in years. First-time financings collapsed too, from $2.6 billion to $900 million. In that squeeze, even promising pre-clinical players get squeezed out.

  • The entire biotech sector entered a harsh winter. It wasn’t just Appia. In 2024, nearly 40% of biotechs had less than a year of cash left—the worst burn-rate metric in six years. Layoffs, shutdowns, and deal cancellations became daily headlines. Appia’s closure belongs to a pattern—not a fluke.

Why It Matters: 

  • Losing a pharma partner can be fatal. The end of the Kite deal didn’t just cut off a revenue stream—it erased external validation that future investors needed. Without that signal, raising capital became nearly impossible.

  • Near-clinic startups are the riskiest place to be. Appia got within weeks of an IND, but that milestone meant nothing once the cash ran out. In biotech, “almost there” has the same value as “nowhere.”

  • The funding freeze is reshaping the map. Appia’s collapse shows how quickly capital markets can erase even well-funded, well-connected companies. It’s not only the weak players going under—the squeeze is hitting ambitious, technically strong teams too.

Trend

Perplexity Revenue Share

A couple days ago, the Wall Street Journal reported that Perplexity is rolling out a publisher revenue-share tied to its new Comet Plus subscription. Translation: AI search finally puts real money on the table for the articles it summarizes. Not a vague “we’ll send traffic,” but a check with your name on it.

Why It Matters:

  • Traffic is collapsing. When Google shows an AI summary, people click links about half as often (8% vs. 15%), according to Pew Research. A recent Digiday piece says premium publishers are seeing 1–25% referral drops since AI Overviews launched. That’s existential.

  • It sets a new baseline for negotiations: Publishers can now say, “Don’t just license training; pay for ongoing usage.” If Perplexity shows this can scale, others will have to explain why they won’t.

  • It’s a test of whether AI can be a real distribution business—not just a scraper with lipstick. If Comet Plus grows, this becomes material revenue. If the browser stalls, the pool stalls. Reality will decide.

What Perplexity Just Did

Perplexity created a $42.5M pool and promised 80% of Comet Plus subscription revenue to publishers. Payouts aren’t based on vague traffic referrals but on specific usage signals: when a publisher’s site is opened in Comet, when its work is cited in an AI answer, or when it powers an agent task. The idea is to connect money directly to how content is used inside the product.

All of this is tied to Comet, Perplexity’s AI-native browser launched in July, which blends search, reading, and task automation. Comet is where the “usage” happens—and where Perplexity can track enough data to make payouts work.

At the same time, this expands on Perplexity’s 2024 Publisher Program, which only shared ad revenue from a special “related questions” box. Back then, it brought on partners like Time, Fortune, Der Spiegel, The Independent, and Blavity. This new model covers much more ground and shifts from advertising to subscription-funded revenue.

The Bigger Context

This wasn’t rolled out in a vacuum. A New York judge just denied Perplexity’s attempt to move or dismiss the Dow Jones lawsuit, which means they’ll have to fight it in court. And the press is full of complaints about AI models reproducing articles nearly verbatim. A revenue-share pool is partly a way to cool down that heat.

It also sets Perplexity apart from competitors. OpenAI has struck big licensing deals with The Financial Times, The Washington Post, and News Corp, but it hasn’t promised ongoing usage-based payments. Google is talking to around 20 publishers about AI licensing pilots for Gemini/AI Overviews, but again, not usage-linked revenue. Anthropic just settled a major class action with book authors—its energy is going into minimizing legal exposure, not building payout rails.

Reactions and Open Questions

Early partners in Perplexity’s ecosystem (like Fortune and Time) are supportive—they helped shape this plan. But others are staying in court. That split tells you how divided publishers are: some want cash now, others want to hold out for better terms or stronger protections.

The open questions are obvious. Will publishers actually get auditable logs of when their work was cited or used, or will payouts remain a black box? Can $42.5M scale into something meaningful across hundreds of partners, or is this just a pilot designed for headlines? And how long until content creators start building “AI-SEO” strategies to maximize citations and revenue?

Help Me Improve Failory

How Was Today's Newsletter?

If this issue was a startup, how would you rate it?

Login or Subscribe to participate in polls.

That's all of this edition.

Cheers,

Nico