Interview

Eric Newcomer on building a media business with scoops, events, and subscriptions — and why AI can't replace journalists with real networks

Jun 17, 2026 with Eric Newcomer

Key Points

  • Newcomer's independent media business generates revenue from three streams, with event sponsorship largest, followed by subscriptions, after launching on Substack in 2020.
  • Scoops and sourced reporting from real networks remain structurally resistant to AI replication, since the information doesn't exist online to be synthesized by language models.
  • Proprietary data sets once sold exclusively to publications now flow directly to audiences through company-owned channels, unwinding the old trade where media got early access in exchange for amplification.
Eric Newcomer on building a media business with scoops, events, and subscriptions — and why AI can't replace journalists with real networks

Eric Newcomer on independent media, AI, and the venture capital landscape

Eric Newcomer built his newsletter and events business on a simple premise: write for people who already get Silicon Valley, without having to explain who Bill Gurley is. After stints at the Macon Telegraph, the Sun Sentinel, the Tampa Bay Times, the New York Times, the Information (where he was Jessica Lessin's first employee), and six years at Bloomberg, he launched Newcomer in 2020 on Substack. Today the business runs on three revenue streams, with event sponsorship as the largest, followed by newsletter subscriptions.

How the business is built

The events side grew out of an early bet on AI. In 2023, when most people were still skeptical, Newcomer put on the first Cerebral Valley AI Summit out of a Hayes Valley office space. The timing proved right: Ali Ghodsi connected with Naveen Rao, and Mosaic ML was acquired for $1.3 billion, generating press that validated the summit's brand. Newcomer is now expanding internationally, with a mid-year Cerebral Valley event in London.

Subscriptions follow editorial credibility rather than viral reach. Lists and surveys, like a wing-led ranking of up-and-coming startups, tend to stay in front of the paywall because sharing is the point. Gated content, like an LP survey on venture firm performance, converts better precisely because readers know what they're paying for.

I did write that I wanted contrarian optimism... Sometimes that's positive, sometimes that's negative. The best takes are ones where you go talk to people who know and then you're like, yeah, I talk to people... there's so much information that informs your work that is just not anywhere online. It truly only exists in a small group chat or places like that.

The AI threat to journalism

Newcomer acknowledges at least one Substack competitor ranked above him who appears to be using AI to generate content. His defense isn't the writing itself but the sourcing behind it. Journalism that comes from real human networks, from knowing what other people are actually thinking, is harder to replicate than prose style. Scoops are structurally AI-resistant because the information doesn't exist online to be synthesized.

That said, he draws no hard line on AI as a tool. He uses Claude as a copy editor, catching errors including million-versus-billion mistakes. He's also building AI into operational infrastructure: an events ticketing platform and internal story-tracking tools. His honest question is whether polishing a paragraph with AI is actually a disservice to readers, and he doesn't think it is.

The VC landscape

The conversation turns to which venture firms actually hold top-of-mind status with founders today. Newcomer names Sequoia first but is careful to separate two things that often get conflated: LP returns and founder appeal. A firm can generate enormous cash returns by writing large checks into labs at high valuations without being the firm founders want at the table in a competitive Series A.

The more interesting signal, in his view, is who wins when a founder has ten term sheets. Newcomer has run versions of this with a "Founders Choice" survey and an LP-driven ranking of managers. The conclusion he gestures toward is that firms like Thrive and Greenoaks have real momentum with founders, and that Spark Capital's early Anthropic positioning has been meaningfully good for their reputation. Benchmark he sees as rebuilding credibly, though the Travis Kalanick history creates ongoing noise.

Data going direct

One structural pressure on media is that proprietary data sets that once fed Wall Street Journal stories are now being published directly. Ramp's AI spend data is the example he reaches for: companies have realized that strong, original data is its own content marketing, and if they distribute it themselves through podcasts and social media, they don't need to embargo it to a publication to get reach. The old model was to give media exclusive early access in exchange for amplification. That trade is quietly unwinding.

The defensible position for independent journalism sits where the information genuinely doesn't exist anywhere online. It lives in small group chats, in relationships built over years, in the kind of trust that lets a source tell you something they'd never post.

Every deal, every interview. 5 minutes.

TBPN Digest delivers summaries of the latest fundraises, interviews and tech news from TBPN, every weekday.