AI agents are rapidly turning product leaders and builders into, dare I say it, mini-CEOs. The traditional relationship between headcount, capabilities, and shipping velocity is fundamentally changing.
In a world where building software is becoming easier, faster, and cheaper, the competitive advantage is no longer simply how much your team can ship. It’s increasingly about judgment: making exceptional choices about what to build, what matters, and what to completely ignore.
And that’s a theme you’ll find throughout the nine things that got my attention this past month.
We’ll look inside Anthropic’s S-1, explore what it takes to succeed as a PM in the AI era, consider the possibility of one-person companies, revisit the fundamentals of finding product-market fit, and see what Duolingo and Y Combinator can teach us about marketing and launching products.
Different stories, but a common question: When building becomes easier, what actually becomes more valuable?
Nine things. Let’s get into it.
1/ Anthropic’s S-1 tells us a lot about where AI is going
Anthropic filed its S-1 in October, revealing some truly wild numbers.
The doc showed $16.3 billion in revenue in just the first six months of 2026 (a figure that climbed to $33.24 billion by the end of September). This explosive growth is backed by more than 11,900 enterprise customers spending at least $100,000 annually, including over 1,900 customers spending $1 million or more.
For context, the company had approximately 200 customers spending at least $100,000 annually at the end of 2024. That’s a ridiculous customer expansion in less than two years, especially for a company that was only founded in 2021.
“This seems to be the greatest business in the history of entrepreneurship… They’re aiming to raise, $100 billion, making it the largest IPO ever.”
— Jason Calacanis, All-In Podcast
The All-In podcast crew spent a good portion of a recent episode unpacking Anthropic’s financials, enterprise adoption, capital requirements, and the potential implications of an IPO at such a historic valuation of $1.5 to $2 trillion.
The scale and speed of enterprise adoption is an incredible anomaly. For years, traditional enterprise software companies spent considerable time and money convincing organizations to adopt new workflows. With frontier AI, we’re watching companies leap from brief experimentation to significant, multi-million-dollar spending in an extraordinarily short window.
But there’s another side to this story. The filing highlights that while top-line revenue is exploding, Anthropic faces staggering infrastructure needs. The company disclosed a monumental $895 billion in future compute and data center commitments with partners like Amazon and Google.
Because of these massive infrastructure costs, Anthropic disclosed a headline net loss of $104 billion for the first nine months of 2026, though the vast majority of that stems from non-cash accounting charges related to the surging value of its convertible notes. Even its narrower operating loss came in at $1.3 billion for the same nine-month period.
AI is simultaneously becoming one of the biggest software markets we’ve ever seen and one of the most capital-intensive businesses ever built. These enormous infrastructure commitments, absolute dependence on compute capacity, and fierce competition between frontier models create structural risks that traditional SaaS companies rarely had to navigate at this scale.
That tension is worth paying attention to, especially as founders, executives, and product leaders increasingly build their own businesses on top of these models.
Watch: All-In Podcast — Anthropic S-1 Shocker