The artificial intelligence revolution is moving at breakneck speed, and if you are trying to value tech stocks right now, the ground is shifting beneath your feet. In a revealing new interview, Anthropic CEO Dario Amodei pulls back the curtain on the exponential growth of AI, the massive compute crunch, and what it all means for the future of software. This episode is an absolute must-listen for retail investors trying to separate the real, long-term value generators from the fleeting hype.

When evaluating the tech landscape, the insights from the very people building the foundation are invaluable. Amodei’s interview isn't just about the science of AI; it's a masterclass in shifting business models and economic disruption. Let's dive into the core lessons every value investor needs to understand right now.

Following the release of powerful coding tools like Claude Cowork, the market saw a massive sell-off in traditional software, dubbed the "SaaSpocalypse." Amodei confirms that the fundamental definition of a software moat is changing forever. If a company's only competitive advantage is the complexity of writing its code, that advantage is effectively dead.

However, this doesn't mean software is a bad investment; it means the criteria for value have shifted. The new moats are entrenched customer relationships, proprietary data, and deep domain expertise. As Amodei bluntly puts it: "If your moat is, we wrote this complex software that no one else can write, like good luck, you're not gonna be able to defend that. But I think folks have customer relationships. Folks have know how of how the field works." He predicts the software industry will actually get larger, but complacent incumbents will be wiped out.

While competitors focused on flashy, consumer-facing chatbots, Anthropic deliberately targeted the enterprise sector. For a value investor, this is a crucial distinction. Consumer AI often relies on an advertising-driven model that prioritizes engagement and addiction—a model fraught with regulatory and ethical landmines.

Enterprise AI, on the other hand, is built on long-term contracts, systemic trust, and solving tangible economic problems like drug discovery or energy efficiency. This aligns Anthropic’s business model perfectly with its safety-first values, creating a more sustainable revenue base. "Consumer can have this, you know, almost this gimmicky aspect to it," Amodei explains. "With enterprise... you build a relationship where you work with a company for many years, you deliver on what you say... and they basically trust you."

With rumors of a nearly trillion-dollar valuation for a five-year-old startup, it's easy to cry "bubble." But Amodei provides a rational, math-driven explanation for these staggering numbers: the sheer cost of computing power required to stay on the frontier. Anthropic planned for a 10x annualized growth in compute, but saw a massive 3x growth in revenue in a single quarter—which equates to an 80x annualized pace.

Raising massive rounds of capital from giants like Amazon and Google isn't about padding the balance sheet; it's a required buffer to secure limited physical infrastructure. As an investor, it's important to realize that the fundamental business looks strong, but the capital expenditures required to stay ahead are astronomical. > "Raising money is kind of the buffer against this cone of uncertainty. So it's a totally rational thing to do. It's a very small dilution to the business."

Anthropic was one of the first AI companies to sign a contract with the US Department of Defense, a move that surprised some given Amodei's historically anti-war stance. The rationale is highly pragmatic: preventing authoritarian regimes like China and Russia from dominating the AI landscape.

However, they didn't just hand over the keys. Anthropic drew strict "red lines," outright banning the use of its models for mass surveillance or fully autonomous weapons. This selective government partnership shows a mature approach to scaling a defense-adjacent business without compromising core brand values. > "We should use this technology in every way except the ways that undermine our own values... It's not worth democracies winning if democracies do those things."

Perhaps the most shocking revelation is regarding Anthropic's "Mythos" model. It is so exceptionally good at finding and exploiting zero-day vulnerabilities in code that early testers called it a "super weapon." Instead of releasing it to the public for a massive PR and commercial win, Anthropic held it back, taking a massive financial hit to prioritize global cybersecurity.

They are currently using Mythos quietly with defenders to patch the internet's vulnerabilities before bad actors can exploit them. This highlights a critical risk factor for investors looking at the broader tech space: the threat landscape is about to expand exponentially. "We have suffered enormously commercially from not releasing this model," Amodei admits, but notes that they are trying to create a world where "hopefully six months or a year from now, we have a much more secure internet ecosystem."

Amodei doesn't shy away from the scary math, maintaining that there is a 10 to 25% chance of civilizational collapse due to rogue AI. Because of this, he believes no single entity—not a corporation, and not even the government—should have unchecked power over this technology.

To counter this, Anthropic implemented a "Long Term Benefit Trust," an independent body that can actually fire Amodei and controls the majority of the board. For value investors focused on corporate governance, this is a fascinating, unprecedented structure designed to ensure long-term stability over short-term corner-cutting. > "I'm scared of companies having it, but I'm also scared of government having it. And then the companies need to provide checks on government, and the government needs to provide checks on companies."

The "SaaSpocalypse" isn't the end of software; it's a great filtering event. As AI continues to accelerate, the actionable advice for retail value investors is clear: it is time to rigorously audit your tech and software holdings. Do the companies you own rely solely on the complexity of their code as a moat? If so, they are incredibly vulnerable. Pivot your focus toward businesses with deep, proprietary domain expertise and sticky, long-term enterprise relationships. The companies that use AI to make their workforce more productive—without losing their core human relationships—will be the ultimate winners in this expanding pie.

For more of my insights on this topic, be sure to follow me.