The Anthropic IPO is on! We looked under the hood, here are the Red Flags...

On the surface, there’s plenty to like about Anthropic. 

• First-half 2026 revenue growth of 14x.
  
• Gross margins of 85%, up from just 12% a year earlier. 
  
• A private valuation that has risen 16x since March 2025, and $65 billion raised in its latest round. 
  
• Not to mention Anthropic is at the forefront of what could be the most consequential technological revolution in human history.

So, it's easy to see how this is set to potentially be the largest IPO in history. What a time to be alive. 

But our investment team has spent a lot of time on the business in recent weeks looking under the hood, and this is where the concerns start to pile up quickly. 

Biggest Anthropic Red Flags

Before I go on, I should emphasise we’re style agnostic - we invest across value, growth and quality - so we don’t immediately ignore or screen out these big IPOs just because we're anti-hype or because we are instinctively against anything popular. 

But there are striking red flags that provide significant cause for concern. These are the ones that really stand out to me. 

Locked-in costs

At the top of the pile is roughly $895 billion in committed future spending, overwhelmingly related to compute. That's a remarkable long-term financial commitment against a revenue base where 79% is consumption-based, rather than traditional subscription revenue. While some customers have minimum spending commitments, future demand remains inherently uncertain.

Back to losses

The company swung from a $974 million operating profit in the second quarter to a $567 million loss in the third. 

Potential junk rating

Blackstone now expects the debt to be rated BB - junk, not investment grade. 

TAM hype

Anthropic's prospectus points to more than $40 trillion in global economic activity associated with knowledge work. That's larger than the entire US economy. But there's a world of difference between the economic value of work AI could transform and the revenue AI companies can actually capture.

Porous moat

There are structural issues too. Stanford University's 2026 AI Index Report found that the best open-weight model (often free to download and use), was only 3.3% behind the leading proprietary model.

Talent costs and leakage

Personnel costs have risen almost fivefold in a year. Why? Because AI researchers hold the precious intellectual  property that makes AI possible in their grey matter. Matt Deitke, a 24-year-old researcher is reportedly earning US$250m over 4 years at Meta. Andrew Tulloch a former OpenAI researcher was reportedtly offered US$1.5b over 6+ years although Meta disputed that figure.

Founder control

Anthropic has established a special entity called Founder LLC, controlled collectively by its seven co-founders. Founder LLC holds a single Class F share carrying 50.1% of shareholder voting power. This means the founders can retain control even if they sell down much of their economic ownership.

No AI payoff, yet

As highlighted by Apollo there remains big question marks about measurable AI profitability gains across the S&P 493 (the S&P 500 minus the Magnificent 7).

Cost per Unit of Intelligence is Crashing

The cost of AI intelligence is falling at an extraordinary rate, with some estimates suggesting a 13-fold annual decline. While this is wonderful for consumers, it raises serious questions about pricing power. If those cost reductions are passed through to customers, AI usage must grow enormously just to sustain revenue growth.

And... we all may die

And in its own prospectus, Anthropic warns of “existential risks” to life as we know it. So there is that...

All up, the team found 14 Red Flags in total. The full list can be seen here:

AI: 'Asymmetry Inverted'

I don’t know whether Anthropic will prove to be a wonderful investment or an expensive lesson. Early reports are that the book is already covered so we would expect it to trade well in the very short-term. But after 25 years in markets, I’ve learned that avoiding disasters can be just as important as finding winners.

For us, the job of a fund manager is to protect investors from outsized downside risk - as I've said before, our risk systems are designed to act as the proverbial flak jacket for our portfolios and ultimately our clients hard-earned capital. 

At the risk of sounding cliche, Warren Buffett put it simply: Rule number one is don’t lose money. Rule number two is don’t forget rule number one.

If you would like to learn more about investing in our global investment fund, The Plato Global Alpha Fund, click here to visit our website or search for the ASX ticker 'PGA1' wherever you invest.  

The Plato Global Alpha Fund has returned +23.7% p.a. after fees since inception on 1 September 2021 (as at 30 September 2026).*

Managed Fund

Plato Global Alpha Fund

Global Shares

View

ETF

Plato Global Alpha Complex ETF (PGA1)

Global Shares

View

*Past performance not reliably indicative of future returns. Read PDS and TMD before investing available at plato.com.au. All investments carry risk.