By Rachel Waterhouse, CEO, Australian Shareholders’ Association
Most investors expect a prospectus to warn about competition, financial losses and uncertain growth. Anthropic goes considerably further, acknowledging concerns that sufficiently advanced artificial intelligence could pose “catastrophic or existential risks to humanity” without appropriate safeguards.
That is an extraordinary risk for a company to put before prospective shareholders. But for investors attracted by the growth of AI, it also raises a familiar question: what should you understand before putting your money behind the story?
Anthropic, the company behind Claude, combines extraordinary technological promise, rapid financial growth and an unusual distribution of corporate power.
The attraction is easy to understand. AI is moving from experimentation into everyday business use, with the potential to improve productivity, accelerate innovation and reshape how companies operate. Anthropic is positioning Claude at the centre of that shift.
You can believe AI will transform the economy without concluding that every AI company is a good investment at any price.
Anthropic develops Claude’s AI models and products, increasingly aimed at performing complex work. If these tools improve productivity, customers have good reasons to keep paying.
For investors, the next question is how much of that economic benefit Anthropic can capture. Customers may retain much of the productivity gain, competition may push prices down, and developing increasingly capable models requires enormous continuing investment.
Growth is not the same as value
Anthropic’s revenue reached US$33.24 billion in the nine months to September 2026, compared with US$2.55 billion a year earlier. Its operating loss narrowed from US$5.22 billion to US$1.26 billion. Computing and infrastructure expenses alone reached US$18.81 billion over the latest nine-month period.
Those figures demonstrate remarkable expansion and improving operating performance. They do not answer an investor’s most basic question: what is this business worth?
A valuation reflects expectations about the future. The more ambitious those expectations, the less room there may be for disappointment. Relevant considerations include sustainable profitability, funding requirements, competition and the cost of remaining at the technological frontier.
Growth, governance, and control
Under the governance arrangements set out in its IPO prospectus, Anthropic’s Long-Term Benefit Trust generally has the right to elect and remove directors constituting a board majority. A founder entity is allocated 50.1 per cent of voting power on matters submitted to shareholders generally.
As a public benefit corporation, Anthropic must also balance shareholders’ financial interests with those materially affected by its conduct and its stated public benefit purpose.
The structure is designed in part to protect Anthropic’s long-term mission from short-term commercial pressure. There is a credible argument for doing so. Developing powerful AI responsibly may sometimes require patience, including decisions that constrain near-term revenue.
For investors, the question is whether the alternative accountability mechanisms are sufficiently strong.
Do directors have the skills and independence to challenge management? Do incentives support durable performance? Who decides when commercial opportunity conflicts with safety?
These are questions for scrutiny, not conclusions that Anthropic has failed those tests.
Its own risk disclosures nevertheless raise the stakes. A safety policy needs authority, resources and a willingness to enforce it. Who can stop deployment? How does the board test management’s assurances? Where does human accountability ultimately sit?
What should shareholders expect?
These issues extend well beyond Anthropic. As AI moves deeper into companies, shareholders may reasonably ask boards where it is being used, what decisions it can influence, what information it can access and where human judgement remains essential.
And there is another question worth watching: does AI strengthen accountability or make it easier to diffuse it?
When an automated system contributes to a poor decision, “the technology did it” cannot become an acceptable explanation. Someone still chose the system, approved its use and decided how much authority to give it.
That is why Anthropic is such an interesting investment case. The technology may be new and the potential enormous, but the disciplines investors use to understand a company remain remarkably familiar.
For shareholders, understanding the business means looking beyond the technology: at the financial expectations, incentives, shareholder rights and ultimately who has the power to make decisions.
AI may change what companies can do.
It should not change our expectation that those with power can be held accountable.