Lessons from Anthropic’s IPO (including the risks)

We know that news suggesting imminent disasters inevitably tends to be more sensational than news suggesting future scenarios characterised by innovative solutions and opportunities for all. The rule is always the same: negative news is news; positive news is simply not news. This rule always holds true in a daily reality built on constant scaremongering, but it is particularly true when the news concerns a special world known as artificial intelligence. When negative news filters through from the world of AI (robots will be eating our brains in a few years’ time and will probably kill us all), that news has the power to dominate the front pages of the news. When non-negative news emerges from the world of AI, that news tends to be hidden, suppressed or watered down.

The most interesting case study to focus on is that concerning the launch of one of the most significant IPOs in the history of the US stock market: that of Anthropic, one of the world’s leading technology companies. Its CEO, Dario Amodei, made headlines a few days ago for the nightmare scenarios that could unfold if artificial intelligence were to spiral out of control, but he received little attention for the dream scenarios that could materialise thanks to the implementation of AI-related tools. These dream scenarios were set out in black and white by Amodei himself in the report through which the head of Anthropic is seeking to persuade investors to back the company’s IPO.

In the world of AI, expressing concern about what is being developed has become a necessary prerequisite for being seen as reassuring, and even in the proposal presented by Amodei there is considerable self-criticism directed at the sector. But beyond this, there is a detail that has escaped many people’s attention, which coincides with a long and compelling list of reasons why AI could change our lives for the better in the future. Amodei explains that AI – provided, of course, it doesn’t kill us all first – will help us in the field of medicine, speeding up the discovery of drugs, enabling new treatments for cancer, Alzheimer’s, heart disease and genetic disorders, detecting epidemics earlier, improving diagnoses and providing personalised therapies. He points out that in poorer countries, AI can offer millions of people access to the expertise of doctors, lawyers, accountants and agronomists; it can improve the distribution of medicines and help tackle malnutrition. Finally, he writes that, within the public sector, AI can simplify tax systems, planning permission processes, social security and healthcare services; it can reduce waiting times; it can improve the efficiency of the state; and, incidentally, it can help strengthen democracies by defending infrastructure against cyber-attacks more effectively than is currently the case.

Trying to convince investors (and observers too) that a technology which the developers themselves consider potentially dangerous offers more reasons to be optimistic about the future than reasons to be concerned will be no mean feat. But Anthropic’s real challenge may well be another: to demonstrate that investing in a company with high revenues (16.3 billion dollars in the first half of 2026, up 1,300 per cent on the previous year), which has significant losses (80.7 billion), modest operating profits (974 million in the second quarter of 2026) and has committed to spending 265.6 billion dollars on infrastructure and computing capacity, is a gamble capable of exciting not only artificial intelligence but also human investment portfolios. Anthropic has made headlines for the potential catastrophes that could be caused by artificial intelligence. It would also deserve to do so for all the benefits that artificial intelligence could bring us if used simply to build a better future.