It's a strange moment in the AI revolution. On Wednesday, Anthropic priced its blockbuster initial public offering at $83 per share, giving the company a valuation of approximately $1.5 trillion.
When trading began Thursday, shares of ANT opened at $111.50, roughly 34% above the offering price, briefly pushing the company's implied valuation to approximately $2 trillion. But much of that early enthusiasm faded during the trading session as shares surrendered a substantial portion of their gains.
Meanwhile, another kind of investment is attracting attention: Americans are stocking up on generators, bottled water, canned goods, battery-powered radios, and other supplies in anticipation of tech and infrastructure disruptions.
The contrast is striking. Some Americans are betting extraordinary sums on an increasingly sophisticated digital future. Others are investing in ways to get by when that future temporarily stops working.
And recent events have given them multiple reasons to consider that idea.
Over the past week, major outages have disrupted Microsoft services, banking transactions, airline operations, and even the National Security Agency's internal communications network. The disruptions have generated widespread anxiety, along with speculation about cyberattacks and other potential threats.
The Washington Post reported Thursday that the disruptions have prompted some Americans to embrace emergency preparedness, including people who previously gave little thought to such precautions. One IT specialist purchased a hand-crank radio and additional water. Her father bought a portable generator.
There is, however, no established evidence that the outages were coordinated or resulted from a widespread cyberattack. Officials have pointed to separate technical failures, software problems, and maintenance activities.
But here's where the story takes an especially interesting turn: AI may be contributing indirectly to some of the very disruptions that have people questioning the reliability of modern technology.
AI-powered tools have dramatically accelerated the discovery of security vulnerabilities in existing software. That's generally good news: Finding weaknesses before criminal networks can exploit them makes systems safer.
But identifying vulnerabilities faster also creates pressure to fix them faster, sometimes before organizations have fully tested the consequences of those fixes.
As Fortune reported Wednesday, the rush to install security patches has become a growing concern among cybersecurity experts. Director of National Intelligence Jay Clayton linked the National Security Agency's recent network disruption to software updates following the discovery of vulnerabilities with AI assistance. The causes of other outages remain less certain, and there's no basis for attributing all of them to AI.
It's an extraordinary irony. Technology that helps us discover and prevent failures may also, through the speed and complexity of implementing its findings, contribute to new ones.
None of this means the AI revolution is failing, or that Americans should prepare for an imminent collapse of digital infrastructure. AI is delivering real benefits, and cybersecurity improvements remain essential.
But the past week's events raise a question that deserves at least as much attention as the trillions being invested in new AI capabilities:
As we build increasingly intelligent systems, are we doing enough to ensure that the infrastructure supporting them remains reliable?
That may explain why, at this particular moment, investing in a promising AI company and buying a backup generator can both seem like sensible preparations for the future.
