Claude Just Wiped $73 Billion from Tech Stocks. It's Not Done Yet

Here's something that would've sounded ridiculous a few years ago: an AI company published a blog post, and investors lost $40 billion in a single afternoon.

That's not a typo. That's what happened to IBM on February 23, 2026.

And it wasn't just IBM. CrowdStrike dropped nearly 10% that same day. Datadog fell 11%. Zscaler, Atlassian, Accenture: all took serious hits. In less than a week, a wave of AI announcements from Anthropic, the company behind the Claude chatbot, wiped an estimated $73 billion from the market caps of some of tech's biggest names.

So what actually happened? And more importantly, what does it mean for your portfolio?

Let's walk through it, piece by piece.

If you've never heard of COBOL, you're not alone. It stands for Common Business-Oriented Language, and it was invented in 1959. Yes, 1959. The year Alaska became a state.

Here's the wild part: COBOL still runs roughly 95% of ATM transactions in the United States. It processes insurance claims, handles government payroll, and keeps airlines in the air. Hundreds of billions of lines of COBOL code are running right now, today, on IBM's mainframes.

IBM has spent decades profiting from this. Think of it like this: you've got a house with a very complicated, very old electrical system. Nobody knows how it works anymore — except IBM's team of specialists. So every time you need even a small repair, you pay IBM's premium rates. For years, that was the deal.

Then Anthropic dropped a bombshell.

On February 23, Anthropic announced that Claude Code, its AI coding tool, can now modernize COBOL. It can map dependencies across thousands of lines of code, document old workflows, identify risks, and essentially do in hours what used to take human consultants months.

IBM's stock fell 13.2% the same day. Its worst drop since the dot-com crash in October 2000. Shareholders lost more than $40 billion.

Here's what a lot of headlines missed: the COBOL news wasn't even the first hit that week.

Two days earlier, on February 21, Anthropic had launched another feature: Claude Code Security. This one scans software for vulnerabilities and patches them automatically. That announcement hit cybersecurity stocks hard. CrowdStrike fell nearly 10%. Datadog dropped 11%. Zscaler shed over 10%.

The logic? If an AI can continuously scan and fix code, why do companies need as many dedicated cybersecurity platforms?

Now, whether that fear is fully justified is a fair debate and we'll get to that. But on Wall Street, fear moves first and asks questions later.

By the end of the week, the full damage looked like this:

That's a lot of red. And for investors who were holding these names going into February, it stings.

Just when you thought this week couldn't get more interesting, Anthropic walked into a very public fight with the U.S. Department of Defense.

Here's the short version: Anthropic has a $200 million contract with the Pentagon and has been the first AI company to have its model deployed on the military's classified networks. Defense Secretary Pete Hegseth set a Friday deadline, demanding that Anthropic allow the military to use Claude for "all lawful purposes," with no restrictions.

Anthropic said NO.

Specifically, Anthropic CEO Dario Amodei said the company "cannot in good conscience" agree to terms that could allow Claude to be used for mass surveillance of Americans or in fully autonomous weapons, weapons that can kill without human oversight.

The Pentagon threatened to label Anthropic a "supply chain risk" — a designation normally saved for foreign adversaries. They also threatened to invoke the Cold War-era Defense Production Act to force compliance.

Amodei's response was clear: these threats "do not change our position."

Then something unexpected happened. More than 330 employees from Google and OpenAI, Anthropic's biggest competitors, published an open letter titled "We Will Not Be Divided." They called on their own companies to stand alongside Anthropic and refuse similar demands. OpenAI CEO Sam Altman backed the same red lines.

Retired Air Force General Jack Shanahan put it bluntly: "Painting a bullseye on Anthropic garners spicy headlines, but everyone loses in the end."

This isn't just a tech story anymore. It's a question of who controls AI in America: private companies or the federal government. And the answer is going to matter for a lot more than Anthropic's contract value.

Fair question. Here's the honest answer: the market is currently in a "sell first, ask questions later" mode on anything that looks like a legacy tech moat. And some of that fear is legitimate, but some of it is overdone.

What's likely overblown: Translating COBOL to Java is one step. Moving off IBM's entire mainframe infrastructure: the data layers, the transaction processors, the disaster recovery systems built over 50 years is a completely different challenge. Several analysts, including teams at Evercore and Jefferies, have called out the IBM reaction as excessive. UBS upgraded IBM after the selloff.

What's real and worth watching: AI is genuinely compressing timelines for work that used to take months. That doesn't mean IBM disappears overnight. But it does mean the consulting model, charging for hundreds of human hours, is under pressure. That's not going away.

The Pentagon standoff: If Anthropic loses its $200 million contract, that's a hit. But the bigger risk is reputational, and so far, the market seems to be watching this more as a policy story than a valuation story. If it escalates, that changes.

For investors with exposure to legacy IT consulting, cybersecurity platforms, or enterprise software, this is a moment worth paying attention to. Not to panic but to understand what's actually shifting under the surface.

One AI company is reshaping how Wall Street thinks about entire sectors. That's rare. It's uncomfortable. And it's happening fast.

IBM may not be dead. CrowdStrike may not be toast. But the questions Anthropic is forcing investors to ask: about moats, about pricing power, about what AI can and can't replace—are real questions. And they're not going away.

The investors who come out ahead in this environment won't be the ones who panic-sold IBM at $223. They'll be the ones who understood what changed, what didn't, and positioned accordingly.

That's what we're here to help you figure out.