HiddenLayer's $100M Shows AI Safety Is Now Big Business

HiddenLayer, a company that protects AI systems from attacks, just raised $100 million. Why? Because businesses are terrified of AI going wrong. Hackers can trick AI into revealing private data, and AI tools can make costly mistakes. As more companies use AI every day, they need bodyguards for it. That market is exploding — from about $1.5 billion last year to nearly $3 billion this year, according to analysts. Banks and giant tech companies are HiddenLayer's biggest customers, along with the U.S. government.

Three years ago, people doubted AI attacks were even real. Now it's a crisis. AI has become more powerful and more independent, acting on its own to do jobs. That means more ways for things to break. HiddenLayer originally protected regular computer models, but now it also watches AI "agents" — programs that can take actions by themselves. It stops attacks called prompt injection, where someone slyly tells an AI to ignore its rules, and it checks open-source AI models to make sure they aren't secretly carrying hidden malicious code.

One of HiddenLayer's customers is apparently a huge AI company with 700 million weekly users — think OpenAI or Anthropic. This shows even the biggest names need extra security. But there's a catch: security tech like this can get expensive fast, and big tech platforms like Microsoft or OpenAI might eventually bundle similar protection into their own products. For now, startups like HiddenLayer are cashing in on our growing fear of what AI can do wrong.

For regular people, this is good news. It means the AI tools touching your bank account, health care, or job applications are getting more guardrails. Less chance of AI leaking your personal data or making a huge, unplanned decision. But it also means security will keep playing catch-up, because AI keeps getting smarter — and so do the tricks people use to break it.