The AI security tax
Real talk: AI is a double-edged sword for your portfolio. While banks are betting on tech to boost productivity, they’re also getting hit with a massive 'AI tax' just to stay secure. Bank of America CEO Brian Moynihan revealed that the firm had to sprint to patch deep-seated software vulnerabilities that were exposed by Anthropic’s powerful Mythos model, which hit the scene earlier this year.
To plug these holes, BofA brought in 1,000 external contractors to assist their 3,000-person in-house team. The cost for this emergency cleanup and related regulatory compliance in the third quarter alone? A cool 'couple hundred million plus.'
The profit vs. patch struggle
It’s giving a high-stakes game of cat and mouse. Moynihan emphasized that the bank has to stay ahead of these AI models, which are proving suspiciously good at finding decades-old code flaws. However, BofA is trying to play it cool with investors—especially since the company’s shares took a dive of up to 8% on Wednesday. The CEO claims these remediation costs are temporary and will decline once the legacy systems are patched.
Meanwhile, BofA is still pushing for AI-driven gains. They’ve got about 130 internal projects running and expect AI-related savings to climb from $300 million this year to $700 million by 2027. It's a classic case of spending money to make (and save) money, though the firm also cut its headcount by roughly 3,400 employees over the last year.
Why it matters
When big players like BofA spend hundreds of millions just to defend their digital walls from new AI tech, it proves that the 'AI revolution' isn't just about revenue growth—it’s a massive, expensive defensive upgrade that will keep companies' margins tight while they struggle to secure their infrastructure.
