AI Tool Spending Grows 10x as Companies Track Real Returns

Companies pouring money into AI coding tools are about to get something they've desperately needed: proof that it's working.

Anthropic just launched a suite of reporting tools that show businesses exactly what they're getting from their AI spending. Managers can now see their team's output over any date range, track costs down to the penny, and identify exactly where workflows stumble.

The timing couldn't be better. AI coding tool spending has exploded 10 times since January, now eating up about 2% of total engineering costs at many companies. That's growth most CFOs never planned for, according to data from Jellyfish, which tracks software development at 276,000 companies.

"It's not enough to just spend a bunch anymore," says Nicholas Arcolano, chief scientist at Jellyfish. "You now have to optimize it and make it work."

The push for transparency is paying off for Anthropic. The company hit profitability in the second quarter, with existing customers spending nine times more year over year. Their AI coding tool, Claude Code, accounts for nearly a third of all AI-proposed code changes across major companies.

But here's the surprising part: most Fortune 500 companies still aren't seeing major gains from AI despite the spending surge. These new reports aim to change that by offering specific recommendations for improvement, all while promising not to use company data for AI training.

The Bright Side

What makes this moment special isn't just better tracking. It's watching a technology mature from experimental toy to essential business tool in real time.

Companies are discovering that switching AI providers isn't like changing email apps. Once teams standardize their workflows around one system, the switching costs add up fast. Every AI tool has its quirks, its own way of handling files and processes.

Anthropic's enterprise customers are sticking around not because competitors don't exist, but because the tools already work well enough that change feels risky. The company embeds consultants inside customer organizations to help teams succeed, and these new self-service reports extend that hands-on approach to everyone.

The investment in making AI actually useful, rather than just impressive, signals something bigger: we're moving past the hype phase into the making-it-work phase. That's when technology stops being a gamble and starts being a genuine productivity boost.

For the thousands of companies trying to figure out if their AI spending makes sense, having clear answers beats blind faith every time.