If you’re building production bots on Claude’s API, you need to know that Anthropic’s support system might ghost you when money’s on the line.
I’ve spent the last month staring at $180 in unexplained charges on my Anthropic account. No response. No explanation. Just silence from a company that’s supposed to be leading the AI revolution with “responsible” practices.
What Actually Happened
In early March, my billing dashboard lit up with charges I didn’t recognize. I’m a Claude Max subscriber, so I’m used to predictable monthly costs. Between March 3-5, something went sideways. The charges appeared, and my attempts to get answers disappeared into the void.
As of April 2026, I’m still waiting. Other users have reported similar billing issues during the same timeframe, which suggests this wasn’t an isolated glitch affecting just my account.
Why This Matters for Bot Builders
When you’re running production systems, billing predictability isn’t a nice-to-have. It’s fundamental infrastructure. I architect bot systems that handle thousands of API calls daily. My clients depend on accurate cost projections. When charges appear out of nowhere and support goes dark, it creates a cascade of problems.
First, there’s the immediate financial hit. $180 might not sound catastrophic, but multiply that across multiple projects or scale it up to enterprise usage, and you’re looking at serious budget chaos.
Second, there’s the trust issue. If I can’t get a straight answer about what I’m being charged for, how can I confidently recommend Claude to clients? How do I explain to a stakeholder that our AI provider might randomly bill us and then refuse to discuss it?
The Payment System Adds Insult to Injury
When I tried to update my payment methods, the system rejected three different credit cards—Mastercard, Visa, and American Express. I even attempted to use Stripe Link. Every single option failed with the same error.
This isn’t a “my card was declined” situation. This is a pattern that screams system-level problems on Anthropic’s end. Yet there’s no acknowledgment, no status page update, no communication whatsoever.
The Bigger Picture
Anthropic recently pushed out version 2.1.88 of Claude Code with an accidental file inclusion that caused issues. They’re also making bold predictions about powerful AI systems appearing by late 2026 or early 2027. They’re clearly focused on the future and the big technical challenges.
But here’s what gets lost in that forward momentum: the basics still matter. You can’t build trust in advanced AI systems if you can’t handle basic customer service. You can’t claim to be a responsible AI company if you leave paying customers hanging for over a month on billing disputes.
What Bot Builders Should Do
I’m not saying abandon Claude’s API. The technology is solid, and for many use cases, it’s still the best option available. But you need to plan for this reality:
- Set up billing alerts that notify you immediately of unexpected charges
- Keep detailed logs of your API usage so you can cross-reference against bills
- Consider maintaining accounts with multiple AI providers as a hedge
- Document everything when issues arise, because support might not
The AI space is moving fast, and companies are racing to ship new capabilities. But operational excellence—including responsive support and transparent billing—can’t be an afterthought. When you’re asking developers and businesses to build their infrastructure on your platform, you’re making a promise that goes beyond the API’s capabilities.
I’m still waiting for Anthropic to honor that promise. My $180 is still unexplained. And my confidence in recommending them for production systems has taken a serious hit.
đź•’ Published: