Разработка ПО
Менеджмент
Для компаний
Для стартапов
Архитектура ПО
Employees spend millions on Claude without limits, while management discovers the losses. A real story and takeaways for entrepreneurs implementing AI.
Оглавление
What happened
An AI consulting company shared a shocking story with Axios: one of its clients spent $500 million on Claude licenses from Anthropic in just one month. The reason is banal to the point of genius — employees received unlimited access to tools without any quotas or controls. The result was predictable: people used expensive AI for tasks that can be solved in seconds without it — for example, checking the weather. The story instantly spread across social media, where users compared the amount spent to buying five private jets or two superyachts.
This is not an isolated case. Microsoft abandoned Claude Code in favor of GitHub Copilot CLI. Uber exhausted its Claude Code budget back in April, although the year had only just begun. Uber COO Andrew Macdonald stated: there is no connection between increased AI usage and real benefit for customers. Amazon shut down its internal AI token usage ranking, and Vice President Dave Treadwell asked employees to stop using AI for the sake of using it.
How this is useful for business
The $500 million story is not a joke, but a warning. The main conclusion: without a cost control system, AI turns from an efficiency tool into a budget black hole. Companies that handed out access to employees thoughtlessly are now receiving bills that are disproportionate to the results. It is especially telling that even giants like Uber and Microsoft are suffering — which means the problem is systemic.
The positive side: the AI services market is beginning to mature. Clients are stopping the chase for volume and starting to calculate ROI. This creates opportunities for those who know how to work with AI effectively, not extensively.
How to make money from this
First, demand for AI audits is growing. Companies do not understand where the money spent on AI is going and need experts who can optimize costs. Second, there is demand for employee training — not “how to use AI,” but “how to use AI wisely and for the right tasks.” Third, the need for AI spending control and analytics tools is growing. This is a completely new niche with low competition.
Business ideas
1. AI consulting for cost optimization: auditing current processes, identifying overspending zones, implementing quotas and monitoring systems. Revenue — from $5,000 per project.
2. AI budget control platform: a SaaS tool that tracks AI usage in a company in real time and shows where spending is inefficient. Subscription $200-2,000 per month.
3. Corporate training “AI with benefit”: courses for employees that teach them to determine when AI is needed and when they can do without it. One-time trainings from $3,000.
4. AI process implementation agency: you help companies build workflows with AI where every request is justified and measurable. Contracts from $20,000 per month.
5. Marketplace of ready-made AI prompts for business: structured templates for typical tasks — marketing, sales, analytics. Subscription $50-500 per month.
Risks and limitations
The main risk is that the AI market is changing quickly. Tools that are expensive today may become cheaper tomorrow or be replaced. Competition in consulting is already growing, and in a year or two the niche may become oversaturated. There is also a risk that companies will overestimate their capabilities and start selling services without real expertise, which will undermine trust in the market as a whole.
7-day action plan
Day 1-2: Study cases of AI overspending in open sources, collect data on typical mistakes. Day 3: Define your niche — consulting, software, or training. Day 4-5: Prepare your first commercial proposal or product prototype. Day 6: Test demand — post the offer on LinkedIn or specialized platforms. Day 7: Collect feedback and adjust your positioning. The main thing is to start before the market fills up.
Original news: Fast Company · See other news in the news section.
