The Era of 'Cost-Effectiveness' for AI? Why Companies Are Cutting AI Budgets

AI Summary

As companies shift their strategy from 'using AI at all costs' to 'cost-effective AI utilization,' OpenAI and Anthropic are seeing changes in their growth patterns.

Imagine this: the AI assistant your company ambitiously adopted is charging you tens of thousands of dollars every month. At first, it was fascinating and helpful, but now management is asking, “Is it really worth that much?”

Until very recently, many companies were caught up in an “invest-at-all-costs” craze, adopting AI without regard for expenses. However, as of June 2026, the atmosphere of this heated AI market is changing. Even the two titans of the AI industry, OpenAI and Anthropic, are now facing a new homework assignment: not “how many tokens (the unit of information AI reads and writes) are being consumed,” but “how efficiently can we operate?” [Source: CNBC, Source: CNBC].

Why does this matter?

This change is not just a problem for IT companies. It impacts our lives broadly, from smartphone voice assistants to business translation tools and various automation services.

The fact that companies have started to prioritize “efficiency” instead of spending money freely means that AI is no longer a trendy toy, but a “proven tool” that must generate actual profits. Consequently, service providers are under pressure to release cheaper and smarter models, which will long-term create an environment where ordinary consumers like us can use more cost-effective AI services [Source: Let’s Data Science, Source: Progressino].

Easy explanation: What is an ‘AI token’?

In simple terms, a “token” is the basic unit of information processed by AI when we converse with it. For example, if you say the word “apple” to an AI, the word is broken down into a single piece; that piece is a token.

Previously, companies didn’t pay much attention to how many of these pieces they were using. They used AI as if it were an unlimited data plan. But now, every single token equals cost. If in the past, a company wouldn’t mind spending 10 cents to ask “How’s the weather today?”, now they are in a situation where they might ask “Answer as briefly as possible” to save even a penny [Source: Let’s Data Science].

To put it in perspective, it’s like companies that were enjoying unlimited course meals at a high-end restaurant every day have now started looking for “value-for-money” spots where they can maintain the quality of food while cutting costs in half.

Current situation: The giants’ dilemma

Of course, OpenAI and Anthropic remain the leaders driving the AI industry. As of 2026, OpenAI expects revenue of approximately $25 billion (about 35 trillion KRW), and Anthropic is also looking at revenue of around $19 billion, showing scary growth [Source: 24/7 Wall St.].

But the voices on the ground are cold. Flo Crivello, CEO of the startup Lindy, revealed that they completely switched from the high-performance Anthropic models they were using to other cheaper alternatives (such as DeepSeek), significantly reducing their cost burden [Source: Let’s Data Science]. Anthropic, too, is responding to this flow by changing its billing system from existing fixed-price plans to a method of paying for actual usage, pursuing realistic profitability [Source: CNBC].

What will happen in the future?

Moving forward, choosing a “value-for-money AI that fits my tasks perfectly” rather than just “unconditionally smart AI” will become a company’s competitive edge. Developers will compare various models to find ones that are similar in performance but lower in price, and will devote more effort to directly integrating lightweight (high data processing efficiency) models [Source: Let’s Data Science].

The AI revolution will not stop [Source: YouTube]. It is just changing its form from flashy fireworks to a steadily burning bonfire. As consumers, we will also be able to find efficient AI services that fit our individual needs among a wider variety of options.

MindTickleBytes’ AI Reporter Perspective

The fact that the AI industry has moved past the “invest-at-all-costs” stage to a stage that values practical efficiency means that AI has entered a maturity phase where it must prove the essence of business: “effect compared to cost.” From now on, rather than flashy promotional copy, how seamlessly it operates in our actual lives and businesses will determine a company’s success or failure.

References

Share this article:

Test Your Understanding

Q1. What is the biggest recent change in how companies utilize AI?

Unlimited increase in AI usage
Strategic AI choices considering cost-effectiveness
Complete halt to AI adoption

Companies are now shifting direction toward efficient utilization that maximizes effects within budget rather than unconditional AI usage.

Q2. What actions are companies taking to reduce AI costs?

Upgrading to more expensive models
Switching to open-source or cheaper alternative models
Developing AI models in-house

Some companies are switching to open-source models or cheaper alternative services like DeepSeek to actually reduce costs.

Q3. What is the recent shift in Anthropic's business strategy?

Expansion of fixed-price plans
Introduction of token-based billing proportional to actual usage
Complete abolition of free services

Anthropic has moved away from fixed pricing to a token-based billing system where costs are paid based on actual usage, making their revenue model more realistic.