Giants Digging Deep into the Red to Invest in AI: What Should Small Businesses Do?

Alibaba’s core profit for the January-March 2025 period has plummeted by 84% compared to the same period last year. Normally, such a figure would raise eyebrows and prompt questions like, “Are they okay?” However, Alibaba is actually accelerating its investments in AI and cloud services.

During the same period, xAI (Elon Musk) is operating around 50 gas turbines at its data center in Mississippi, while a data center in Utah has been approved that is more than twice the size of Manhattan. The required power is 9GW, exceeding the total power consumption of the entire state of Utah.

Meta, Google, Microsoft, and Amazon — all are announcing annual investments in AI infrastructure amounting to trillions of yen. Meta anticipates its AI-related capital expenditures to reach up to $65 billion (approximately 9.7 trillion yen) by 2025.

Watching this “fight to the death in the red”, small business owners might be thinking one thing:

“So, what should we do?”

To cut to the chase: Don’t get caught up in it. Just pick the fruit.

—

Why Are Giants Willing to Sacrifice Profits to Invest in AI?

First, let’s understand the structure.

The reason Alibaba continues to invest in AI despite an 84% drop in profits is simple: AI infrastructure is a ‘location rental business’.

Companies that control cloud and AI infrastructure can charge usage fees from all businesses that provide services on top of it. In real estate terms, they become the “landowners”. This is why Alibaba, Microsoft, and Google are willing to dig into short-term losses to secure infrastructure.

xAI’s operation of 50 gas turbines to run its data center follows the same logic. Even with the risks of power costs and environmental lawsuits, they want to secure a foundation for training AI models. The fact that the massive data center in Utah is moving forward despite local opposition is because they are playing a game where “the one who secures the location first wins”.

This is not a game for small businesses. It is a ring exclusive to companies that can invest trillions of yen annually.

—

What Becomes Cheaper in the Giants’ Fight?

This is where it gets important for small businesses.

As the giants bleed in competition, the prices of AI services used by small businesses dramatically decrease. This is the essence of what is happening structurally.

Let’s look at some specific numbers:

• GPT-4 API usage fee: When it launched in March 2023, it was about $30 per 1M tokens → As of 2025, with GPT-4o, it’s $2.5. A reduction to about 1/12 in just two years.
• Image generation: In 2023, hiring a professional designer for a visual would cost between 5,000 and 30,000 yen, but with AI tools, it can be used for a monthly fee of 2,000 to 3,000 yen.
• Transcription: Outsourcing transcription for one hour of audio would cost 5,000 to 10,000 yen. With AI transcription, it can be done for a few dozen to a few hundred yen.
• Translation: Hiring a professional translator would cost 10 to 20 yen per character. With AI, a first draft can be produced for almost zero yen.

Thanks to Alibaba, Google, and Microsoft competing with the willingness to incur losses, the costs of tools available to small businesses are decreasing almost monthly.

Understanding this structure makes it clear what small businesses should do.

—

The One Thing Small Businesses Should Do: Integrate ‘Cheaper Options’ into Operations

There is no need to participate in the giants’ investment competition. What needs to be done is to constantly update a ‘list of cheaper options’ and apply them to the company’s operations.

Here are some specific examples:

1. Automating Inquiry Responses

In a local manufacturing company, they replaced inquiry responses received via phone and fax with an AI chatbot. The introduction cost was around 10,000 to 30,000 yen per month. Previously, developing a dedicated system would have cost over 3 million yen. They freed up the equivalent of one part-time employee’s response time, allowing that time to be redirected to customer visits.

2. Creating Sales Materials and Proposals

Sales representatives who previously took half a day to create a proposal can now generate a draft with AI and finish it themselves in just one hour. If they create 20 proposals a month, they save about 80 hours. This is nearly equivalent to the labor time of one sales representative.

3. Recruitment and Job Listings

The market rate for outsourcing job listings is around 30,000 to 50,000 yen per piece. By creating a draft with AI and adjusting it in-house, the cost can be reduced to almost zero yen. If they produce 10 listings a year, that alone results in a savings of 300,000 to 500,000 yen.

4. Automating Meeting Minutes and Daily Reports

If meeting recordings are sent to AI, a summary with meeting minutes can be produced in 5 minutes. If daily reports are structured through voice input and AI formatting, the burden on the field is significantly reduced. The complaints of “it’s a waste of time to write daily reports” can now be a thing of the past.

What all these examples have in common is the approach of “replacing only the parts of the tasks that can be sufficiently handled by AI”. There is no need to hand everything over to AI. Retain the parts that require human judgment and relationship building, and delegate only the ‘tasks’ to AI.

—

From ‘Building In-House’ to ‘Choosing and Using’

Another trap that small businesses often fall into is the misconception that they must develop a proprietary AI.

Developing a custom AI model from scratch requires a minimum investment of several million to tens of millions of yen. Just the personnel costs for data scientists can range from 8 to 15 million yen annually. It is extremely difficult for small businesses to recoup this investment.

However, we are now in an era where high-performance AI created by giants can be accessed via API for a monthly fee of several thousand to tens of thousands of yen.

Thanks to the competition among OpenAI, Google, Anthropic, and Alibaba’s Qwen, small businesses can take the position of “choosing”. There is no need to build it themselves. They only need to choose which one to use.

Alibaba’s decision to bolster its AI infrastructure even at the cost of an 84% profit drop is actually good news for small businesses. The more intense the competition, the higher the quality of services and the lower the prices.

—

Environmental Costs: Should Small Businesses Be Concerned?

The gas turbine issue at xAI and the water resource problems of the Utah data center are indeed serious topics. The 9GW power consumption, environmental lawsuits, and local opposition are risks that giant tech companies should bear, and discussions about regulations will likely accelerate in the future.

However, these are not concerns that small business owners need to worry about at this moment.

For small businesses, “environmental consideration” means not wasting computational resources. This means using AI only for necessary tasks and not for unnecessary ones. This is good for the environment and also good for costs. It may seem obvious, but the ability to make the judgment of ‘using it only where necessary’ is a strength of small businesses. Large corporations are often too big for this judgment to be made quickly.

—

Conclusion: Watch the Giants’ Fight from the ‘Audience’

Alibaba’s 84% profit drop. xAI’s 50 gas turbines. Utah’s 9GW data center. Meta’s $65 billion annual investment.

These are all stories of giants fighting in the ring.

What small businesses should do is not to step into the ring. They should sit in the audience, pick up the tools that have become cheaper as a result of the fight, and bring them back to their own operations.

Specifically:

1. Once a month, take inventory of ‘things that have become cheaper due to AI’. Track price fluctuations of new tools and services.
2. Try AI first in one area where the most time is spent on ‘tasks’. Meeting minutes, translation, inquiry responses, document creation — any of these will do.
3. Don’t fall for the temptation of ‘proprietary AI’. Use off-the-shelf products in areas where they are sufficient.
4. Redirect saved time to ‘things only humans can do’. Building relationships with customers, making on-the-spot judgments, and challenging new initiatives.

Using the infrastructure that giants have invested trillions of yen in for a monthly fee of several thousand yen is the most rational AI strategy for small businesses in 2025.

Let the giants handle the competition in the red.