Large Corporations Reduce Workforce While SMEs Face Labor Shortages: What is Happening?

Oracle has announced a reduction of 21,000 jobs, which accounts for about 13% of its total workforce.

In factories, robots are taking over production lines, and AI agents are conducting job interviews.

On the other hand, small and medium-sized enterprises (SMEs) in rural areas are struggling with labor shortages. “Even when we post job openings, no one applies,” and “the new hire we finally managed to recruit quit after three months” — such stories have become commonplace.

The phrases “reducing workforce” and “labor shortages” may seem contradictory at first glance. However, this is not a contradiction; it is two sides of the same structural change.

AI is rapidly sorting out “jobs that no longer require human involvement” from “jobs that only humans can perform.” Large corporations are cutting the former, while SMEs are struggling to fill the latter. That’s the crux of the matter.

The question is whether we can think about “what we should do next” after understanding this structure. This is where the fork in the road lies.

—

Oracle’s Reduction of 21,000 Jobs — What Has Become ‘Unnecessary’?

Oracle’s workforce reduction, implemented in early 2025, affects approximately 21,000 employees, or about 13% of its total workforce of around 141,000. According to reports, the cuts primarily target areas that can be replaced by AI and automation, such as cloud infrastructure management and customer support.

It is important to note that Oracle did not simply save costs by reducing its workforce. The company has significantly increased its capital investment in AI infrastructure. In other words, the reality is that “labor costs have been redirected towards AI investments.”

Let’s consider annual labor costs. Assuming the average salary of an Oracle employee is $100,000 (approximately 15 million yen), the reduction of 21,000 employees translates to an annual cost of about $2.1 billion (approximately 315 billion yen). This scale of funds is flowing into AI infrastructure and GPU clusters.

For large corporations, this is a rational decision. Humans need rest and make mistakes, while AI operates 24/7 and scales efficiently. If the same quality of work can be maintained, it is a natural conclusion for management to replace human labor with AI.

However, the question to consider is not “Should we do what Oracle is doing?” but rather “Is the talent that Oracle has cut the same as the talent we need?”

—

Factory Automation — The Nature of ‘Jobs That Should Be Done by Humans’ is Changing

The automation of manufacturing is also accelerating. According to the International Federation of Robotics (IFR) 2024 report, the number of operational industrial robots worldwide has surpassed approximately 4.2 million units. Notably, China accounts for more than half of the global installations, introducing about 290,000 new units annually.

On the other hand, an interesting phenomenon is occurring. In some factories in India, workers are being equipped with wearable cameras, and AI is analyzing work patterns and procedures. The goal is not “surveillance” but rather “standardization” and “optimization” of work processes.

Here lies a hint.

Robots excel at “repeating predetermined movements accurately.” However, what is truly valuable on the ground is the ability to “notice abnormalities,” “change setups,” and “make judgments based on the situation.” As automation progresses, the work required from humans is shifting from ‘tasks’ to ‘judgment.’

Voices from local manufacturing sites reflect this: “We no longer need people for simple tasks. However, we are overwhelmingly short of individuals who can handle machine troubles and oversee the entire process.”

In other words, the reduction in workforce and the labor shortage are due to a change in the quality of required talent. It is not a matter of quantity but of quality.

—

AI Agent Interviews — Breaking the Conventional Wisdom of Recruitment Costs

AI agent interviews, being developed by companies like FikaJobs, aim to transform the very structure of recruitment.

Recall the traditional recruitment process for SMEs. Hundreds of thousands of yen spent on job advertisements, the time spent by HR personnel on document screening, scheduling interviews, and conducting interviews — hiring one person can cost between 500,000 to 1,000,000 yen and take several weeks.

What changes when AI agent interviews become practical?

• Initial screening becomes automated. Job seekers record a three-minute video, and AI analyzes their suitability, communication skills, and motivations. HR personnel no longer need to read through 100 resumes.
• Interviews can be conducted 24/7. Job seekers can attend interviews at midnight or on weekends. This significantly lowers the barriers for applicants to rural SMEs.
• Recruitment costs can dramatically decrease. The cost of initial interviews could potentially drop to a few hundred yen per person.

This is great news for SMEs. While large corporations can develop their own recruitment AI, SMEs can utilize the same systems for a monthly fee of several tens of thousands of yen. The ‘economies of scale’ in recruitment are collapsing.

However, there are pitfalls. AI cannot make intuitive judgments like “Would I want to work with this person?” Final interviews should still be conducted by humans, and onboarding after hiring needs to be more thorough than ever.

Reinvest the time saved through AI into ‘relationship building’ that only humans can do. This is the correct approach.

—

The Reality of ‘Job Reductions’ and ‘Labor Shortages’

When we line up the three movements we have seen so far — Oracle’s mass layoffs, factory automation, and AI interviews — a structure begins to emerge.

AI is lowering the costs of ‘routine tasks.’

Data entry, report creation, initial inquiry responses, routine quality inspections, and document screening. The costs of these jobs have been reduced to one-tenth, or in some cases one-hundredth, by AI.

Large corporations are reducing the workforce that handled these “cost-reduced tasks.” This is rational.

On the other hand, what is lacking in SMEs is not “people to perform routine tasks.” What is needed are individuals who can make judgments on-site, build trust with customers, and handle multiple tasks across different areas — in other words, people doing jobs that AI cannot (yet) replace.

These two points are separate issues. Confusing them can lead to incorrect judgments.

There is no need to fear that “AI will take away human jobs.” However, it is no longer acceptable to have humans perform jobs that AI can handle.

—

What SMEs Should Do Now — Three Concrete Strategies

So, what should rural SMEs do?

1. First, Inventory ‘Jobs That Can Be Assigned to AI’

Divide internal tasks into “jobs that require judgment” and “jobs with predetermined procedures.” Start replacing the latter with AI or automation tools. Tasks like invoice processing, meeting minutes creation, and initial inquiry responses can be automated starting today with tools costing a few thousand to tens of thousands of yen per month.

2. Integrate AI into the Recruitment Process

For SMEs without dedicated HR personnel, the benefits of AI screening are significant. Automating applicant responses alone can reduce the time that CEOs and team leaders spend on recruitment tasks by more than half. This freed-up time can be focused on the final interviews to determine “Would I want to work with this person?”

3. Concentrate Human Resources on ‘Jobs That Only Humans Can Do’

Reinvest the time and costs saved through AI into employee training, customer engagement, and on-site improvements. This will become the true competitive advantage for SMEs. As large corporations cut jobs and replace them with AI, the ability to concentrate resources in areas where ‘human power makes a difference’ is actually a strength of SMEs.

—

This is Not a Crisis, but a Transition

The phrases “reducing workforce” and “labor shortages” are not contradictory.

The definition of necessary talent is simply changing.

Oracle cut 21,000 jobs because those tasks can now be performed by AI. The labor shortage in rural SMEs exists because there are still many jobs that AI cannot do.

There is no need to mimic large corporations and think, “Let’s also reduce our workforce with AI.” After all, SMEs are already short on labor.

What needs to be done is to eliminate ‘jobs that do not require human involvement’ with AI and concentrate everyone on ‘jobs that only humans can do.’

The SMEs that quickly recognize this structural transformation will win in the next decade. Conversely, those companies that think “It’s fine as it is now” will be left behind by both large corporations and AI.

Start with one task. Find one job that can be automated starting tomorrow and give it a try.