The turning point has arrived. In 2025, AI's contribution to US GDP growth has surpassed that of human consumers. Meanwhile, US spending on AI infrastructure is rapidly approaching expenditures on human workers, with approximately 500,000 jobs disappearing within a year.

A striking chart circulating online shows US spending on AI infrastructure nearly matching expenditures for human labor. This development coincides with the loss of roughly 500,000 jobs over the past year. Most alarmingly, last quarter marked the first time AI capital expenditure exceeded consumer spending as a GDP driver.

We are witnessing the inflection point for white-collar employment - the fifth industrial revolution has begun. This year's US graduates face unprecedented challenges, with computer science graduates submitting thousands of applications only to receive offers from coffee shops.

**Data Center Spending Surpasses Office Buildings**

Official data shows US data center construction reached historic highs in June, with annualized investment exceeding $40 billion for the first time - a 28% increase from last year and 190% growth since ChatGPT's launch three years ago.

The trajectory shows data center construction spending (yellow line) rapidly approaching office building construction funding (green line). Industry observers note that data center cranes now outnumber office cranes, silicon-based infrastructure is winning the square footage competition, and planning committees prioritize cooling systems over parking ratios.

Construction industry professionals confirm this shift. While data center construction was theoretical just years ago, builders now report primarily constructing data centers rather than physical offices for human employees. Human workers are making way for AI.

**500,000 Tech Jobs Lost in 2025**

Recent analysis by researcher David Shapiro concludes that AI has eliminated 500,000 jobs in 2025. Using methodology similar to MIT economist David Autor's approach to tracking manufacturing job losses, Shapiro examined total employment versus economic output to determine productivity gains.

In Q2 2025, real GDP grew 3% while total weekly hours worked declined 0.3%. This gap indicates AI filled positions that would otherwise require additional human labor. While official AI-related layoffs number 20,000-30,000, the actual impact ranges from 300,000-500,000 jobs lost through position elimination, non-replacement of departing workers, and AI substitution.

**AI Drives GDP More Than Consumer Spending**

Second-quarter GDP data reveals a stunning phenomenon: AI capital expenditure's contribution to US GDP growth has far exceeded consumer spending this year. Historically, consumer spending drove 70% of US economic growth. However, two sectors comprising 6% of GDP (information processing equipment and software) now play larger roles in economic expansion than traditional consumption.

This capital expenditure surge proves AI has been consuming the American economy throughout the first half of 2025.

**Graduates Fall into Employment Black Hole**

Simultaneously, American graduates face an exceptionally bleak job market. July job additions hit pandemic-era lows, with graduate unemployment soaring to 5.3% - the worst in nearly a decade. Submitting 1,000 applications for two interviews has become commonplace.

Federal Reserve Bank of New York data confirms graduate unemployment at 5.3% compared to 4% overall unemployment. This represents the most challenging job market for graduates since 2015. July added only 73,000 jobs, far below economist projections, while long-term unemployment (over 27 weeks) increased by 179,000 to 1.8 million people.

Contributing factors include economic uncertainty from policy changes and AI's impact on customer service and basic software development roles.

**Computer Science Graduates Hit Hardest**

Adam Mitchell, a Georgia State University computer science graduate with three years internship experience and web development background at State Farm, exemplifies the crisis. Despite submitting over 100 applications in seven months, he received only two interviews. His sole job offer came from Starbucks for a 4 AM shift position. Even Costco hourly positions and State Farm customer service roles rejected him.

With savings depleted and substantial student loans, Mitchell represents technology workers - among the hardest hit demographics. According to tech recruitment site TrueUp, over 400 employers including Meta, Intel, and Cisco announced 130,000 layoffs through 2025.

**Degrees Becoming Worthless**

Anthony Young graduated from a technical college hoping to work at a nearby nuclear plant where his wife is employed. After over a year of searching, he remains unemployed. His two-year degree feels worthless, and local opportunities are scarce with meat processing plants and Michelin tire facilities cutting staff.

Young decided to return to school for electrical technician certification and applied for a five-year electrician apprenticeship program requiring a two-hour commute each way. Despite his wife's $90,000 salary, rising costs make survival difficult. Apartment rent increased from $600 to $1,000 monthly, grocery expenses rose from $80 to $180 weekly, while he carries $20,000 in student loans plus increasing insurance, gas, and utility costs.

**Graduate School: A Costly Mistake**

Sabrina Highfield, 25, earned over $70,000 as a project manager and analyst before returning to school in 2024 hoping to increase her salary. After earning a design master's degree from University of Texas at Austin this spring, she discovered her mistake. Following 1,000+ job applications yielding two interviews, she returned to live with her grandmother, taking a food assistance management position paying $35,000 annually - completely unrelated to her field.

She suspects her hiring resulted from the manager knowing her grandmother. Despite customizing resumes and cover letters for each application, nothing worked. Her current salary represents a significant step backward from her career goals.

**Few Success Stories**

Limited success stories include Jaylah Dorman finding clinical research work at a private practice and Saida Lopez-Rosales securing elementary teaching at her internship school. Dorman attributes success to networking and high-demand field specialization. Lopez-Rosales, earning $55,000, remains financially pessimistic, living with parents due to $1,000+ monthly rent costs.

"Everyone said you'd find work immediately after graduation," Lopez-Rosales recalls. "But the world has changed."

Anthropic CEO Dario Amodei's prediction that "50% of entry-level white-collar jobs will be eliminated by AI within five years" appears increasingly prophetic.