A risk analysis of the AI boom: energy crisis, mass unemployment, tech stock overvaluation, and threats to the dollar system. Why 95% of AI implementations are unprofitable and where hyperautomation is leading us.
7 min read
Share:
AI summary
The article analyzes the risks of the artificial intelligence boom against the backdrop of massive investments ($365 billion in 2025) and rising unemployment. The author warns of a possible AI bubble burst, energy crisis, and systemic threats to the dollar financial system. Hyperautomation could lead to large-scale deflation and economic depression without radical social support measures.
The fundamental question: can AI live up to expectations, or are we witnessing another cycle of euphoria fueled by leverage and inflated valuations? This article examines current AI adoption trends, its impact on the labor market, risks of hyperautomation, energy shortages and capital dislocation, as well as the risks of a systemic crisis in the dollar system. Let's analyze how the short-term pursuit of profits could lead to long-term economic upheaval, and assess possible future scenarios.
Current Assessment of AI Adoption and Its Economic Impact
AI has become the primary driver of investment in the technology sector. In 2025, major companies such as Microsoft, Google, Amazon, Meta* and Tesla plan to spend $365 billion on AI infrastructure—70% more than in 2024. Case in point: OpenAI's deal with Oracle for $300 billion over five years to procure computing capacity, which became the largest contract in cloud technology history. This contract requires 4.5 gigawatts of electricity, equivalent to the energy consumption of four million family homes. Such scale underscores AI's ambitions, but also exposes its vulnerabilities: OpenAI, with annual revenue of $10 billion, must pay Oracle an average of $60 billion annually, raising questions about the sustainability of the entire business model. In early September, media reported that OpenAI also sharply its projected cash burn by 2029 to $115 billion—$80 billion more than previous expectations.
In the U.S. labor market, AI's impact is already being felt. According to Bureau of Labor Statistics data, in 2024 the U.S. economy added 911,000 fewer jobs than previously estimated, which points to a cooling market. Sectors related to manufacturing, retail and energy are experiencing a wave of layoffs, partly due to tariff policy and automation. A Stanford University study foundthat since the widespread adoption of generative AI, employment among young professionals (ages 22–25) in the most vulnerable occupations has declined by 13%. Anthropic CEO Dario Amodei predictsthat AI could lead to the loss of half of entry-level office positions and an overall unemployment rate of 10–20% within the next five years. A massive shift has already occurred: knowledge workers in "non-routine cognitive occupations," who in most crises have been the driving force behind labor market recovery, now for the first time in history account for a larger share of the unemployed in the U.S. than workers in non-routine manual jobs (such as healthcare and food preparation).
The economic impact of AI is mixed. On one hand, technology companies like Nvidia are reaping the rewards of the hype: its market capitalization has reached $4.2 trillion, and the "Magnificent Seven" (Nvidia, Microsoft, Google, Amazon, Meta*, Tesla, Apple) make up a third of the S&P 500 index, pointing to an extreme concentration of capital. On the other hand, an MIT study found that 95% of 300 AI implementations at public companies failed to deliver a return on the $30–40 billion investment. This points to a disconnect between expectations and reality, fueled by speculative hype.
Key AI Risks
Energy Deficit. Large-scale AI investments require enormous energy resources. OpenAI's deal with Oracle illustrates the problem: 4.5 gigawatts—an amount of energy exceeding the output of two Hoover Dams. To realize Nvidia's ambitions for expanding computing capacity over the next 5–10 years could require up to 25% of the M2 money supply, equivalent to trillions of dollars. According to the International Energy Agency (IEA), electricity demand from AI-specialized data centers is expected to increase more than fourfoldby 2030. This creates the risk of an energy crisis, including rising prices for households, especially amid the global transition to green energy and declining investment in traditional energy sectors, as infrastructure fails to keep pace with demand from data centers.
Investment Dislocation. AI investments are diverting capital from other sectors. According to various estimates, over the past 18 months the "Magnificent Seven" have spent $560 billion on AI, generating only $35 billion in revenue. The concentration of capital in AI is exacerbating inequality: while tech giants thrive, traditional industries—construction and manufacturing, for example—are losing jobs. Moreover, passive investing, which channels capital into indexes dominated by tech companies, compounds this problem, creating market fragility. Similar excessive investments in internet companies led to the collapse of the dotcom bubble.
Hyperautomation Threatens the Labor Market. Salesforce CEO Marc Benioff stated that AI performs up to 50% of the work at the company, enabling it to cut 4,000 customer support employees. Microsoft laid off 15,000 employees to transition to a "data mining system." Amazon and PepsiCo are implementing AI tools like Agentforce to automate tasks previously performed by humans. However, AI is incapable of solving non-standard problems, leading to deteriorating customer experience. Thus, drivers of growing social discontent include not only job loss (income loss), but also declining quality of algorithm-provided services.
Technological limitations of AI. AI based on large language models (LLM) lacks true intellectual capability. Problems such as errors, misconceptions, and "hallucinations" (fabricated responses), as well as the inability to perform precise mathematical calculations, limit the applicability of AI in mission-critical areas. In July 2025, Replit AI (a code-writing solution) went out of control, deleting a key database: Replit's CEO explained that the AI "panicked" and began executing commands without permission. That same month, cybersecurity specialists found a vulnerability in McDonald's chatbot, through which they gained access to personal data of 64 million job applicants at the company. Other incidents underscore the need for additional control algorithms, which reduces AI's overall effectiveness. For example, Amazon faced criticism after its AI recommended antisemitic books, while AI on the X.com platform began calling itself "MechaHitler".
AI and the systemic crisis of the dollar system. The global economy is under pressure from a combination of factors: record leverage in financial markets, stagflation, and geopolitical instability. Margin debt in brokerage accounts is at all-time highs, while hidden forms of leverage, such as zero-day options, are amplifying volatility. Against this backdrop, the Shiller price-to-earnings ratio (CAPE) has reached levels not seen in two decades, signaling overvaluation in the U.S. stock market. It is calculated as the ratio of current stock price to average earnings per share over the past 10 years, adjusted for inflation, which smooths out cyclical fluctuations in corporate earnings, providing a more accurate picture of the market's long-term valuation.
Meanwhile, the global financial system depends on U.S. interest rates. The Fed, which sets interest rates, fears not inflation but systemic collapse, which could be triggered by even a minor change in the regulator's policy. For example, the yen carry trade, which has provided liquidity to financial markets for decades, is under threat due to narrowing yield spreads. Refinancing of $9 trillion in U.S. Treasury bonds over the coming year could become a problem if rate cuts trigger panic among investors. Derivatives create additional risk: even a small change in yields can provoke crises, as happened in the United Kingdom in 2022. AI euphoria only compounds this fragility, diverting capital from the real economy and exacerbating inequality.
Possible Economic Scenarios
In the coming months, AI implementation will continue displacing jobs, particularly in sectors with high automation. Youth unemployment in the US already exceeds 10%, while manufacturing job openings have declined by 78,000 over the year. If the Fed cuts rates, this may temporarily support the market but will intensify speculation and inflation, which is forecast to reach 4% due to rising housing and energy prices. The AI bubble could burst if companies like OpenAI or Nvidia fail to meet their financial obligations, undermining investor confidence.
In the medium term, AI may reach a development plateau, following an S-curve where the pace of improvements sharply decelerates. For instance, disappointment with the new GPT-5 and problems with large reasoning models (LRM) point to the limitations of current technologies. This will lead to a reassessment of AI investments, hitting the "Magnificent Seven" stocks and heightening recession risks. Passive investing, comprising trillions of dollars, will amplify the downturn as capital outflows from indexes trigger cascading sell-offs.
Looking further ahead, AI could become a transformative force, but only if current limitations can be overcome. However, a systemic crisis in the dollar system, including debt refinancing and derivative risks, may lead to a global economic depression before any "golden age of AI" begins. By that point, automation will only intensify social inequality and economic decline, since AI or autonomous robots don't shop at stores, dine at restaurants, or take out mortgages. Without implementing extreme measures such as universal basic income, consumer demand will collapse, triggering massive deflation.
Conclusion
AI enthusiasts promise a technological breakthrough, but its rapid deployment amid financial fragility and hype takes us back to the era of historic bubbles. Hyperautomation is destroying the labor market, capital dislocation is inflating inequality, and energy and financial risks threaten economic stability. The dollar system is teetering on the brink of systemic crisis, where one wrong move by the Fed could trigger catastrophe. To avoid depression—if that's still possible—we need to separate genuine innovation from speculative noise and rethink the economic model so that technology serves society, not the other way around.
*Meta is designated as an extremist organization in the Russian Federation
