SOCIAL, POLITICAL AND ECONOMIC SITUATION

Polarization

The United States is a country politically divided in half, where supporters of the Democratic and the Republican parties view the other as a threat to themselves and to existence of the nation. This climate of tension culminated in well-known events such as the attack on Capitol Hill on January 6, 2021, and continued through to the assassination of Minnesota Democratic representative Melissa Hortman and her husband Mark, and a few months later, the murder of Charlie Kirk.

Deindustrialization

This social fracture and the violence that has stemmed from it finds its roots in the outsourcing process of American companies, which progressively led to the physical and industrial abandonment of what was once the Manufacturing Belt, which by the end of the 20th century had become the Rust Belt. Year after year, the United States has seen itself weaken socially and strategically, with a population progressively losing jobs in central geographic areas, far from the coasts — all while providing work and know-how to a geopolitical rival like China.

Overlapping this deindustrialization were other crises that have profoundly marked the social fabric of the United States:

• Economic stagnation, stagnant median income over the last 20 years
• US progressive involvement in international conflicts since 2001
• The fentanyl crisis, with over 80,000 drug overdose deaths in 2024

TRUMP TARIFF POLICY

It is within this context of social fracture, deindustrialization, geopolitical competition, and economic uncertainty that the Trump administration's tariffs were introduced starting in April 2025:

• A baseline tariff of 10% on nearly all imports
• China: 20% tariffs apply specifically to goods linked to the fentanyl crisis, which will be reduced starting November 10. Export controls on advanced chips, supercomputers, and semiconductor manufacturing equipment to China since October 2022, and increased controls in 2025
• Tariffs regarding steel and aluminum increased to 50% globally, except for the United Kingdom where they stand at 25%
• Automobiles and parts: A 25% tariff applies to most foreign-made vehicles and parts
• India: 50% tariff (10% baseline plus 40%) as a penalty for continued importation of Russian oil. However, negotiations are underway that could reduce the tariff to 15%
• Brazil: 50% tariffs (10% baseline plus 40%) the United States considered the actions of the Brazilian government a threat to U.S. national security and foreign policy which lead to a diplomatic and political dispute. Aircraft and spare parts, orange juice, and certain petroleum products, which are considered key Brazilian export goods, are exempted from additional tariffs

Among the effects of the tariffs we can observe: extremely high volatility in financial markets, major investments—or announcements of supposed such investments—that many companies are willing to make to avoid being cut off from the world's largest market, rising construction costs (+4.7% year-over-year as of Q4 2024), and investments into U.S. territory to avoid trade barriers.

• Reducing dependence on supply chains vulnerable to geopolitical shocks, as demonstrated during the COVID-19 pandemic with the semiconductor shortage
• Slowing China's economic progress and investments. According to various think tanks and analysts, China has already surpassed the USA in some key economic indicators. China certainly has greater manufacturing output (28% of the global total versus 16% for the USA) and a higher GDP measured in purchasing power parity (PPP)

Legitimacy of the Tariffs

There is a question of legitimacy regarding the President of the United States' authority to impose tariffs, on which the Court is called to rule.The issue concerns the fact that a large portion of the tariffs were imposed by President Trump by invoking the International Emergency Economic Powers Act (IEEPA) of 1977—a law designed to manage security threats during times of emergency. What is being contested is whether the trade deficit and the fentanyl crisis constitute a genuine emergency that justifies the president using such broad powers.It is important to note that the tariffs on steel, aluminum, and automobiles, imposed under other trade laws (Section 232 of the Trade Expansion Act), are not the subject of this Supreme Court case. Oral arguments were heard on November 5, 2025, with several justices—both conservative and liberal—expressing skepticism about the tariffs' legitimacy. A decision is expected by June 2026, but given the expedited schedule, a ruling could come sooner. According to some commentators, there is a strong probability that a verdict could arrive by January 2026.

The S&P 500 recovered from the sharp drop triggered by the April 2025 tariffs. Markets remained more volatile than before and highlighted uncertainty over long-term economic and policy implications.

EPOCHAL CHANGE IN THE INVESTMENT LANDSCAPE

Investments in artificial intelligence and digital infrastructure are crucial in this economic, political, and commercial context. These investments represent an epochal shift in the United States economic growth model , as for the first time in 2025 private investment has surpassed consumer spending as the main contributor to growth. The closest historical parallel was during the WWII period when government military spending reached 36-40% of GDP, but this was in war production—there was no comparable private investment. Consumer spending was rationed, not constrained by investment.

The announced figures represent the largest mobilization of private capital since the construction of the railroads in the 1880s, but they raise questions about sustainability and the actual capacity to implement such ambitious projects. Big Tech company spending adds approximately 100 basis points to U.S. growth in 2025, based on Morgan Stanley analysis. The $364 billion investment by Big Tech in 2025 will have a cascading effect:

• Direct impact in the creation of approximately 620,000 jobs directly in Big Tech areas of expertise. Total impact, including the supply chain effect, in the creation of 2.5 million jobs
• $72.8 billion in the construction of data centers and commercial facilities
• $291.2 billion for servers and related equipment
• $21 billion in downstream investments to strengthen sectors such as data processing, electronics manufacturing, and communications equipment

The massive investments by technology companies tell us that they have completely altered the corporate investment landscape. Ten years ago, the five largest companies were in telecommunications and the energy industry (AT&T, Chevron, ExxonMobil, General Electric, Verizon). Today, the six largest companies all belong to Tech (Nvidia, Microsoft, Apple, Alphabet, Amazon, Meta) and are responsible for:

• ~31.7% of S&P 500 market capitalization
• ~30-35% of S&P 500 capex by 2025

APPLE

• Investment of more than $500 billion in the USA over four years (announced February 2025)
• Plans to hire around 20,000 people over the next four years, primarily in R&D, silicon engineering, software development, and AI/machine learning
• Apple's investment differs from other Big Tech companies through its focus on physical manufacturing alongside digital infrastructure
• New Silicon Initiative program preparing students for careers in hardware engineering and chip design, expanded to eight universities including Georgia Tech and UCLA
• Initiatives focused on AI, silicon engineering, and skills development for students and workers
• Expanding data center capacity in North Carolina, Iowa, Oregon, Arizona, and Nevada to support Apple Intelligence
• New Apple Manufacturing Academy in Detroit, Michigan (opened August 19, 2025) in partnership with Michigan State University, offering free in-person and online training to small and medium-sized businesses nationwide on smart manufacturing techniques, AI implementation, project management, and manufacturing process optimization

Apple's $500 billion commitment is a continuation of an ongoing investment routine for the company: the company announced $430 billion in 2021 and $350 billion in 2018, so this represents a natural progression rather than a dramatic new initiative prompted solely by tariff concerns.

CRITICAL ISSUES

This structural dependence on a single sector creates unprecedented systemic vulnerability.

This structural dependence on a single sector creates unprecedented systemic vulnerability. Without these tech-related categories, GDP growth would have been just 0.1% annualized—and for the first time ever, AI data center buildout has contributed more to GDP growth than consumer spending.

TECHNOLOGICAL REQUIREMENTS

The massive investments conceal extraordinary resource requirements for operating AI-related infrastructure:

Computing Power

Meta alone plans to bring 1 gigawatt of compute online in 2025 with 1.3 million GPUs, and is building additional facilities including a 5-gigawatt cluster (Hyperion Solution) expected by 2028. OpenAI's Stargate project requires 10 gigawatts—equivalent to the electricity consumption of New York City during peak summer demand. A single gigawatt-scale data center consumes as much electricity as 800,000-900,000 homes.

Energy Infrastructure

U.S. data center power demand is projected to more than double by 2035, from 35 GW to 78 GW. Data centers are expected to account for almost half of U.S. electricity demand growth through 2030. The IEA projects global data center electricity demand will exceed 1,000 TWh by 2030—equivalent to Japan's entire electricity consumption. Wholesale electricity prices have increased up to 267% near data center clusters, with residential bills expected to rise 8-25% in high-demand regions.

Advanced Semiconductor Manufacturing

Cutting-edge chip production depends on a fragile supply chain centered on TSMC (Taiwan), ASML (Netherlands), creating concentrated geopolitical risk.

Data Center Infrastructure

AI workloads require sophisticated cooling systems that consume 7-30% of facility energy. High-density GPU racks generate so much heat that traditional air cooling is inadequate, requiring advanced liquid cooling solutions including direct-to-chip cooling and immersion systems.

Skills Shortage

• Global AI talent demand exceeds supply by 3.2:1
• 68% of companies face moderate to extreme AI talent shortages
• 58% of data center operators struggle to fill open positions
• 325,000 new full-time data center positions needed globally by end of 2025
• A projected 700,000 AI job shortage by 2027
• AI roles command 67% higher salaries than traditional software positions

REGIONAL ECONOMIC DEVELOPMENT

Investments are concentrating and expanding across many states. Historically, California has been the center of the tech world and will continue to play a dominant role, attracting approximately 16% of the national tech workforce (down from 19% in 2020). However, in this cycle Texas has emerged in recent years as a major hub, reaching approximately 6-7% of the tech market, with costs 20-35% lower compared to Silicon Valley and higher ROI. In Texas, we can find many hubs:

• Stargate's 875-acre data center complex in Abilene
• Apple's chip design center (Austin)
• Apple's AI server manufacturing facility in Houston (250,000 square feet)

Other locations that are emerging or consolidating include:

• Seattle/Washington State: 56% senior talent concentration, Amazon/AWS cloud dominance and Microsoft AI integration
• New York Metro area: Highest entry-level concentration (26%), fintech innovation, ideal for fintech ventures and enterprise SaaS
• Boston/Massachusetts: Data science and biotech innovation center. Data scientist concentration 5.38% vs. 1.29% national average
• Virginia/Washington DC Metro: Government tech and cybersecurity hub, second-largest cybersecurity workforce nationally
• Colorado Springs, Colorado: Cybersecurity specialization leader with 7.5% of tech workforce concentrated in this field

Manufacturing Decline Zones

• Ohio: Rust Belt Epicenter. Traditional assembly workers lack robotics/systems integration expertise
• Michigan: Automotive sector transformation, EV transition creating winners and losers
• Illinois: Agricultural equipment decline, farm equipment manufacturing under pressure. Geographic divide: Chicago (thriving fintech, consulting, tech scenes) vs. Downstate (manufacturing-dependent communities suffering)
• Iowa: Agricultural manufacturing crisis, severe contraction, and limited alternatives

Mixed/Transition Zones

• Indiana: Slight decline but more resilient than Ohio/Michigan
• Wisconsin: Automation adoption without mass layoffs so far, industrial diversity provides buffer
• Arizona: TSMC semiconductor investment represents major job creator ($165B total investment, 6,000+ direct jobs planned), though operational profitability only achieved in 2025 after initial losses

SKILLS AND COMPENSATION TRENDS

The AI investment boom is creating a significant shift in skills and specialized labor demand. There is exponential growth of roles tech-related: AI and Machine Learning specialists, data scientists, cybersecurity experts, cloud architects, and automation engineers. Tech investments are having a dual effect: increasing demand for high-skill positions (grow of automation and systems integration roles) and a decline in traditional manufacturing jobs. U.S. manufacturing lost 87,000 jobs in 2024, continuing a long-term structural decline of 26% since 2000. This is fundamentally a restructuring rather than cyclical job market fluctuations. The skills gap remains the most significant barrier to business transformation, cited by 63% of employers, with nearly 40% of skills required on the job set to change by 2030.

The difference between job losses in manufacturing and job gains in new automation is not just about total job numbers but a profound mismatch of skills required in emerging sectors. This creates simultaneous labour shortage and unemployment in different sectors. According to the WEF, if the global workforce were represented by 100 people, 59 would need reskilling or upskilling by 2030, but 11 are unlikely to receive it—placing over 120 million workers at risk of redundancy.

Job demand typically concentrates in specific geographic areas. Eight metro areas—San Francisco, San Jose, Austin, Boston, Seattle, Los Angeles, New York, and Washington D.C.—accounted for nearly half of all tech sector job creation between 2015 and 2019. Meanwhile, manufacturing job losses have been most pronounced in the Northeast and Midwest creating geographic divergence and regional economic disparities.

Moving displaced workers from traditional roles to systems integration positions isn't quick. Career transitions typically require 4-9 months of formal training, but the complete transition often takes 18-36 months. This creates a transitional unemployment challenge even when enough new jobs are being created overall.

CAN THE US WORKFORCE MEET THE DEMAND?

There are several workforce that the United State will challenge in order to meet CHIPS and Science Act objectives. Engineering shortage could reach as high as 300,000 by the end of the decade according to McKinsey & Company. The Semiconductor Industry Association projects the industry's workforce will grow by nearly 115,000 jobs by 2030, but approximately 67,000 of these technical positions risk going unfilled. The White House estimates the country needs 90,000 to 100,000 more semiconductor technicians. The broader U.S. tech-economy faces a shortage of 1.4 million technicians, computer scientists, and engineers by 2030.

As tech industry investments surge, the shortage of qualified workers will become a significant challenge. According to the World Economic Forum’s Future of Jobs Report 2025, technology-related jobs will experience the strongest growth over the next decade:

• Big Data Specialist ~113%
• Fintech Engineer ~93%
• AI & Machine Learning Specialist ~82%
• Software & Applications Developer ~57%

Approximately 71% of H-1B visa approvals in FY 2024 went to Indian nationals. Combined with China's 12% share, these two countries accounted for nearly 84% of all H-1B issuances (283,397 plus 46,680 approvals). Historically, the U.S. tech industry has heavily relied on foreign talent from India, China, South Korea, and Japan. Major tech companies like Amazon, Google, and Meta have been among the leading H-1B employers, 15% of Meta workforce alone consists of H-1B workers.

Since September 21, 2025, the Trump administration changed the H-1B visas fee from previously $2,000-$5,000 to $100,000 fee. It's worth noticing that this fee applies only to new petitions outside the United States and does not affect renewals of current visa holders. From a forward-looking perspective, F-1 student visa issuances in May 2025 declined by 22% year over year. Indian students saw the steepest decline, with approximately 40% fewer visas issued. Consequently, there is growing concern that major tech companies could be severely affected by worker shortages in this sector if immigration policy does not change.

Companies and policymakers are pursuing several strategies to address this challenge:

1. Near-shoring to Latin America. The Latin American IT outsourcing market is projected to growt at a 9% CAGR. Companies are adopting a hybrid model: US firms take care of architecture and design, and Latin American teams manage implementation and testing. There are several advantages including cost savings, cultural similarity, and time zone alignment. The region hosts over 2 million tech professionals, with Mexico, Brazil, and Colombia leading in talent availability (Stargate Argentina and the $25 billion AI data center project in Patagonia)
2. Skills-Based Hiring & Education Reform. 89% of tech companies report satisfaction with hiring workers based on their abilities rather than traditional educational credentials. Workers without a traditional college degree are 20% more likely to remain in their positions compared to degree-holding peers. Companies like Provalus, featured in the Wall Street Journal in October 2025, are bringing technology jobs to rural America by hiring based on attitude rather than technical skills and providing robust training programs.
3. Up-skilling and Re-skilling. A survey from 2023 IBM Institute for Business Value questioned 3,000 C-suite executives across 20 industries and 28 countries. The results estimated that, due to the integration of AI and automation into operations, 40% of the global workforce will need re-skilling over the next three years.
4. Remote Work & Global Talent. Some companies are expanding international remote work arrangements and establishing operations in countries with more favorable immigration policies.

Several nations are actively competing to attract talent: China launched its K-class visa in September 2025 to target young STEM (Science, Technology, Engineering, and Mathematics) professionals. South Korea introduced the K-Tech Pass in April 2025, and the UK, Germany, and New Zealand are also loosening visa rules to attract skilled migrants.

Research from the Peterson Institute for International Economics found that between 1990 and 2010, rising numbers of H-1B holders contributed to 30-50% of all productivity growth in the US economy. Without addressing either the domestic skills gap through education reform or maintaining pathways for international talent, new immigration restrictions may threaten the major investments projected through 2029. Several implications may emerge:

• Shifting tech operations and innovation overseas
• Slowing AI adoption and implementation
• Losing competitive advantage to countries with more flexible talent policies

The trade-off between protecting domestic workers and maintaining access to global talent remains one of the biggest challenges facing the tech industry from now through 2029.

POLITICAL SUSTAINMENT

Over the past decade the United States governments supported the investment of the tech industry, underlining that AI is a central geopolitical issue with the potential to undermine the competitive edge role of the U.S. in the world. The Trump administration, in its second term, has pursued an aggressive policy of deregulation and AI investment. To achieve these objectives, it has focused on three pillars:

1. Removal of regulatory barriers. The administration revoked Executive Order 14110, President Biden's October 2023 directive on "Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence" on January 20, 2025. Three days later Trump administration issued a new Executive Order "Removing Barriers to American Leadership in Artificial Intelligence"
2. Accelerating data center permits and ensuring low-cost energy. In July 2025, the administration issued an Executive Order on "Accelerating Federal Permitting of Data Center Infrastructure," which directs federal agencies to streamline environmental reviews, and utilize federal lands for data center development: Idaho National Laboratory, Oak Ridge Reservation, Paducah Gaseous Diffusion Plant, and Savannah River Site. This approach includes revoking Biden-era requirements related to DEI and climate considerations for data center projects on federal land.
3. Strengthening diplomatic and security leadership through strategic positioning and technological protection. The administration has taken a mixed approach to semiconductor export controls on China. Actions tightening controls include adding 140 Chinese entities to the Entity List in December 2024 and requiring Nvidia to apply for export licenses for its H20 GPU to China in April 2025. However, the administration also rescinded the Biden-era "AI Diffusion Rule" and, most recently in December 2025, redefined Washington's export control approach by allowing Nvidia to sell H200 chips (and AMD to sell similar products) to China in exchange for a 15-25% revenue share to the U.S. government. Cybersecurity standards remain focused on threats from adversarial nations, with enhanced enforcement to protect AI from theft, malicious use, and hostile actors.

These 2025 policies from the Trump administration build upon the CHIPS and Science Act (Biden, 2022), which was designed to reduce dependence on Taiwan and Asia more broadly for semiconductors. The Act aims to increase the U.S. share of global leading-edge logic chip production from 0% today to approximately 20% by 2030 and 28-30% by 2032. Additionally, it seeks to increase overall U.S. semiconductor manufacturing capacity from the current 10-12% of global production.

However there are several uncertainties regarding politics and international relations:

• Tension between federal deregulation and state regulation. The Trump administration has drafted an executive order that would block states from enforcing AI regulations and direct the Department of Justice to establish an AI Litigation Task Force
• International regulatory divergence (USA, EU, and China). The EU is particularly moving toward stricter regulatory frameworks. The EU's Artificial Intelligence Act, adopted in March 2024, imposes comprehensive rules on AI development with strong emphasis on safety, transparency, and ethics. This divergence could complicate efforts to establish global AI governance standards and limit U.S. influence in shaping international norms
• Potential for future administration to reverse course. Given the significant policy shifts under the Trump administration on AI governance, businesses operating in this space face uncertainty about long-term regulatory stability

INVESTMENT IMPACT ON U.S. ECONOMY

For the first time in United States history, the singular AI and data center sector is sustaining the entire national economic growth, creating a structural dependency and raising questions about the sustainability of the American economy.

As previously described, Harvard economist Jason Furman's analysis, published on September 27, 2025, revealed a striking economic reality: investment in information-processing equipment and software—largely tied to AI data centers—represented only 4% of U.S. GDP in the first half of 2025, yet it accounted for 92% of GDP growth during that period. Excluding these technology-related categories, GDP growth would have been just 0.1% on an annualized basis—a near standstill that underlines the increasingly pivotal role of high-tech infrastructure in shaping macroeconomic outcomes.

This finding was further underscored by Renaissance Macro Research, which estimated that by August 2025, the dollar value contributed to GDP growth by AI data-center buildout had surpassed U.S. consumer spending for the first time ever—remarkable considering consumer spending typically accounts for two-thirds of GDP.

The concentration of growth in a single sector has raised significant concerns:

Structural dependency: Tech giants such as Microsoft, Google, Amazon, Meta, and Nvidia have poured tens of billions of dollars into building and upgrading data centers, responding to explosive demand for artificial intelligence and large language models. According to Morgan Stanley's Chief Investment Officer Lisa Shallet, "hyperscaler capex on data center and related items has risen fourfold and is nearing $400 billion annually," with the top 10 spenders accounting for nearly a third of all spending. She noted that data center-linked spending is adding roughly 100 basis points to U.S. real GDP growth.

Economic vulnerability: Other sectors—from manufacturing and real estate to retail and services—contributed little or even detracted from overall output in the first half of 2025. Job creation has slowed, raising concerns that without technology investment, the U.S. economy would have slipped into recession. U.S. manufacturing has been in recession for over two years.

Sustainability questions: The Bank of England warned in October 2025 that market valuations for AI companies are "increasingly irrational" and "appear stretched, particularly for technology companies focused on Artificial Intelligence." This concentration within market indices "leaves equity markets particularly exposed should expectations around the impact of AI become less optimistic."

Furman himself acknowledged that without the AI boom, "we would probably have lower interest rates [and] electricity prices, thus some additional growth in other sectors. In very rough terms that could maybe make up about half of what we got from the AI boom." Nevertheless, the structural reliance on a single sector for nearly all economic growth represents an unprecedented situation in U.S. economic history.

Despite massive AI investments:

• Consumer economic sentiment is at its worst since 1997 (according to Deloitte's 2025 Holiday Survey, 57% expect the economy to weaken — the most pessimistic outlook since tracking began in 1997, worse than even the 2008 Great Recession's 54%)
• Seasonal retail hiring is projected to fall to its lowest level since the 2009 recession (under 500,000 positions expected in Q4 2025, the smallest seasonal gain in 16 years)
• Manufacturing production has been in contraction for eight consecutive months (ISM Manufacturing PMI at 48.7 in October 2025, following a brief 2-month expansion after 26 straight months of contraction)

Various explanations exist for this disconnect, but clearly despite GDP expansion, the average American is not benefiting due to: (1) concentration of benefits in particular geographic regions where data centers and tech infrastructure are located, (2) wages accruing primarily to high-skill workers in specialized tech roles, and (3) the long-term nature of these investments, which have not yet been completed and thus haven't generated their full economic multiplier effects.

PRODUCTIVITY REVOLUTION

We are still in an early implementation phase of AI, and therefore future effects are hypothetical and depend on several milestons being achieved, such as data center construction to meet enterprise demand. Goldman Sachs Research estimates that generative AI, once fully adopted and incorporated into regular production, will raise U.S. labor productivity by around 15% over roughly 10 years. During this transition, unemployment may temporarily increase by approximately 0.5 percentage points during the AI implementation period, though such displacement typically disappears within two years as new jobs are created.

In August 2025, MIT's NANDA (Networked Agents and Decentralized Architecture) initiative found that 95% of enterprise AI pilot programs fail to deliver measurable financial returns and remain stuck at early stages, unable to scale. Meanwhile, approximately 21% of U.S. workers reported using AI at work as of late 2025, up from 16% a year prior (Pew Research).

Productivity in 2024 grew approximately 1.3-1.5%, which is within the historical range (long-term average approximately 1.5-2.1%), meaning no AI-driven productivity improvement is yet detectable at the macro level. If there is mass adoption of AI, this could lead to productivity improvements in the future.

The current tech investment model involves substantial upfront investments for future gains — productivity effects from AI fit within this framework. Historical precedents for productivity gains from general-purpose technologies suggest waiting 5-10 years at minimum (Goldman Sachs estimate for AI), though earlier technologies like electricity took 20-40 years before productivity booms materialized. On this point, it should be noted that some sectors are slowly integrating AI — including healthcare, education, and government — which represent approximately 40% of the economy. Meanwhile, sectors like technology and finance, representing approximately 30% of the economy, are adopting AI much more rapidly.

INFLATION

Potential Inflationary Pressures

The most plausible cause of inflation increase will be electricity consumption, with AI systems and data centers currently responsible for approximately 15-21% of total data center power consumption, projected to reach 35-50% by 2030. Wholesale electricity prices have already increased by up to 267% over the past five years in areas near data centers (Bloomberg, September 2025). A study from Carnegie Mellon University and North Carolina State University projects that data centers and cryptocurrency mining could be responsible for an average 8% increase in U.S. electricity bills by 2030, with some regions in Northern and Central Virginia potentially exceeding 25%. Companies building data centers are competing for scarce infrastructure, driving up prices for essential electrical equipment such as transformers (up 80-400% since 2020), switches and breakers, with natural gas turbines largely sold out through the end of the decade.

If AI adoption produces productivity improvements, this could reduce production costs across various sectors, potentially creating deflationary pressure that would offset inflation caused by rising energy costs. However, this deflationary effect is expected in the long run, while near-term inflationary pressures from infrastructure build-out may persist.

Sustainability Question

The impact on inflation will depend on several outcomes:

1. Whether energy supply can meet demand, considering both renewable energy and fossil fuels
2. Whether investments will lead to productivity increases
3. Whether this represents a speculative bubble that could collapse

FUTURE OUTLOOK

Companies building AI infrastructure today are positioning themselves to capture economic value from AI in the coming decades. AI infrastructure construction parallels both 19th-century railroad construction and early 20th-century electrification in terms of the scope of disruptiove innovation each brought. At its peak, U.S. railroad construction consumed around 5% of GDP, representing one of the largest infrastructure buildouts in history. Similarly, AI-related capital expenditures now account for approximately 5% of GDP. AI has all the characteristics of a general-purpose technology (GPT); the OECD confirmed in June 2025 that generative AI exhibits the defining GPT characteristics: pervasiveness, continuous improvement over time, and innovation spawning.

Economists are questioning the sustainability of current growth patterns. The concentration of growth in technology infrastructure investment raises questions in terms of long-term stability. If tech industry investments slow, the entire U.S. economy could face an abrupt slowdown. In contrast, Apollo Global Management Chief Economist Torsten Sløk has noted that economists have consistently underestimated economic resilience throughout 2025, arguing that the AI transformation might be more sustainable than feared—though he has also separately warned that AI stock valuations show signs of bubble-like conditions similar to the 1999 dot-com era. Another fundamental factor is the geographic concentration in a few specific zones, which leads to a concentration of benefits but distributed economic pain if the market were to conclude that growth is unsustainable, leading to economic stagnation.

A fundamental question is verifying whether the business model is profitable for companies. We face large investments for the next 3-4 years; if all goes well, the first economic returns could begin to materialize in the next 2 years. However, evidence suggests significant uncertainty: Morgan Stanley projects GenAI revenue could surpass $1 trillion by 2028 with positive ROI as soon as 2025 for some companies. Yet MIT research (August 2025) found 95% of enterprise AI pilot programs fail to deliver measurable financial returns, and J.P. Morgan estimates the AI industry needs $650 billion in annual revenue to deliver a mere 10% return on investments through 2030. The timeline for realizing AI gains varies significantly across business sectors, with average benefits taking several years to materialize at the enterprise level, while infrastructure-level returns may require 5-10+ years.

It is worth noting that some companies with disruptive innovations have required extended periods of investment before achieving profitability. Tesla, for instance, took 17 years to post a positive annual net income (2003–2020), yet its stock appreciated substantially throughout that period despite persistent losses. The companies analyzed here differ fundamentally: they possess 20 to 50 years of consolidated financial history, stable revenue streams, and world-leading core businesses. The key consideration remains how the market evaluates this evolving dynamic.

CRITICAL QUESTIONS AND RELATED KPIs

1. How Long Will It Last? If AI investment continues for 5+ years: Sustainable growth. If it slows within 2-3 years: Recession probable.
2. When Do Productivity Gains Arrive? If they materialize 2026-2028: Investment justified. If delayed beyond 2030: Overcapacity and probably wasted capital.
3. Can the Consumer Economy Recover? If wages grow and confidence returns: Balanced growth possible. If stagnation persists: AI alone cannot sustain the economy.
4. Is Geographic Inequality Sustainable? If benefits spread (infrastructure deployment nationwide): politically stable. If concentration persists: political backlash, policy restrictions.
5. How will rapid depreciation of AI infrastructure affect long-term investment returns?
6. What risks does dependence on a small number of companies pose to the AI ecosystem?
7. What happens if returns don't materialize—will companies cut AI spending?
8. Could cost reductions from in-house GPU production (such as Google's TPU) disrupt the current investment model?

THE 5-10 YEAR OUTLOOK

Optimistic Case:

• Adoption accelerates (30% → 70% of workers using AI)
• Productivity gains compound
• 2030: +0.5-1.0 percentage points annual productivity growth
• Justifies current investment

Pessimistic Case:

• Adoption plateaus (many workers/sectors resist)
• Gains limited to narrow use cases
• 2030: +0.1-0.3 percentage points (marginal improvement)
• Investment partially wasted

Jerome Powell's mandate as Federal Reserve Chair will expire in May 2026 in May 2026, after which a Chair more aligned with Trump's policies is expected to be appointed—one who could further lower interest rates and thus channel capital into the U.S. stock market. Speculation about who might replace Powell has already been underway since mid-2025, with an announcement potentially coming by year-end. This represents a potential buy signal for stocks, particularly if coupled with a possible December rate cut, though Fed officials remain divided on this decision.

Many investment institutions have expressed concerns about a potential AI-related financial bubble, similar to the dotcom bubble. Fed Chair Jerome Powell has spoken favorably about AI's business model, saying that unlike dotcom-era firms, today's AI companies 'actually have earnings' and 'business models and profits.' He emphasized this distinguishes AI investments from the speculative excesses of the late 1990s.

CONCLUSIONS

Tech investments and the transition toward AI implementation across all sectors of society represent a historic moment for the United States, with profound implications for GDP growth, employment, regional development, and America's global competitiveness. While critical concerns remain regarding the sustainability and geographic concentration of these investments, there is a growing sense that AI infrastructure is becoming as fundamental to modern economies as electricity and telecommunications were to the previous economic cycle. Meanwhile, numerous macroeconomic, societal, and geopolitical factors require monitoring: revenues, unemployment, geopolitical tensions, and more.

Tech AI-related private equities are trading at elevated multiples, consistent with periods of intense thematic investment—an inevitable consequence when capital chases what may be a once-in-a-century technological shift. This is a capital-intensive industry with massive investment requirements, and valuations reflect those expectations of outsized returns. Over the next five years, the pace of investment will likely moderate as the cost of AI infrastructure declines—driven in part by custom silicon alternatives to Nvidia GPUs that major tech companies are developing in-house. However, these efforts could be rendered obsolete if Nvidia's ongoing R&D produces a breakthrough product that fundamentally reshapes the competitive landscape. The central question is whether returns will ultimately justify this period of unprecedented capital deployment. Meanwhile, demand for AI infrastructure continues to grow as an increasing number of workflows become dependent on AI-assisted processes.

Like the automobile, electricity, and the personal computer before them, AI should be considered a general-purpose technology that is revolutionizing society. How exactly society will be transformed remains difficult to predict. Just as no one in the 1990s could have anticipated how transformative technology would become—the extent to which it would reshape every aspect of life—today we struggle to imagine a future that remains uncertain and perpetually under construction. Despite this uncertainty, there appears to be room for venture capital investments, given the speed at which innovations emerge, develop, and propagate through society.

The deployment of AI infrastructure across national and international territories involves many intermediate stages, each presenting potential pitfalls. Every aspect—from employment to energy supply to national and state policies—represents a potential point of friction that must be carefully managed. Whether this transformation will be fully realized remains uncertain given the complexity and scale involved. What is clear, however, is that debate and speculation about AI will dominate the next decade of politics, economics, and finance at both national and global levels.

The AI infrastructure boom represents both America's greatest economic opportunity and its greatest vulnerability. Success requires not just massive capital—which investors have committed—but resolution of critical labor shortages (particularly in construction, where worker scarcity could delay most data center projects), realization of uncertain productivity gains, and management of the boom-bust cycles that have characterized every previous technology cycle. The current policy of rapid, deregulated buildout may position the U.S. ahead of China in the short term, but it also concentrates unprecedented risk in a single sector of the economy.