The AI Race: Why More Will Lose Than Win & Surprising Winners Like Water Cooling Systems

Artificial intelligence is advancing at a breakneck pace, sparking a global race among companies, countries, and researchers to claim leadership. The promise of AI transforming industries and economies is immense, but the reality is that more participants will lose than win in this competition. The barriers to success are high, the costs are steep, and the risks are significant. Yet, amid this intense race, some unexpected sectors are emerging as winners—one of the most surprising being water cooling systems for data centers.

This post explores why the AI race will produce more losers than winners, backed by numbers and examples. It also highlights how innovations in cooling technology are quietly benefiting from AI’s growth, offering a fresh perspective on who gains from this technological surge.

Data center with advanced water cooling systems

Data centers rely increasingly on water cooling systems to manage AI workloads efficiently.

The High Stakes of the AI Race

The AI race is not just about developing smarter algorithms. It involves massive investments in hardware, software, talent, and infrastructure. The AI race is no longer a speculative trend; it is a multi-trillion dollar infrastructure reality. According to Gartner, global AI spending is projected to hit $2.52 trillion in 2026. Despite this, only a handful of companies capture the majority of AI’s economic value.

Why Most Will Lose

• Capital Intensity: Building AI models like GPT-4 or similar large language models requires billions of dollars. OpenAI’s GPT-4 reportedly cost over $100 million to train. Many startups and smaller players cannot afford such expenses.

• Talent Shortage: AI expertise is scarce. The demand for AI researchers and engineers outpaces supply, pushing salaries to six figures and beyond. This limits who can build competitive AI systems.

• Data Access: High-quality, diverse datasets are essential for training AI. Large tech firms have an advantage due to their vast user bases and data collection capabilities.

• Infrastructure Demands: Running AI workloads requires powerful GPUs and specialized hardware. The energy consumption of training a single large AI model can equal the lifetime emissions of five cars, according to a 2021 study from the University of Massachusetts Amherst.

  
  

These factors create a steep barrier to entry. Many companies will invest heavily but fail to keep up with the pace of innovation or scale their solutions effectively.

The Environmental and Operational Challenge

AI’s growth is not just a financial challenge but an environmental one. Data centers powering AI consume enormous amounts of electricity and generate significant heat. For example, Google’s data centers use about 12 terawatt-hours of electricity annually, roughly 1% of global electricity consumption.

Managing this heat is critical. Overheated servers can fail, causing downtime and data loss. Traditional air cooling systems are often insufficient for the intense workloads AI demands.

Water cooling system components in a data center

Water cooling systems help data centers maintain optimal temperatures for AI processing.

Water Cooling Systems: The Unexpected Winners

As AI workloads grow, water cooling systems have become a vital solution for data centers. Unlike air cooling, water cooling transfers heat more efficiently, reducing energy consumption and improving server performance.

Why Water Cooling Wins

• Energy Efficiency: Water cooling can reduce cooling energy use by up to 40% compared to traditional air cooling. This translates to millions of dollars saved annually for large data centers.

• Higher Density Support: AI hardware often packs more chips into smaller spaces. Water cooling supports higher server densities without overheating.

• Sustainability: Some water cooling systems recycle water or use natural cooling sources, lowering environmental impact.

• Cost Savings: Although initial installation costs are higher, operational savings and longer hardware life make water cooling economically attractive.

  
  

Real-World Examples

• Microsoft’s Project Natick uses underwater data centers cooled naturally by ocean water, cutting cooling costs dramatically.

• Facebook’s data centers in Sweden use water cooling combined with cold Nordic air, achieving some of the lowest PUE (Power Usage Effectiveness) ratings in the industry, around 1.1 (where 1.0 is ideal).

  
  

These examples show how water cooling is not just a technical fix but a strategic advantage in the AI race.

The Winners and Losers in Numbers

• Top AI Companies: Alphabet, Meta, Nvidia, Microsoft, Amazon, and OpenAI dominate AI development and deployment. They control over 70% of AI patents and invest billions annually.

• Startups: Over 3,000 AI startups worldwide compete, but only about 10% secure significant funding or market traction.

• Cooling Market: The data center cooling market, including water cooling, is projected to reach $15 billion by 2027, growing at 8% annually.

• Winners: With Data center growth surging, we like the cooling sector as a whole. Household names include Vertiv Holdings ($VRT), Modine Manufacturing ($MOD), & one of our favorite global companies, Schneider Electric ($SBGSY- France). A "penny" for our thoughts? How about Asetek ($ASTK) currently sitting at $1.71. Reminder: We do not give financial or single stock advice as our research points to the sector as a whole.

  
  

These figures highlight the concentration of power and resources in AI, and the parallel growth of supporting industries like cooling systems.

What This Means for Businesses and Investors

Businesses entering the AI space must recognize the high risks and costs. Success requires more than just technology—it demands access to capital, data, talent, and infrastructure. Many will invest heavily but fail to achieve competitive AI capabilities.

Investors should consider the broader ecosystem. Companies providing infrastructure solutions, such as water cooling system manufacturers, stand to benefit from AI’s growth even if they are not directly involved in AI development.

Final Thoughts

The AI race is a high-stakes competition with few winners and many losers. The barriers to success are steep, and the costs are enormous. Yet, this race also creates opportunities in unexpected areas like cooling systems for data centers. These systems help manage the massive energy and heat demands of AI, making them crucial players in the AI ecosystem.

Understanding these dynamics helps businesses and investors make informed decisions. The AI race is not just about algorithms and models—it’s about the entire infrastructure that supports them. Those who recognize this will find opportunities beyond the obvious winners.