Microsoft is deepening its commitment to artificial intelligence infrastructure with plans to source electricity from a massive new 2.67-gigawatt natural gas power plant in West Texas, a move that highlights the growing tension between Big Tech’s AI expansion and its climate commitments.

The project, known as Project Kilby, will be developed through a partnership between Microsoft and Chevron under a 20-year power purchase agreement that will provide dedicated electricity for Microsoft’s AI and cloud data center operations.

Chevron described the facility as being “among the largest co-located natural gas power and data center developments in the U.S.,” underscoring the scale of the energy demands now being created by the artificial intelligence boom.

A shift taking place across the technology industry has seen companies race to secure reliable electricity supplies for increasingly power-hungry AI workloads. While technology giants have spent years championing renewable energy and carbon-reduction targets, the explosive growth of AI is forcing many firms to confront a difficult reality: current clean-energy infrastructure often cannot provide the consistent, around-the-clock power that advanced AI systems require.

Project Kilby will rely primarily on two large turbines supplied by GE Vernova, while additional generation capacity will come from Solar Turbines, a subsidiary of Caterpillar.

The choice of natural gas rings loud because Microsoft has positioned itself as one of the world’s most aggressive corporate advocates of sustainability. The company has pledged to become carbon negative by 2030 and has invested heavily in renewable energy projects, carbon removal technologies, and environmental initiatives.

Artificial intelligence data centers consume significantly more electricity than traditional cloud computing facilities. Training and operating advanced AI models requires thousands of specialized chips running continuously, creating enormous power demands that can rival those of major industrial facilities.

Industry analysts view access to electricity as one of the most important competitive advantages in the AI race. Technology companies are no longer competing solely for chips and talent; they are also competing for energy. That dynamic has triggered a surge in partnerships between technology firms and energy providers. Across the United States, utilities, oil and gas companies, nuclear developers, and renewable energy producers are all seeking to capitalize on growing demand from AI operators.

For Microsoft, the attraction of a dedicated power source is straightforward. Unlike wind and solar generation, natural gas plants can provide stable baseload power regardless of weather conditions. That reliability is valuable for AI data centers, where interruptions can disrupt critical computing operations and reduce utilization of expensive hardware.

The arrangement tilts toward co-located power generation, where energy facilities are built specifically to serve nearby data centers rather than relying entirely on regional electricity grids. Such projects can reduce transmission constraints and provide greater certainty for both power suppliers and technology companies.

However, the environmental implications are substantial.

According to the Environmental Integrity Project, Project Kilby could emit more than 13 million tons of carbon dioxide, alongside approximately 3,200 tons of conventional air pollutants and 278,000 pounds of hazardous air pollutants over its operating life.

Those figures are likely to intensify scrutiny of the environmental cost of AI development. The industry has increasingly faced questions about whether sustainability goals remain achievable as AI deployment accelerates. Data centers already account for a growing share of global electricity consumption, and forecasts suggest power demand from AI could increase dramatically over the next decade.

Microsoft’s decision may signal a broader reassessment taking place within the technology sector. While renewable energy remains a long-term objective, companies now appear willing to prioritize energy security and computing capacity in the near term.

The move also highlights the changing fortunes of the natural gas industry. For years, many investors expected technology companies to become major drivers of renewable energy adoption. Instead, AI has created a new source of demand for gas-fired generation, extending the relevance of fossil fuels even as clean-energy investments continue to grow.

As companies including OpenAI, Anthropic, Google, Meta, and Amazon expand AI infrastructure, the pressure on energy systems is expected to intensify. Utilities and energy producers are increasingly taking a position as essential partners in the next phase of the AI economy.

Project Kilby, therefore, represents more than a power agreement. It is largely seen as an illustration of a new reality emerging in the technology sector: the race to dominate artificial intelligence is becoming inseparable from the race to secure electricity.

In one of the most dramatic corporate turnarounds in South Korea’s modern business history, SK Hynix has overtaken long-dominant Samsung Electronics to become the country’s most valuable listed company, a milestone powered almost entirely by its commanding position in the specialized memory chips fueling the artificial intelligence boom.

This comes weeks after the chipmaker hit $1 trillion market valuation.

Shares of SK Hynix closed up 5.6% on Monday, lifting its market capitalization to 2,080.4 trillion won ($1.35 trillion). Samsung’s stock eased 0.14%, leaving it with a market value of 2,066.7 trillion won, excluding preferred shares. Including those shares, Samsung’s total valuation stood at 2,246.4 trillion won, but the symbolic shift at the top of the leaderboard marks a profound change in the fortunes of two companies that have defined South Korea’s technological prowess for decades.

The reversal is all the more remarkable given SK Hynix’s brush with collapse just two decades ago. In 2002, then-Hynix Semiconductor was crippled by debt from an aggressive expansion and nearly sold to Micron in a deal that ultimately fell through. The company spent nearly a decade under creditor control, its shares plunging as low as 135 won in 2003 and earning it the derisive label of a “penny stock.”

Today, SK Hynix stands as the world’s most valuable memory chipmaker and a central player in the AI supply chain. Its dominance in high-bandwidth memory (HBM) chips, specialized, vertically stacked components that deliver faster performance and lower power consumption for AI processors, has transformed it from a commodity producer into an indispensable partner for the likes of Nvidia and Google.

“The emergence of customized AI memory fundamentally changed the industry’s economics and allowed SK Hynix to establish itself as the market leader,” said Kim Sunwoo, a senior analyst at Meritz Securities.

By 2025, SK Hynix had captured 61% of the global HBM market, far ahead of Samsung’s 17% and Micron’s 21%. Unlike conventional DRAM, HBM chips are tightly integrated with AI accelerators, creating high barriers to entry and giving leading suppliers significant pricing power and profitability.

A Deliberate Bet on the Future During Industry Downturns

SK Hynix’s success traces back to a strategic decision to keep investing heavily in HBM technology even during the severe memory downturn of 2023, when the company posted a record annual operating loss of 7.73 trillion won. That bet paid off spectacularly as the AI frenzy took hold. In 2024, SK Hynix reported an operating profit of 23.5 trillion won, a record at the time, and its shares have skyrocketed more than 340% this year alone.

SK Group Chairman Chey Tae-won, who faced internal opposition when the conglomerate acquired Hynix years ago, reflected on the vision in a book published in January.

“What I really wanted to accomplish when we acquired Hynix was to transform it from a commodity memory producer into a mainstream semiconductor company whose products are indispensable. In the past, it did not matter whether memory came from Hynix, Samsung or Micron. They were interchangeable commodity products. HBM is different. If SK Hynix’s HBM is replaced with another product, the AI system may not function properly. What used to be a peripheral component has become a core component,” he said.

This focus has allowed SK Hynix to thrive in an industry where AI has fundamentally altered the economics. While Samsung maintains a broader portfolio that includes logic chips, smartphones, and consumer electronics, SK Hynix’s narrower but deeper specialization in memory has proven highly advantageous in the current cycle.

Analysts now see Samsung’s long-held position as the world’s largest DRAM producer coming under increasing threat. Bank of America estimates SK Hynix’s monthly DRAM output will reach about 589,000 wafers this year, compared with Samsung’s roughly 691,000. However, SK Hynix is projected to expand DRAM production by 38% between 2025 and 2028, versus just 17.5% growth at its rival. That would narrow the production gap to less than 10% by 2028 from around 23% this year — a significant achievement given Samsung’s larger manufacturing base.

“Previously, the difference in manufacturing scale meant there was simply no way for rivals to close the profitability gap with Samsung,” Kim said.

SK Hynix’s ascent highlights how the global AI surge is reshaping not just individual companies but entire national economies. South Korea, already heavily dependent on semiconductors for its export performance, now has two firms at the absolute pinnacle of the industry, both riding the wave of hyperscaler spending from Microsoft, Google, Meta, and others.

The development also comes as SK Hynix prepares for a U.S. listing on Nasdaq, a move expected to broaden its investor base and further elevate its global profile. The timing could hardly be better, as investor appetite for AI-related plays remains robust despite periodic market volatility tied to geopolitical events like the Iran conflict.

For Samsung, industry analysts believe the shift serves as a reminder of the need to adapt. While the company remains a technology behemoth with unmatched manufacturing scale and diversification, its memory business faces intensifying competition in the most profitable segments of the AI era. Samsung has responded by accelerating its own HBM development, but SK Hynix’s early lead and focused execution have given it a clear edge in the current cycle.