OpenAI’s ambitious Stargate project, a cornerstone of the AI infrastructure race, has secured preliminary agreements with South Korean semiconductor powerhouses Samsung Electronics and SK Hynix to supply massive volumes of DRAM wafers, potentially devouring up to 40% of global production by 2029.

The deals, formalized as letters of intent in October 2025 following high-stakes meetings involving OpenAI CEO Sam Altman, South Korean President Yoon Suk Yeol, and executives from both firms, underscore the project’s unprecedented scale and its ripple effects on global tech supply chains.

The partnerships go beyond mere supply: Samsung and SK Hynix, which together command about 70% of the global DRAM market and 80% of high-bandwidth memory (HBM), have committed to ramping production to meet OpenAI’s projected demand of 900,000 DRAM wafer starts per month.

This volume encompasses commodity DDR5, low-power LPDDR4/LPDDR5, premium HBM for AI accelerators, and specialty DRAM types. Notably, initial deliveries will consist of undiced wafers—raw silicon before cutting into individual chips—granting Stargate flexibility in customization and processing. But questions linger on who will handle dicing, packaging into DRAM chips, HBM stacks, or modules, with speculation pointing to potential in-house or third-party fabs.

To contextualize the scale, global 300mm fab capacity is forecasted to hit 10 million wafer starts per month (WSPM) in 2025, per TechInsights. DRAM’s share stood at 22% (about 2.07 million WSPM) in 2024, with analysts like those at TrendForce projecting an 8.7% growth to roughly 2.25 million WSPM in 2025—meaning Stargate could claim around 40% of that capacity at peak. Broader market forecasts from Yole Group predict DRAM revenues surging to $129 billion in 2025, up from prior years, driven by AI demand, with total memory market revenues nearing $190 billion, including NAND.

However, this concentration has fueled warnings of shortages: Counterpoint Research notes 50% price hikes in 2025, with further increases into 2026, impacting consumer electronics like PCs, smartphones, and GPUs.

Deloitte’s 2025 Semiconductor Outlook highlights AI and data centers as key drivers for chip sales growth, even as PC and mobile demand lags.

Stargate, a $500 billion collaboration controlled by OpenAI, Oracle, and SoftBank, aims to deploy multiple gigawatt-scale AI data centers globally to train next-generation models. Launched in January 2025 with a request for proposals from partners, the project has accelerated rapidly.

The flagship site in Abilene, Texas, went operational early, providing 1 gigawatt of capacity. By mid-2025, Oracle pledged an additional 4.5 gigawatts, pushing totals over 5 gigawatts. September brought announcements of five new U.S. sites, escalating investments to $400 billion, and capacity toward 7 gigawatts. In October, Michigan joined with a multi-billion-dollar campus developed by Related Digital, set for construction in early 2026. December saw ground broken on a $15 billion, four-building complex in Wisconsin by Vantage Data Centers, offering nearly 1 gigawatt, and Michigan regulators approving 1.4 gigawatts of power from DTE Energy—despite local opposition over energy costs and grid strain.

These facilities demand enormous resources: thousands of servers per site, each packed with Nvidia Blackwell GPUs and ancillary gear like cooling systems and power delivery—potentially necessitating dedicated power plants, as Altman has toured Asia-Pacific to secure. The Samsung ecosystem’s involvement deepens the alliance. Samsung SDS will co-design and operate Stargate sites in Korea, provide consulting for integrating OpenAI models into enterprise systems, and act as a reseller for services like ChatGPT Enterprise in South Korea.

Samsung C&T and Samsung Heavy Industries are pioneering floating data centers—offshore platforms leveraging maritime tech for superior cooling efficiency and lower emissions, though engineering hurdles remain. Plans include two Korean “Stargate-style” facilities starting at 20 megawatts, positioning South Korea as an AI hub amid high local adoption of OpenAI tools.

Altman lauded Korea’s “incredible tech talent, world-class infrastructure, strong government support, and thriving AI ecosystem” during his visit.

Layoffs have become one of the most defining and unsettling features of the U.S. labor market in 2025, with artificial intelligence now sitting at the center of corporate explanations for sweeping job cuts.

Across technology, finance, enterprise software, and cybersecurity, companies are increasingly framing workforce reductions as a necessary consequence of rapid AI adoption, even as questions grow over whether automation is the sole or even primary driver.

Figures from consulting firm Challenger, Gray & Christmas, published by CNBC, show that AI-linked layoffs reached nearly 55,000 jobs in the United States this year. In total, employers announced about 1.17 million job cuts through 2025, the highest annual level since the height of the Covid-19 pandemic in 2020, when layoffs surged to 2.2 million. The scale and persistence of the cuts suggest a structural reset rather than a temporary correction.

The pace of job losses has remained elevated into the final quarter of the year. Employers announced roughly 153,000 job cuts in October, followed by more than 71,000 in November. AI was explicitly cited in over 6,000 of the November cuts alone, according to Challenger, underscoring how frequently the technology now features in restructuring announcements.

These layoffs are unfolding against a challenging economic backdrop. Inflation continues to erode purchasing power, tariffs have added to input costs, and borrowing remains expensive after years of aggressive interest rate hikes. For many firms, especially large multinationals under pressure to protect margins, AI has emerged as an appealing lever. Automation promises immediate efficiency gains, lower payroll costs, and faster execution, particularly in roles involving repetitive cognitive tasks.

Academic research is reinforcing those incentives. CNBC cited a study released by the Massachusetts Institute of Technology in November, which found that AI systems can already perform tasks equivalent to about 11.7% of jobs across the U.S. labor market. The researchers estimated potential wage savings of up to $1.2 trillion, with the largest exposure in finance, healthcare, legal services, and other professional sectors where routine analysis, documentation, and customer interaction are common.

Yet not everyone agrees that AI alone explains the scale of the layoffs. Fabian Stephany, assistant professor of AI and work at the Oxford Internet Institute, has argued that many companies are overstating the role of automation. He has said that firms that expanded aggressively during the pandemic-era boom are now reversing course, using AI as a convenient narrative.

“Many companies significantly overhired during Covid,” Stephany previously told CNBC. “Instead of saying ‘we miscalculated two or three years ago,’ they can now point to AI as the reason.”

In his view, a portion of the layoffs represents a delayed market correction rather than a sudden technological displacement.

Even so, corporate leaders have been unusually candid about AI’s role in reshaping their organizations.

Amazon announced in October the largest round of layoffs in its history, cutting 14,000 corporate roles as it redirects spending toward what it describes as its “biggest bets,” with AI at the top of the list. In a blog post, Beth Galetti, the company’s senior vice president of people experience and technology, said the shift was necessary to reduce layers, speed up decision-making, and give teams more ownership.

CEO Andy Jassy had earlier warned employees that AI would fundamentally change Amazon’s workforce. He said the company would need “fewer people doing some of the jobs that are being done today, and more people doing other types of jobs,” signaling a long-term rebalancing rather than a short-term cutback.

Microsoft has taken a similar path. The company has cut around 15,000 jobs in 2025, including 9,000 announced in July. CEO Satya Nadella framed the layoffs as part of a broader effort to reposition Microsoft around AI, describing a transition from a traditional software model to what he called an “intelligence engine.”

In an internal memo, Nadella emphasized that AI was not just another product line but a foundational shift.

“It’s about building tools that empower everyone to create their own tools,” he wrote, arguing that this approach required a different organizational structure and skill mix.

At Salesforce, the impact has been particularly visible in customer support. CEO Marc Benioff confirmed in September that about 4,000 customer support roles had been eliminated, reducing headcount in that division from roughly 9,000 to 5,000. Benioff has said AI is already handling up to half of the company’s workload, allowing it to operate with far fewer staff than before.

IBM’s experience highlights the uneven nature of the transition. CEO Arvind Krishna said in May that AI chatbots had replaced a few hundred human resources roles, but he stressed that the company was simultaneously hiring in areas that demand higher-level judgment, including software engineering, sales, and marketing. Still, IBM announced a 1% global workforce reduction in November, potentially affecting nearly 3,000 employees.

Cybersecurity firm CrowdStrike was among the most explicit in linking layoffs to AI. In May, it said it would cut about 500 jobs, or 5% of its workforce. CEO George Kurtz described AI as a “force multiplier” that flattens hiring needs and accelerates everything from product development to customer engagement, reducing the number of people required to scale the business.

Workday, an HR software company, moved early in the year. In February, it announced plans to cut about 8.5% of its workforce, or roughly 1,750 jobs, to free up resources for AI investment. CEO Carl Eschenbach said the decision was about prioritization, signaling that AI spending had become more important than maintaining existing headcount.

Collectively, these moves point to a deeper shift in how companies value labor. While executives insist that AI will create new roles over time, particularly for engineers, data scientists, and AI specialists, the near-term effect has been a sharp contraction in white-collar and support roles that were once considered relatively insulated from automation.

The result has been growing anxiety about job security and reskilling for workers. For policymakers, it raises difficult questions about how quickly labor markets can adapt and whether existing safety nets are adequate. As 2025 draws to a close, AI’s promise of productivity and innovation is becoming inseparable from its social cost, with layoffs serving as the most visible sign of a labor market being reshaped in real time.