AI Profit Rift Fuels Samsung Strike.

Mazen Al Sabbagh

— 6 min read
Share

TL;DR

  • Samsung at Brink: South Korea's Prime Minister warns of "fatal damage" as Monday marks the final negotiation opportunity before a May 21 strike by 50,000+ chip workers. The core issue: an AI-driven profit surge that benefits unevenly.
  • BlackRock Anchors SpaceX IPO: The firm is in discussions for a $5-10B stake in SpaceX's projected $1.75T IPO, an unprecedented institutional commitment. The S-1 filing is anticipated as early as Wednesday for a June 12 listing, implicitly valuing xAI.
  • Triple-Front Monday: Critical AI-related decisions converge on May 19: Google I/O unveils next-gen Gemini, Musk v. Altman jury deliberations commence, and Samsung faces its strike deadline.
  • Vatican AI Mandate: Pope Leo XIV establishes the Vatican AI Commission, its first formal body to address AI's global impact, preceding the Magnifica Humanitas encyclical on AI and peace.
  • AI Employment Rift: Conflicting data emerges: Bloomberg/BLS indicates 10M exposed jobs declining, Goldman reports 16,000 monthly losses, yet CNBC finds 56% of companies citing AI layoffs experienced stock declines.

Lead Story: AI Profit Rift Fuels Samsung Strike

South Korea's Prime Minister, Kim Min-seok, issued an exceptional public warning Saturday, stating that a Samsung chip worker strike would inflict "fatal damage" and "unimaginable" economic harm. Reuters confirmed the government's readiness to explore "all options including emergency arbitration." Samsung's executive leadership engaged in rare direct talks with union representatives Friday evening.

The impending May 21 walkout, involving over 50,000 employees for 18 days, carries substantial economic weight. Samsung has already initiated emergency protocols, withdrawing 360,000 wafers from its Pyeongtaek DRAM lines in a controlled "warm-down." Daily losses are projected between $700 million and $2 billion, with total damages potentially reaching 100 trillion won if materials are rendered unusable.

Monday's direct negotiations represent the final viable opportunity to avert industrial action. The union demands a 15% share of operating profit in performance payouts and the removal of a 50% base-salary bonus cap. Samsung has offered a one-time 2026 payment but resists fundamental adjustments to its bonus structure.

A Reuters weekend exclusive illuminates the undercurrent of the dispute: the AI boom itself. Samsung's Q1 2026 operating profit surged nearly eightfold, propelled by HBM memory chips—sold out through 2026 due to Nvidia demand. However, these gains were not uniformly distributed, largely favoring HBM workers over logic and foundry engineers. This internal disparity is exacerbated by rival SK Hynix, which settled with its union last September to allocate 10% of operating profit directly to workers, translating to $460,000-$477,000 per employee this year. Samsung's workforce is acutely aware of this competitive compensation benchmark.

Micron and SK Hynix equities have already seen upward movement on anticipated supply scarcity. Should this strike proceed, it would severely constrain the HBM memory supply chain, which is critical for every major AI chip currently in production, at a time when DRAM contract prices have already experienced a 90-95% quarter-over-quarter increase.

In Other News

BlackRock in talks for $5-10B SpaceX IPO anchor stake. The Information reported Saturday that BlackRock, the world's largest asset manager, is negotiating one of the largest single institutional commitments to any IPO in history. SpaceX shareholders have authorized a 5-for-1 stock split ahead of its June 12 Nasdaq listing under the ticker $SPCX. The public S-1 prospectus is expected as early as Wednesday, with roadshows commencing the week of June 8. This IPO, targeting a $1.75-2T valuation and a $75B raise, would be the largest ever—and given xAI's subsidiary status, it effectively represents the first public market valuation of Musk's AI ventures. CNBC's Jim Cramer has cautioned that the IPO could function as a "liquidity black hole" for the broader market.

Pope Leo XIV formally establishes Vatican AI Commission. The Pope signed a rescript Saturday creating an Interdicasterial Commission on Artificial Intelligence, signifying the Vatican's first official cross-departmental body dedicated to studying AI's human impact. On Sunday, during the 60th World Day of Social Communications, Leo XIV addressed pilgrims in St. Peter's Square, cautioning against AI's potential to diminish "human voices and faces." The commission's formation precedes his inaugural encyclical, Magnifica Humanitas, expected by month's end, which will address AI, autonomous weaponry, and global peace. AP noted its motivation by "the increase in usage of AI, its potential effects on human beings and on humanity as a whole."

Recursive Superintelligence emerges from stealth with $650M. Former Salesforce Chief Scientist Richard Socher launched the lab Wednesday, securing a $4.65B valuation with backing from Google Ventures, Greycroft, AMD Ventures, and Nvidia. The team, comprising talent from OpenAI, DeepMind, Meta AI, and Uber AI, includes the notable AI researcher Peter Norvig. Their stated mission: to develop AI systems capable of self-improvement without direct human intervention. Socher's emphasis on product delivery over purely academic output, combined with Nvidia's direct investment, signals significant institutional validation for this "automated AI research" paradigm.

OpenAI quietly acquired voice-cloning startup Weights.gg. The NYT's Mike Isaac reported that OpenAI completed the acquisition of the six-person startup earlier this year. Weights.gg had previously hosted a library of celebrity AI voice clones—including Taylor Swift, Samuel L. Jackson, and Trump—before ceasing its consumer operations on April 1. OpenAI publicly maintains that its internal voice-cloning tool remains "too risky" for general release. The divergence between stated policy and acquisition strategy in this sensitive area warrants close observation.

X / Social Pulse

Microsoft AI CEO Mustafa Suleyman's February FT claim that white-collar work would be "fully automated within 12 to 18 months" has re-entered a viral cycle this weekend. Fortune, LiveMint, and several other publications have released new analytical pieces over Saturday and Sunday. This resurgence coincides with fresh data: Bloomberg/BLS indicating a decline in 10 million AI-exposed jobs, Goldman Sachs estimating a displacement of 16,000 jobs per month, and the Detroit Big Three reducing over 20,000 salaried roles while simultaneously increasing AI engineering hires. A CNBC analysis published today offers a counter-narrative: of 23 S&P 500 companies announcing AI-linked layoffs, 56% experienced subsequent stock declines—Nike down 35%, Salesforce down 32%. BBC's Raspberry Pi CEO, Eben Upton, further countered, warning that overstating AI job destruction risks deterring young talent from tech careers. The discourse is rapidly bifurcating into two distinct camps: those asserting immediate, widespread displacement, and those cautioning against industry alienation.

One to Watch

Monday, May 19, is set to be the most consequential single day on the AI calendar this year. Google I/O commences at 10 AM PT, with anticipated reveals of Gemini 3.2 Flash, the continuous Gemini Spark agent, and the "Googlebook"—Google's new Gemini-centric laptop platform, which debuted at the Android Show. Leaked benchmarks suggest Gemini 3.2 Flash could achieve 92% of GPT-5.5's coding capability at a 15-20x lower cost; if accurate, this would fundamentally reprice the inference market overnight. Concurrently, jury deliberations begin in the Musk v. Altman trial (an advisory verdict is expected by week's end), and Samsung holds its final labor negotiations before the May 21 strike deadline. Nvidia's earnings report follows on Tuesday. This week will largely define the contours of Q2 2026.

Quick Hits

  • Anthropic's $30B round at $900B valuation has agreed terms, per the FT, with closing anticipated this month and IPO discussions targeting October 2026.
  • Cerebras (CBRS) concluded its first trading week at approximately $298, down from its $350 open but still up 61% from the $185 IPO price. Its 86% revenue concentration from two UAE-linked customers remains a primary risk factor.
  • AI VC funding reached $255.5B in Q1 2026 alone, surpassing the full-year 2025 total, according to PitchBook data.
  • April CPI registered a hot 3.8%, contributing to Friday's downturn in AI stocks—SoundHound -5.7%, HubSpot -4.3%, Akamai -5.4%. This presents a headwind for rate-sensitive names entering earnings season.
  • Bill Ackman's Pershing Square disclosed a new Microsoft stake on Friday, labeling MSFT a "dominant" AI beneficiary and driving the stock up +3% to $422.

The week ahead will unfold at an accelerated pace. Monday alone presents Google I/O, the commencement of Musk v. Altman jury deliberations, and Samsung's final opportunity to prevent a strike that could send ripples throughout global AI chip supply chains. Nvidia's earnings on Tuesday will either affirm or challenge the premise of accelerating AI infrastructure expenditure. Meanwhile, Anthropic is positioned to close a funding round that could value it above OpenAI for the first time. The critical inquiry is not whether significant events will transpire this week, but rather which of these simultaneous developments will most profoundly redefine the operational landscape.

Sources

Lock in. M. mazen@thorterminal.com

Read more