📨 AI's Speculation Costs

Markets moved billions on weak speculation while AI's hidden costs surfaced across hiring, skills, and supply chains. Meanwhile, geopolitical tensions reshaped chip exports and defense contracting.

Watch the 3-min video overview:

The Citrini report's AI unemployment predictions sent the Dow down 800 points despite offering little new analysis. A $6M karaoke company claiming AI shipping efficiencies erased billions from logistics stocks. Even Nvidia's 73% revenue jump couldn't calm investor anxiety — the stock fell 5%. AI markets now move on narrative quality, not fundamentals [Wired].

The Trump administration is drafting rules requiring foreign nations to invest in US AI data centers or provide security guarantees to receive exports of 200,000+ chips. Even small installations under 1,000 chips could need licenses. The framework mirrors Middle East deals and gives leverage to negotiate US investments [Reuters].

China's 5-year plan positions AI as a job creator to offset 300 million retirements and persistent youth unemployment, contradicting IMF predictions that AI will affect 40% of global jobs. Academics warn "job destruction often precedes creation," while robotaxis and AI agents already threaten Chinese employment despite small-scale deployment [Reuters].

Leopold Aschenbrenner's AGI-focused hedge fund grew from $1.5B to $5.5B in months, now betting heavily on electricity generation and computing infrastructure over AI model builders [Fortune].

Wrtn, a Korean interactive AI storytelling platform, is approaching $100M in annual revenue with 5M monthly users in Korea and Japan. The company's growth reflects demand for AI companionship in markets with high social isolation and aging populations.

Claude overtook ChatGPT to become the top free app in Apple's US App Store following Anthropic's public dispute with the Pentagon over AI safety guardrails [TechCrunch]. The surge accelerates a broader trend: a16z report shows Claude and Gemini grew paid subscriptions over 200% last year, narrowing ChatGPT's lead.

Global markets are dealing with 3 simultaneous crises: Middle East war driving oil prices up, AI threatening established companies while raising infrastructure costs, and rising defaults in private lending. Unlike past downturns, no single policy response addresses all 3, making recovery less predictable [Bloomberg].

Anthropic filed two lawsuits challenging the Defense Department's designation of the company as a supply-chain risk. The label followed Anthropic's refusal to allow its AI for mass surveillance or fully autonomous weapons [TechCrunch]. Five national security law experts say Anthropic has strong case against Pentagon blacklisting [Reuters].

Two studies documented AI's hidden costs: developers using AI for learning scored 17% lower on comprehension tests with no speed improvement [arXiv], while hiring of workers aged 22-25 in AI-exposed fields dropped 14% since ChatGPT launched [Anthropic].

Oracle is planning thousands of layoffs and freezing hiring in its cloud division as AI data center expansion drains cash. The company is raising $50B through debt and equity, but Wall Street projects negative cash flow until 2030. Oracle's stock has fallen 54% from its September 2025 peak [YahooFinance].

Tech giants are buying memory chips for AI data centers at premiums, leaving fewer chips for consumer devices. DRAM spot prices jumped nearly 700% in a year; NAND is rising sharply. HP says memory now costs 35% of laptop materials, up from 15-18% last quarter. Relief unlikely before 2027 [Bloomberg].

Companies are buying security tools, content production capabilities, and compute capacity — while top labs continue competing for viral AI talent through acqui-hires”:

AMI Labs (Yann LeCun) — AI researcher raised $1.03B seed round at $3.5B valuation to build "world models” — AI systems with reasoning, planning, and persistent memory beyond token prediction [NYT].

Arda (Bob McGrew) — Ex-OpenAI chief research officer raising $70M at $700M valuation to build AI platform automating factory floors with robots [WSJ].

Legora — Swedish legal AI platform valued at $5.6B after $550M round; expanding to US where it will compete directly with Harvey [CNBC].

The Pentagon can likely use AI to surveil Americans despite OpenAI's amended contract prohibiting it. Fourth Amendment protections don't cover publicly available information or commercial data the government purchases. AI transforms unregulated individual data points into detailed profiles at scale. OpenAI's contractual safeguards may not override "lawful purposes" clauses, and technical enforcement mechanisms remain unclear [MIT Tech Rev].

Caitlin Kalinowski, OpenAI's robotics and hardware lead, resigned March 7 citing concerns the Pentagon agreement was "rushed without guardrails defined." She argued surveillance of Americans without judicial oversight and lethal autonomy without human authorization deserved more deliberation [Politico].

Norwegian startup Aikido will deploy a 100kW subsea data center powered by floating offshore wind this year, scaling to 10-12MW by 2028. The small unit will live in the submerged pods of a floating offshore wind turbine. If all goes well, the company hopes to build a larger version to deploy off the coast of the U.K. in 2028 [TechCrunch].

👀 AI narrative volatility is the new risk: Flimsy speculation now moves billions. Watch for opportunistic actors exploiting market hypersensitivity.

👀 Chip diplomacy as infrastructure policy: Trump's export rules force allies to invest billions in US data centers for chip access — or find alternatives that accelerate China's self-sufficiency.

👀 China's high-stakes labor experiment: AI must replace 300 million retiring workers without destroying more jobs than it creates — academics warn the math doesn't work.

📅 April 9-10, 2026 | 🌐 Marina Bay Sands, Singapore

Tech summit connecting AI pioneers, frontier startups, investors, and digital leaders across sectors. The event drives strategic partnerships and breakthrough innovations from Singapore, Asia's innovation capital.

"The test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function."

— F. Scott Fitzgerald

Amid the recent AI apocalyptic predictions, Fitzgerald's old test for intelligence feels newly relevant: AI represents both disruption and opportunity. Technology destabilizes, humanity adapts — but the timeline matters. Those caught in the transition pay the cost. The question for investors: are you positioned for the adaptation, or exposed to the disruption?

That’s it for this week.

The CLNM Capital Team

Want to share this newsletter with a friend? Use this link: https://clnm.beehiiv.com/

[Edited with the assistance of LLMs]