📨 AI’s Power Play: Nuclear Bets & Data Risks

This week, the AI sector faces a critical pivot as massive energy demands drive a nuclear and gas infrastructure boom, while frontier model developers secure record-breaking capital amid intensifying scrutiny over data governance and workforce restructuring.

New research shows data centers have nearly tripled US demand for gas-fired power in 2 years. Gas projects explicitly linked to data centers increased 25-fold since 2024, now accounting for over a third of new demand. This massive build-out, if completed, could increase the US gas fleet by 50%, raising significant concerns about greenhouse gas emissions and regulatory oversight. [Wired]

The massive energy demands of AI data centers are driving unprecedented investment in next-generation nuclear power. These advanced plants are seen as a potential solution to provide the huge computational appetite with a reliable energy supply, being potentially cheaper to construct and safer to operate than traditional nuclear facilities. This trend highlights a critical infrastructure shift for the AI sector. [MIT Tech Review]

Anthropic CEO Dario Amodei projects powerful AI within 1-2 years could match Nobel Prize winners across fields. Key risks identified: autonomous AI rebellion, bioweapon creation (models already approaching dangerous capability thresholds), authoritarian surveillance states (particularly China), and 50% displacement of entry-level white-collar jobs within 1-5 years. Amodei advocates chip export controls, transparency regulations, and Constitutional AI training while acknowledging 10-20% annual GDP growth potential if risks are managed. [Amodei’s Essay]

Amazon reported over one million instances of child sexual abuse material found in its non-proprietary AI training data in 2025. This highlights the grave risks of the rapid AI data race and the challenge of amassing large datasets without proper safeguards. Unlike peers, Amazon has not provided source details, hindering law enforcement and raising concerns about data governance and regulatory compliance. [Bloomberg]

SoftBank in talks for $30B OpenAI investment as part of $100B round valuing company at ~$830B. Anthropic closed oversubscribed $10B round, potentially reaching $20B with Microsoft/Nvidia participation [TechInAsia]. Anthropic significantly raised revenue forecasts to $18B (2025) and $55B (2026), signaling aggressive growth expectations [CNBC]. Combined $110-120B in fresh capital demonstrates unprecedented investor confidence in frontier AI model development despite profitability questions [Reuters].

Big Tech earnings reveal investors now demand revenue growth to justify AI spending. Microsoft dropped 10% despite $37.5B quarterly spending, as Azure growth barely exceeded expectations. Microsoft faces additional pressure: OpenAI represents 45% of its cloud backlog ($280B at risk) while losing ground to competitors. [Reuters]

This week highlights AI companies with notable activity in legal tech, enterprise sales automation, and strategic expansion into Japan's regulated sectors:

Synthesia's $200M raise at $4B valuation (Google Ventures-led) signals enterprise AI training demand [Synthesia]. Ricursive Intelligence's $300M Series A at $4B marks rising capital flow to Nvidia chip challengers [Pitchbook]. Former Sequoia partner's Blockit secures $5M seed for AI scheduling automation [X]. Healthcare AI divide emerges: OpenAI/Anthropic deploy consumer chatbots as entry points [e.g. ChatGPTHealth], but analysts argue real $110B+ value lies in infrastructure: Microsoft/Google embedding AI into hospital systems and electronic health records where platforms outperform products in regulated environments [Forbes].

Pinterest announced layoffs of under 15% of its workforce (impacting hundreds of workers) to reallocate resources to AI-focused roles and prioritize AI-powered products. This restructuring, expected to incur up to $45 million in pretax charges, signals a clear strategic shift toward AI integration in its core offerings, following a trend seen across the tech industry. [APnews]

A new study utilizing AI to analyze 1.2 million satellite images found that global floating algae blooms are rapidly expanding. Macroalgae, including Sargassum, increased by 13.4% annually in the tropical Atlantic, signaling a major "regime shift." This AI-driven discovery underscores the critical investment opportunity in AI for large-scale environmental monitoring and its impact on marine industries. [University South Florida]

👀 Energy Infrastructure Scaling: Monitor the conversion of gas-fired power proposals into operational sites as data centers face a 25-fold increase in energy demand and potential turbine shortages.

👀 Data Governance Regulation: Watch for new compliance standards following Amazon’s massive data contamination report, which may force AI companies to provide greater transparency regarding external training sources.

👀 Workforce AI Integration: Track whether Pinterest’s 15% staff reduction for AI reallocation sets a precedent for other mid-cap tech firms seeking to optimize margins for computational spending.

📅 February 3, 2026 | 🌐 Virtual Livestream (9 AM–8 PM PST)

The 2nd annual Cisco AI Summit features live conversations with AI's influential leaders including Jensen Huang (NVIDIA), Sam Altman (OpenAI), Marc Andreessen (a16z), and executives from Microsoft, AWS, Google, Intel, Anthropic, and Meta. Topics span frontier models, geopolitics, infrastructure, silicon supply chains, and enterprise AI deployment. Free virtual attendance (no registration required).

We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run. Roy Amara

That’s it for this week.

Until next time,

The CLNM Capital

[Edited with the assistance of LLMs]

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