The Download: weight-loss drugs slowing aging and carbon dioxide batteries

The newsletter highlights several technology developments, including drugmakers’ claims that GLP-1 patients show slower aging on molecular clocks, while noting uncertainty about what the findings establish. It also covers Energy Dome’s compressed-CO₂ energy storage, a planned €13 billion Google data center halted by Finland, an EU corporate levy for companies earning more than €100 million, and Absolics’ planned production of glass panels for AI chips this year. These are a collection of reported developments rather than a single market-moving event.

Analysis

The GLP-1 aging narrative is upside optionality, not yet a basis for extending LLY or NVO valuations: molecular clocks can move with weight loss and inflammation without demonstrating slower functional aging or fewer age-related events. The second-order risk is a higher evidentiary bar—if investors price a longevity franchise before validated outcomes, disappointing clinical data could unwind that premium while leaving the core obesity/diabetes thesis intact. Over 1–3 months, watch whether either company commits to controlled studies with clinically meaningful endpoints; over 6–18 months, only replicated outcome data would support a durable market expansion. Falsifier: results showing biomarker shifts do not persist or fail to correlate with health outcomes.

For GOOG, a local permitting setback is more informative as a constraint on the marginal economics and schedule of European data-center capacity than as evidence of a broad demand problem. Delays can raise the value of power access, grid connections, and approved sites, while redirecting capex geographically; they may also constrain AI capacity if substitutes are not ready. One project is insufficient to infer a material consolidated earnings hit. Track project approval, revised timing/capex, and evidence of similar opposition elsewhere. The Absolics/Intel glass-substrate item is early-cycle optionality: qualification and yield, not announced production plans, determine whether it becomes a meaningful bottleneck or efficiency gain. No near-term earnings inference is warranted.

Key Decisions for Investors

• Do not pay a separate longevity premium for LLY or NVO on biomarker claims alone. Treat this as an alert for trial-design and outcome evidence; reconsider only with controlled, replicated data linking clock changes to clinically relevant benefit.
• Keep GOOG exposure sized to the broader business thesis rather than this single-site headline. Escalate to an underweight review if the project is cancelled or delayed materially and comparable permitting resistance appears across multiple data-center markets.
• Watch Absolics and Intel for customer qualification, production yield, and shipment evidence before treating glass substrates as an investable AI-hardware catalyst; absent those milestones, no trade.