Daily signal: energy shock, enterprise AI acceleration, and rising bubble risk

Today’s Perplexity Discover feed shows a very clear market picture: macro risk from energy + micro euphoria around AI. On the macro side, oil and gas are being repriced by war headlines (Hormuz, cross-border strikes, bomber deployments), which then feeds into power prices and inflation expectations in Europe. On the micro side, Big Tech and startups keep shipping agentic products to capture enterprise budgets before financial conditions tighten further.

The first vector matters because this is not just an oil story; it is a broad cost shock. Discover items point to the palm-oil spike via biodiesel economics, the jump in Spanish wholesale power, and pressure across logistics, manufacturing, and households. Reuters, NYT, El País, and Bloomberg broadly align on the same conclusion: even with violent intraday reversals, markets are now embedding a meaningful geopolitical premium in energy and transport inputs.

The second vector is the battle for the enterprise “copilot of work.” Microsoft is pushing Copilot Cowork and new enterprise bundles; Anthropic is shipping multi-agent code review; and tools focused on giving agents fresh context (like Context Hub) highlight that the bottleneck is not just model quality but context quality + workflow integration. That is a business story, not a demo story: cycle-time compression and fewer production errors.

There is a direct connection between both vectors: if energy and inflation stay elevated, investors and CFOs will demand faster measurable ROI from AI projects. In a higher-rate or more fragile financing backdrop, winners are the vendors that prove operating leverage (hours saved, defects avoided, rework reduced), not the ones selling generic “AI transformation” narratives.

A third signal is also visible: the market is separating “consumer AI” from “production AI.” News around Apple delaying its smart-home display due to Siri execution issues shows how quickly consumer timelines slip when reliability is not there. By contrast, B2B spending remains resilient when products plug into critical workflows (engineering, finance, operations) and deliver measurable gains in weeks.

From an investing perspective, this argues for selective rotation rather than blanket “buy AI.” Joseph Stiglitz’s bubble warning is not an automatic crash call, but it is a useful filter: distinguish companies with distribution moats, proprietary data, and execution discipline from companies supported mainly by narrative and multiple expansion.

For Europe (and especially Spain), the tactical message is twofold. First, actively manage near-term energy exposure (input contracts, hedging, procurement strategy), because inflation transmission can move faster than growth. Second, use the AI cycle to unlock productivity in less glamorous sectors—banking, utilities, industrials, public services—where a 5–10% efficiency gain has immediate margin impact.

Base case for the coming weeks: high volatility in energy and geopolitics, with relief rallies on ceasefire headlines that can reverse quickly. Bull case: credible de-escalation, restored shipping flows, and continued AI spend backed by better adoption metrics. Bear case: prolonged conflict, sticky energy inflation, and multiple compression in software/semis as monetization skepticism rises.

The practical signal today is not “buy or sell AI as a block,” but prioritize exposure to verifiable productivity and strong balance sheets while treating energy risk as the dominant macro variable. In short, 2026 looks less like a linear innovation narrative and more like a race between operational execution and cost of capital.

Key sources reviewed (Discover-led, cross-checked): Reuters, BBC, Financial Times, AP, Bloomberg, NYT, El País, The Verge, Fortune, CNBC/PBS, plus the day’s Perplexity Discover reports.