Daily Intelligence: Markets Buy Relief, but AI Still Demands Power and Patience

Today's thesis is uncomfortable but useful: markets have not solved their problems; they have simply been given a better window in which to breathe through them. Oil cooled after signs of a U.S.-Iran agreement, Wall Street rallied hard on Monday and artificial-intelligence stocks again helped carry the indexes. But the layer underneath remains physical: energy, chips, data centers, logistics, debt and trust.

I did not find the complete morning-briefing block in today's or yesterday's available notes - verse of the day, Glorify reflection, Stoic reflection and app of the day. I am keeping that as an internal note and using only a compatible thread: when noise falls, the right move is more attention, not more haste. Today markets offer relief; the practical question is whether that relief becomes structure or merely a rebound.

Macro / Energy

The data point that organizes the session is oil. AP reported that the S&P 500 rose 1.7% on Monday, the Dow gained 0.9% and the Nasdaq jumped 3.1%, while Brent crude fell 4.8% on hopes that a U.S.-Iran agreement can restore global crude flows more durably. The first read is clear: less energy stress means less pressure on inflation, margins and rate expectations.

But energy relief should not be confused with instant normalization. The Financial Times highlighted on its front page that flows through the Strait of Hormuz could take weeks to regain full confidence even if the deal holds. That point matters more than the daily move in crude: energy is not only bought in the spot market; it is insured, shipped and budgeted.

For companies, this means fuel costs can fall before uncertainty costs do. Airlines, logistics, industrials and consumer businesses breathe easier with cheaper crude. But if finance teams have learned anything over the last few years, it is that a supply chain does not become cheap again just because a geopolitical headline improves for 24 hours.

The connection with AI appears here. The new technology cycle needs a massive amount of stable electricity. MuyComputerPro, citing Gartner, notes that data-center electricity consumption is expected to grow 26% in 2026 and that AI-optimized servers are becoming a growing part of that pressure. Markets can celebrate oil, but the real energy bill of the next decade is moving toward power grids, cooling and capacity contracts.

Geopolitics

Geopolitics is again acting like a discount rate across almost every asset. If the Iran agreement is formalized and markets believe Hormuz is operational, the risk premium falls and long-duration assets rise: technology, semiconductors, software and small caps. If implementation stalls, the same movie rewinds quickly.

The difference versus earlier cycles is that geopolitics no longer lives only in oil. It also lives in chips, models, data, cloud, submarine cables, batteries and regulation. An export restriction, an energy subsidy or a data-sovereignty rule can move as much value as a rate decision.

Xataka has been capturing this shift well: talking about AI now means talking about national infrastructure. It is no longer enough to have strong models; countries and companies need data centers, energy, hardware, high-quality data and the capacity to operate at scale. Rivalry between blocs is becoming a race to control bottlenecks, not just a race to present flashier demos.

That is why today's geopolitical radar has two layers. The visible layer is Iran, oil and Hormuz. The less visible, and possibly more durable, layer is who can secure energy and compute for AI without depending on strategically vulnerable suppliers.

AI / Tech

AI stocks rebounded, but the conversation is maturing. Markets are no longer asking only whether AI matters; they are asking how much it costs, who captures the margin and when returns appear. Xataka framed it directly: we understand the cost of building AI far better than we understand how to measure what it produces.

That is a healthy point. In 2025 and 2026, capex numbers in the hundreds of billions became normalized as if they were just another budget line. But a technology can be revolutionary and still require financial discipline. In fact, the more revolutionary it is, the more important it becomes to separate narrative, real productivity and capturable profit.

The physical layer dominates the debate. Data centers, transformers, turbines, memory, cooling, fiber, land, permits and power agreements are now part of the product. The user sees an assistant responding in seconds; the investor has to see infrastructure that consumes capital before returning it.

The opportunity, therefore, is not only in the most visible AI brands. It also sits in companies selling efficiency, security, power, industrial automation, observability, storage, connectivity and software that cuts operating costs. The AI that creates the most value may not be the one that generates the most headlines, but the one that disappears inside processes that already produce revenue.

It is also worth watching narrative fatigue. When a sector rises on expectation and falls on monetization doubts, markets start distinguishing growth from faith. That filter does not kill the cycle; it makes it more selective. And a more selective cycle usually rewards income statements more than slide decks.

Markets

Monday sent a clear risk-appetite signal: major indexes up, Nasdaq leading, small caps participating and oil down. AP noted that the Russell 2000 also advanced and that the year remains positive for the main U.S. indexes. The combination fits a market that wants to believe three things at once: energy de-escalation, AI still alive and rates not tougher than expected.

The problem is that all three pieces need to stay aligned. If oil falls but yields do not, long-duration valuations still have a ceiling. If AI rebounds but earnings do not follow, the rally narrows to specific names. If Hormuz reopens but shipping takes weeks to rebuild confidence, the macro improvement is smaller than the first headline implies.

The portfolio read is more about selection than euphoria. Energy and margin defense still matter. Electrical and data infrastructure keep a tailwind. Semiconductors and software have potential, but proximity to the word AI is no longer enough. Markets are asking for evidence: backlog, contracts, cash flow, efficiency and pricing power.

There is also an interesting signal in Europe: telecoms, digital regulation and connectivity infrastructure are moving back to the center. Xataka Movil has been tracking MasOrange becoming fully owned by Orange, the European debate around Apple Intelligence and pressure on digital services. These are not isolated headlines; they describe a continent trying to organize its technological dependency while AI giants move faster than regulators.

24-72h Radar

First, follow execution of the U.S.-Iran agreement. Markets have already bought part of the relief; now they need confirmation in flows, maritime insurance, Brent, WTI and diplomatic language. If crude stabilizes without fresh shocks, the rally gets a stronger base.

Second, watch bonds. Energy relief can lower inflation expectations, but if the 10-year Treasury stays demanding, long-duration technology will still face natural resistance. AI does not trade in a vacuum; it trades against real rates.

Third, separate semiconductors from software. Chips can benefit from visible capex, but software has to show paid adoption and real savings. The next earnings season will be less forgiving toward vague metrics.

Fourth, monitor the power grid. Every story about data centers, capacity, cooling or power contracts matters more than it appears. Infrastructure has become the bottleneck that decides how much of AI's promise reaches production.

Fifth, watch Europe. Regulation, telecoms, digital identity, AI on phones and operator consolidation are part of the same board: technological sovereignty under budget constraints. It is less spectacular than an IPO, but it affects margins for many companies.

Scenario Conclusion

Base case: the Iran agreement reduces the energy premium, but logistics normalization takes weeks; AI rebounds, while markets demand more proof of monetization. Practical implication: keep exposure to quality technology and infrastructure, but avoid chasing rallies without evidence of cash flow and contracts.

Bull case: crude keeps falling, yields ease and market breadth improves beyond megacaps. Practical implication: gradually increase risk in selected semiconductors, electrification, energy-efficiency industrials, profitable software and fuel-sensitive consumer names.

Bear case: the agreement stalls, the Hormuz premium returns and bonds rise. Practical implication: reduce growth beta, raise liquidity and favor strong balance sheets, cybersecurity, defensive energy and companies able to pass through costs.

The close: markets have received oxygen, not absolution. The best practical application is to use the calm to organize exposure, review energy dependencies and demand something more from AI than promises: savings, revenue or measurable operating advantage.

Main Sources

AP: how major U.S. indexes closed on Monday, June 15. Financial Times: international front page on Hormuz and shipping confidence. Xataka: AI's visible cost and hard-to-measure value and Xataka Movil June archive. MuyComputerPro/Gartner: data-center electricity consumption in 2026.