factolio.com

news & analysis

AI Optimism, War Risk, and the Price of Energy

Listen to this episode

Listen to this episode on RedCircle

Listen to Factolio on:

Spotify  |  Apple Podcasts  |  Amazon Music / Audible  |  iHeartRadio  |  YouTube  |  RedCircle

Global markets are rallying on strong artificial-intelligence expectations even as the U.S.-Iran war disrupts Middle Eastern energy flows. The Nasdaq reached a record while Brent crude fell below $100, but investors may be pricing in diplomatic progress that has not yet become a durable settlement.


Factolio looks at major current events from several AI-generated perspectives. Red Velhouse is the moderator. Sam Dewinski brings historical context, Kate Burvish examines the economic forces and consequences, and Ann Tofado looks at the political dynamics and implications.

Discussion

Kate Burvish:

It is seeing two different time horizons. Investors still expect unusually strong earnings from the companies associated with artificial intelligence. Analysts cited by the Associated Press expected companies in the Standard & Poor’s 500 index, commonly called the S&P 500, to report nearly 29 percent year-over-year earnings growth in the third quarter. At the same time, the latest energy news looked less alarming: Brent fell below $100, Saudi Arabia reportedly restarted its East-West pipeline, and there were reports of U.S.-Iran contact at the United Nations. That combination allows investors to price in easing war risk while continuing to buy the technology story. The vulnerability is that the easing is still an expectation, not a settled fact.

Omar Seidren:

And the technology story is not pure fantasy. There is a real investment cycle underway. The International Energy Agency says five large technology companies spent more than $400 billion in capital expenditure in 2025, with another 75 percent increase expected in 2026. That spending is broader than artificial intelligence alone, but AI infrastructure is a major driver. Chipmakers, cloud providers, networking companies, and data-center builders are selling into genuine demand. The market is not just applauding a clever chatbot demo; it is financing an industrial buildout.

Eric Arcan:

But the market is treating a lower price as if it meant a restored system. Brent at $99 is better than Brent near $110, but it is still about $27 above its prewar level. U.S. gasoline was nearly $4.48 a gallon on September 21, and the Energy Information Administration estimated average oil-production shut-ins of 6.7 million barrels a day in August. Those figures say the immediate panic has moderated. They do not say the physical shock has disappeared.

Red Velhouse:

So the rally may be rational without being secure. Kate, what tells us whether this is a repricing of improving conditions or complacency about risks that have not gone away?

Kate Burvish:

Some of each. Markets are forward-looking, so it is rational to assign a probability to improved tanker movement, pipeline use, and diplomacy before every barrel is flowing normally. But the rally is concentrated. Technology has held up better than energy shares, while higher yields and geopolitical uncertainty have pressured banks and other interest-sensitive sectors. Investors are not declaring the whole economy healthy. They are selecting the part they believe can preserve margins and growth despite higher costs.

Eric Arcan:

And the pipeline needs to be described accurately. Saudi Arabia’s East-West pipeline can move roughly four million barrels a day toward Yanbu on the Red Sea, according to sources briefed on the restart. That is valuable redundancy and can reduce the exposure of some exports to the Strait of Hormuz. But it cannot replace every normal route, every grade of crude, or every shipping requirement. A bypass is a pressure-release valve, not a reconstruction of the global oil system.

Omar Seidren:

That same distinction applies to AI infrastructure. Adding a second route does not mean the whole system has unlimited capacity. We can have enough chips on paper and still run into constraints in networking, cooling, grid connections, or reliable deployment. The impressive part is that the system is scaling rapidly. The less glamorous part is that every layer underneath the model has to scale with it. The market puts a spotlight on the model and a small footnote on the substation.

Red Velhouse:

Omar, let’s stay with that infrastructure point. How do we distinguish genuine AI monetization from companies spending because they are afraid to fall behind?

Omar Seidren:

Look for revenue quality rather than impressive spending totals. Genuine monetization means customers are using systems repeatedly, paying for them, and integrating them into workflows where the productivity gain is measurable. We should also ask whether capital spending produces useful capacity or simply reserves scarce chips and power. The optimistic case is real: better models can automate parts of research, software, design, and operations. But forecasts are not results. A company can build an expensive data center before it knows whether customers will pay enough to justify it.

Kate Burvish:

And the distribution matters. The early benefits flow toward semiconductor firms, hyperscalers, utilities, construction companies, and regions able to host large facilities. The costs are broader: electricity demand, transmission investment, land, and potentially higher financing costs. The Federal Reserve raised its target range to 3.75 to 4 percent on September 16 while saying inflation remained elevated and geopolitical uncertainty was significant. That makes the buildout more expensive and reduces the present value of profits expected far in the future.

Eric Arcan:

Which is why I resist treating AI power demand as an abstract cloud-computing issue. Data centers need firm electricity, transmission, cooling, and a connection schedule. The International Energy Agency projects data-center electricity use to rise from about 485 terawatt-hours in 2025 to roughly 950 terawatt-hours in 2030. In the United States, data centers are expected to account for about half of the increase in electricity demand through 2030. Those are engineering and permitting problems, not just valuation problems.

Red Velhouse:

Eric, what can actually supply that electricity, and where do the next constraints appear?

Eric Arcan:

There will not be one answer. Existing gas generation, renewables, storage, nuclear power, efficiency, and transmission will all contribute, depending on location and timing. The practical question is what can be built reliably before the load arrives. If data centers arrive faster than generation and wires, prices rise or reliability deteriorates. If a region relies heavily on gas, emissions and fuel infrastructure become part of the debate. If it relies heavily on new renewables, storage and transmission must arrive too. The workable strategy is a portfolio, not a slogan.

Omar Seidren:

There is also a subtle efficiency issue. AI systems are becoming more efficient per task, which is excellent. But if efficiency makes a service cheaper and more useful, people use much more of it. Video generation, reasoning-heavy systems, and autonomous or agentic applications can be computationally intense. Energy per query can fall while total electricity consumption rises. That is not a contradiction; adoption can outrun efficiency gains.

Kate Burvish:

And higher energy prices can interrupt that adoption story through household demand. A family paying more for gasoline has less money for discretionary purchases, while transport-intensive businesses face higher costs. If oil remains elevated, inflation can stay sticky enough to delay interest-rate relief. The economy then gets an uneven combination: productive investment in AI on one side, weaker purchasing power and more expensive credit on the other.

Red Velhouse:

Let’s connect that economic pressure to the diplomatic signal. Eric, how much of the oil decline reflects actual supply restoration, and how much reflects traders pricing in an outcome that has not been secured?

Eric Arcan:

Both, but the diplomatic component may be doing more work than the headlines admit. Increased tanker movement, the Saudi pipeline restart, and reported U.S.-Iran contact reduce the probability of the worst immediate scenario. Yet the contact was not a verified ceasefire or agreement. Its substance, authority, and durability remain unclear. The Energy Information Administration’s outlook assumes export constraints persist through the end of 2026 and that most flows return toward pre-conflict averages only in the second quarter of 2027. That is a forecast built on workarounds, not proof that normality has returned.

Kate Burvish:

That is also where financial markets can look detached from households. Investors can buy the probability of de-escalation in seconds. Households experience fuel prices, food and transport costs, and interest rates over months. Even if oil falls another few dollars, inventories may remain under pressure. If attacks resume, prices can rise quickly because spare route capacity and shipping alternatives are limited.

Omar Seidren:

Technology revenues are globally diversified, which helps explain why investors see some insulation from a regional conflict. But the infrastructure is not geographically abstract. Chips, electricity, cables, cooling equipment, and construction all depend on physical supply chains. AI can be globally sold and locally constrained. Software margins may be global, but electrons still have addresses.

Red Velhouse:

If you each had to name the most revealing evidence over the next two weeks, what would you watch?

Kate Burvish:

I would watch bond yields, corporate earnings guidance, and consumer-facing energy prices. If earnings remain strong while yields stabilize, the market can defend high technology valuations. If oil rises again and yields move higher, the combination becomes more damaging: weaker demand, more expensive financing, and less patience for distant promises. I would also watch whether the rally broadens beyond a narrow group of technology companies. Breadth would suggest confidence in the economy; concentration suggests a bet.

Eric Arcan:

I would watch physical evidence: sustained tanker flows through the Strait of Hormuz, actual restoration of production, inventory changes, and whether the Saudi pipeline operates consistently at meaningful volume. A single diplomatic meeting or one lower settlement price is not enough. The system needs repeated safe voyages and rebuilding inventories. Another infrastructure attack could reverse the optimistic energy narrative almost immediately.

Omar Seidren:

For AI, I would watch customer behavior and bottlenecks. Are companies converting pilots into recurring contracts? Are data centers getting connected on schedule? Are power and networking constraints delaying deployments? And do earnings show returns on all that capital, rather than just more capital being announced? I remain optimistic because the capability curve is genuinely moving. But the next phase will be decided less by dazzling demonstrations than by boring reliability, useful integration, and enough electricity to run the thing.

Red Velhouse:

The unresolved issue is whether markets are correctly anticipating two favorable developments at once: durable improvement in Middle Eastern energy flows and profitable, scalable AI demand. So far, neither proposition is fully established. Watch for sustained tanker traffic, inventories and pipeline operations; for an actual diplomatic arrangement rather than a reported contact; and for earnings, bond yields, customer adoption, and data-center power connections. The rally may prove farsighted—or simply early. Sources and references for this discussion are available with the episode at Factolio.com.


Sources and References

These sources supported the factual material used in this discussion. Factolio’s panel discussion is AI-generated from researched evidence and is written in original language.

  1. Reuters, republished by Kitco — Nasdaq sets record high on AI optimism, oil drops below $100 (NEWS)
  2. Associated Press — Wall Street holds near its record after Brent oil falls below $100 per barrel (NEWS)
  3. Associated Press — Global shares are mixed and oil prices stabilize after Trump says US and Iran met at the UN (NEWS)
  4. Reuters, republished by MarketScreener — Saudi Arabia restarts East-West oil pipeline, sources say (NEWS)
  5. Reuters, republished by MarketScreener — US open to meeting Iran at UN, no meeting scheduled yet, Rubio says (NEWS)
  6. Axios — Arab mediators push for U.S.-Iran meeting at U.N. to de-escalate crisis (NEWS)
  7. U.S. Energy Information Administration — Short-Term Energy Outlook, September 2026 (PRIMARY)
  8. U.S. Energy Information Administration — Short-Term Energy Outlook, September 2026: Global Oil Markets and Energy Security (PRIMARY)
  9. Federal Reserve Board — Federal Reserve issues FOMC statement, September 16, 2026 (PRIMARY)
  10. International Energy Agency — Data centre electricity use surged in 2025, even with tightening bottlenecks (PRIMARY)
  11. International Energy Agency — Key Questions on Energy and AI (PRIMARY)
  12. International Energy Agency — Electricity 2026: Demand (PRIMARY)
  13. International Energy Agency — Electricity Mid-Year Update 2026 (PRIMARY)
  14. U.S. Department of Energy — DOE’s Office of Electricity Publishes 2026 Draft National Transmission Needs Study (PRIMARY)