Brent has moved back within touching distance of $100 after Houthi strikes on two Saudi tankers, and you should read that as a financing story as much as an oil story.
The oil market doesn't need a perfect blockade to scare investors. It needs enough missiles, enough rerouted ships, and enough uncertainty around the Bab el-Mandeb Strait to make every barrel look harder to move. That is what happened this week after Yemen's Houthis claimed attacks on the Saudi tankers Encelia and Layla in the Red Sea.
Financial Times reported that Brent rose above $99 a barrel after the attacks, its highest level since May. Prices are up roughly 30% this month. AP and Xinhua both reported the vessel names and the Houthi claim that missiles and drones were used. Xinhua noted a crucial caveat: the group did not provide evidence of the full damage it claimed, and Saudi authorities had not immediately commented.
Keep that caveat. It matters.
Shipping is already reacting as if the risk is real. The Wall Street Journal, citing Kpler ship-tracking data, reported that the Strait of Hormuz had only nine vessel crossings on July 21, while Bab el-Mandeb traffic fell to 29 crossings from 44 two days earlier. Before the Gaza war, that Red Sea route typically saw 65 to 72 daily crossings. You don't need every ship to stop before insurance costs, delivery schedules and oil prices start moving. The math moves first. Goldman Sachs put the risk plainly in a July 20 note reported by Bloomberg and MarketWatch: Brent could top $120 a barrel by the fourth quarter if disruptions through Hormuz persist. Goldman still sees $80 Brent in Q4 as its base case, assuming de-escalation. But the point is no longer academic. A $40 gap between the base case and the stress case is the market admitting it doesn't know which map it is trading.
Oil is now a rates story
The Federal Reserve was already stuck before the tankers were hit. Its April 29 statement kept the federal funds target range at 3.50% to 3.75%, and Reuters reported at the time that traders were betting there would be no rate cuts in 2026. By July, the debate had shifted even further. The Wall Street Journal reported this week that the Fed's 18 policymakers are evenly split on whether to raise rates this year.
That is a bad backdrop for founders who spent 2025 waiting for cheaper money. If you're raising a late-stage round, you don't care about Brent in the abstract. You care because oil feeds transport, food and manufacturing costs, and those costs feed the inflation numbers the Fed is paid to watch. More inflation pressure means fewer rate cuts. Maybe none.
Venture debt gets ugly in that world. Deals priced off SOFR can already land in the high single digits or low double digits once fees and warrants are counted. Add a delayed equity round, a lower revenue multiple, and a board that suddenly wants 24 months of runway, and the math changes fast. Frankly, founders who still model 2026 as a return to 2021 financing conditions are kidding themselves.
The split in venture is already obvious. AI infrastructure companies can still pull capital because investors believe compute demand is durable. Most other startups are fighting harder for smaller rounds, flatter terms and more proof. Oil at $99 doesn't create that divide, but it widens it.
AI has its own energy problem
The energy pressure does not stop at financing. It hits the cost base of the companies investors still want to fund. The International Energy Agency says data centres used about 415 TWh of electricity in 2024 and could reach around 945 TWh by 2030, with AI as the largest driver of the increase. In the United States, the IEA expects data centres to account for nearly half of electricity demand growth through the end of the decade.
Those are not soft numbers. They are the grid budget behind the AI boom.
The Energy Information Administration's latest monthly data showed average U.S. residential electricity prices at 17.30 cents per kilowatt-hour in 2025, up from 13.66 cents in 2021. Consumer Reports reported in March that 3,069 data centres already operate in the U.S., with another 1,489 planned or under construction, and pointed to Meta's 3,650-acre Hyperion campus in Richland Parish, Louisiana, as the kind of project now reshaping local power debates.
PolitiFact's March fact check was careful about the numbers. It found no support for one Florida candidate's claim that AI data centres drive bills up 30% to 40% as a general rule, but it did cite studies projecting an 8% national bill increase by 2030 and up to 25% in data centre-heavy places such as Virginia. One 2024 study it cited put Virginia's possible increase as high as 70%.
So the honest version is not that data centres automatically double your bill. They don't. The honest version is harder for the industry: new power plants, transmission lines and grid upgrades have to be paid for by someone, and households are already suspicious that they will be handed the bill.
The Houthi attacks are a Red Sea story on the surface. Underneath, the startup economy is still built on old-world inputs: fuel, shipping lanes, electricity prices and central bank patience. When those inputs move against you at the same time, the clever financing deck doesn't protect much.
Also read: Intel's stunning 2026 revival faces its biggest test as Q2 earnings land today • SpaceX's Starship Flight 13 is the first real test of a $2 trillion valuation • How the US-Iran War Turned Yemen's Houthis Into the Gulf's Most Dangerous Economic Wildcard
