The Bill Is Coming

There are three forces converging on your business right now and one of them nobody is talking about.

There is a version of this week’s AI conversation that goes like this. Jensen Huang stands on stage at Nvidia’s annual conference and announces a trillion dollars in chip orders through 2027. Meta reportedly plans to cut 20% of its workforce to fund $135 billion in AI spending. Atlassian lays off 1,600 people to “self-fund further investment in AI and enterprise sales.” The market rewards all of it. Stocks go up. The future looks expensive but inevitable.

That is the version most people saw but it is far off from the full picture.

While the AI industry was celebrating the biggest infrastructure bet in corporate history, oil prices were sitting at roughly $100 a barrel. Up from $70 just three weeks ago. The Strait of Hormuz, a narrow waterway between Iran and Oman, has been effectively closed to shipping since late February. You have probably seen the oil headlines. They are dramatic and they matter. But they are not the part of this story that should concern you most.

The part that should concern you is the natural gas.

Twenty percent of the world’s oil passes through the Strait of Hormuz every day. But so does twenty percent of global liquefied natural gas. Qatar, which supplies a fifth of the world’s LNG, halted production and declared force majeure after Iranian drone strikes damaged facilities at Ras Laffan. And natural gas is what actually powers the AI industry.

I am not going to write about the war. But I am going to write about what happens when the single largest technology buildout ever attempted runs headfirst into the most significant energy disruption in years.

AI Runs On Power and Power Just Got A Lot More Expensive.

Data centres run on electricity and a large share of that electricity, particularly in the US and across Europe, comes from gas-fired power plants. When natural gas prices spike, electricity generation costs spike. And when electricity costs spike, the infrastructure behind every AI tool you use gets more expensive to operate.

According to the International Energy Agency, the combined electricity appetite of AI, computing, and cryptocurrency is on track to hit 1,050 terawatt-hours by 2026. That is roughly the entire electricity consumption of Japan. The fuel source underneath much of that demand is natural gas. And right now, a significant share of the world’s gas supply has been stuck on the wrong side of the Strait of Hormuz for nearly a month.

LNG prices across Europe have jumped by more than 60 percent. There is typically a four-to-eight-week delay before rising import prices feed through to domestic energy bills. That means the full impact on US and European data centre costs will start showing up in April and May. The AI industry already faced rising power costs before any of this happened. US retail electricity had already climbed 42 percent compared to 2019 levels. Goldman Sachs estimated that data centre electricity demand alone would add measurable inflationary pressure in both 2026 and 2027. That was before a war shut down a fifth of the world’s gas supply.

The Squeeze From Above

Let me pull the lens out because the energy story does not exist in isolation. It is arriving at the same time as two other forces that are pressing directly on small and mid-sized businesses.

The first is the layoff wave. By mid-March, more tens of thousands of tech roles had been cut across hundreds companies and every one of these companies framed the decision the same way. AI made it possible. Or necessary. Or both.

The second force is the infrastructure arms race. Meta has committed to spending $600 billion on AI data centre infrastructure by 2028. Amazon is projecting $200 billion in data centre spending in 2026 alone. Apple committed $500 billion in US investment over four years, spanning manufacturing, AI infrastructure, and R&D. Nvidia’s Jensen Huang told analysts this week that every company now needs an “OpenClaw strategy” the same way they once needed an internet strategy. He said tokens will be issued alongside laptops when you start a new job.

These are not small moves. They are reshaping how the largest companies on earth operate, how they spend, and what they expect from everyone else in their supply chain.

And here is where it lands on you.

The Squeeze From The Below

If you are running a small business or working inside a mid-sized company that has started relying on AI tools, you are now sitting at the intersection of three converging pressures.

Energy costs are rising, which means the infrastructure that powers your AI tools is getting more expensive to run. The companies providing those tools are spending at levels that require them to either raise prices, cut people, or both. And the labour market is flooding with experienced professionals who were displaced by the same technology you are trying to adopt.

The hyperscalers will absorb higher energy costs for a quarter, maybe two. After that, they adjust. API pricing shifts. Subscription tiers move. The tools you budgeted for in January may not cost the same by summer.

Meanwhile, if you have not started building AI into your operations, the gap between you and competitors who have is widening and partly because the companies laying off thousands of people are doing so on the assumption that smaller, AI-augmented teams can deliver the same output. If they are right, the standard for what a lean team can accomplish just changed permanently.

What I Keep Thinking About

I keep coming back to something Jensen Huang said this week. He described the shift from training AI models to running them in production as “the inflection point of inference.” The idea is that we have moved past the experimental phase. AI is no longer being built. It is being deployed. At scale. Across every industry.

That language matters because it says a lot about who bears the cost. During the building phase, the big labs absorbed the expense. During the deployment phase, everyone pays. The energy bill, the subscription cost, the competitive pressure. All of it distributes outward.

I think a lot of people assumed that AI would just keep getting cheaper. And in some ways it will. Models are becoming more efficient. Competition among providers is fierce. But cheaper models running on more expensive energy, inside companies that are simultaneously cutting headcount and raising infrastructure spend, does not guarantee that the end user saves money.

It probably means the opposite, at least for a while.

So What Do You Actually Do

I will not pretend there is a tidy answer here. But there are a few things I would be doing right now if I were running a business that depends on AI tools, which I am, so this is not hypothetical.

First, I would audit what I am actually paying for AI and what I am getting from it. Which tools are producing measurable returns and which ones are costing me money?

Second, I would look at where my AI stack depends on a single provider and whether I have options if pricing changes. Model-agnostic infrastructure is not just a technical preference. It is a financial hedge.

Third, I would pay close attention to what the big players do with pricing over the next 90 days. If energy costs hit data centre operators the way analysts are projecting, the downstream effects will start appearing in Q2. That is not far away.

And finally, I would stop treating the current moment as a reason to slow down. The squeeze does not reward hesitation. It rewards clarity. Knowing what AI does for your business, knowing what it costs, and being ready to move when conditions change.

All the Zest 🍋

Cien

Cien Solon is a founder and AI transformation strategist working at the intersection of people, platforms, and power. Through LaunchLemonade, she helps organisations design AI systems that are dependable, governable, and human-centred.

---

Sources and further reading

• How soaring oil prices caused by the Iran war impact costs — Al Jazeera

• Iran war threatens prolonged impact on energy markets — Al Jazeera

• Economic impact of the 2026 Iran war — Wikipedia

• Gasoline prices still rising as Iran war stretches into third week — NPR

• How will the Iran conflict impact oil prices? — Goldman Sachs

• Hormuz closure pushed LNG up 60% and is making AI compute more expensive — Abhishek Gautam

• Why the Iran war may have just killed the AI boom — OilPrice.com

• Iran conflict: oil price impacts and inflation — Morgan Stanley

• Electricity prices will keep rising on AI data centre demand — CNBC / Goldman Sachs

• US data centres’ energy use amid the AI boom — Pew Research Centre

• Nvidia GTC 2026: 5 biggest takeaways — TechRepublic

• Nvidia GTC 2026: Jensen Huang sees $1 trillion in orders — CNBC

• Nvidia GTC 2026: agentic AI takes centre stage — CNBC

• Nvidia GTC 2026: 8 takeaways from Huang’s analyst meetings — Constellation Research

• Meta stock climbs on report of planned layoffs to offset AI spending — CNBC

• Meta planning sweeping layoffs as AI costs mount — CNBC / Reuters

• Atlassian slashes 10% of workforce to self-fund investments in AI — CNBC

• An important update on our team — Atlassian

• Apple will spend more than $500 billion in the US over the next four years — Apple Newsroom

• Tech layoffs in first 74 days of 2026 — IBTimes