TODAY'S HEADLINE: Six companies have promised to spend roughly $1.5 trillion on AI. Most of it does not appear where you would look for it.

Everyone tracks capital expenditure. It gets announced on earnings calls, analysts model it, headlines quote it. But a Financial Times analysis this week found something much larger sitting one layer down: purchase commitments approaching $1.5 trillion across Alphabet, Microsoft, Amazon, Nvidia, Oracle, and Meta. Here is what that number is, why it does not look like debt, and why it matters.

Leadership is asking: are we getting value from AI? Which tools are worth the spend? Where are we exposed? Right now, most teams have no idea.

Harmonic Security Usage Explorer changes that.

You get a complete picture of how your organization uses AI, automatically categorized into custom tasks and use cases.

You’ll see the projects being worked on, who’s using what tools, where AI investments are driving value, and where employees are engaging in risky behavior.

CIOs can rationalize spending and cut wasted licenses. CISOs can pinpoint where risk exists and neutralize it. AI committees can show exactly how their efforts are paying off.

See it in action

SECTION 01 // What actually happened

According to the Financial Times, the six largest AI infrastructure buyers have accumulated close to $1.5 trillion in purchase commitments, weighted heavily toward computing infrastructure, chips, data center capacity, and energy. Separately, Goldman Sachs identified roughly another $1.5 trillion in lease commitments across the same landscape.

Alphabet is the clearest illustration. Its purchase commitments rose sharply between the first and second quarters as the company locked in long-term technical infrastructure and energy agreements. These are contracts to buy things in future years, signed now, at agreed terms. They are not spending that has happened. They are spending that must happen.

The analysis: https://techstartups.com/2026/08/14/top-tech-news-today-august-14-2026-apple-anthropic-deepseek-google-ibm-pony-ai-openai-spacex-uber-more/

SECTION 02 // The puzzle

Purchase commitments do not appear on a conventional balance sheet the way borrowed money does. Debt shows up as a liability with a principal, an interest rate, and a maturity. A commitment to buy GPUs in 2028 is a contractual obligation, but it is disclosed in the footnotes rather than counted alongside the debt.

The practical effect is that a company can look moderately leveraged by the standard measures while carrying enormous fixed future obligations. Nothing improper is happening here, and the disclosures are public. The point is simply that if you evaluate these companies using capital expenditure alone, you are looking at the visible tip of a much larger structure.

Source: Financial Times analysis, reported August 14, 2026

Creators lock in holiday calendars 90 days out. Structure commissions, recruit creators, and optimize your creator affiliate strategy before the rush with Levanta's 90-Day Holiday Sprint. Get the Free Guide.

SECTION 03 // The real story

The reason these contracts exist is shortage. GPUs, electricity, construction crews, networking gear, and data center shells are all scarce, and the way you guarantee access to a scarce thing is to sign a long contract for it before someone else does. Every one of these companies has concluded that the risk of not having compute is worse than the risk of overcommitting.

That logic is sound right now. Its weakness is that it runs in one direction. Locking in supply years ahead protects you from shortages but removes your ability to pull back if AI revenue arrives more slowly than expected. The economics start to resemble heavy industry rather than software: enormous upfront commitments, long-lived contracts, and high fixed costs that revenue eventually has to cover.

Lease commitments: Goldman Sachs analysis, reported August 2026

SECTION 04 // Why it matters now

You can see the demand working through the supply chain this week. SMIC, China's largest foundry, is raising prices as its utilization rate hit 93.7 percent in the second quarter, with revenue passing $3 billion for the first time and quarterly profit more than tripling to $479.2 million. Applied Materials reported fiscal third-quarter revenue of $9.12 billion, up 25 percent, and said it aims to roughly double semiconductor system output by 2028.

Note what those two have in common: neither designs a GPU. AI demand is consuming memory, networking, controllers, power management chips, and mature manufacturing nodes, which is why the effects show up in foundries and equipment makers rather than only in Nvidia's numbers. Higher foundry pricing eventually reaches servers, industrial systems, and consumer electronics. Investors also pushed Applied Materials shares down despite the beat, which tells you expectations across this supply chain are already priced for years of extraordinary spending.

SMIC and Applied Materials: DIGITIMES and Wall Street Journal, August 14, 2026

FAQ // Quick answers

What is a purchase commitment?

It is a contract to buy a specified amount of something in the future at agreed terms. In this case that means chips, servers, data center capacity, construction, and electricity. The money has not been spent yet, but the company is obliged to spend it.

How is that different from debt?

Debt is money already borrowed, carried on the balance sheet with a principal and an interest rate. A purchase commitment is a future obligation disclosed in the footnotes. Both represent money the company must pay, but only one appears in the leverage figures most people look at.

Is $1.5 trillion unusual?

The scale is what makes it notable. It is roughly comparable to the lease commitments Goldman Sachs identified separately, so the two together are far larger than the annual capital expenditure numbers that dominate coverage of AI spending.

Why would companies commit this much before knowing returns?

Because compute is scarce and contracts are how you secure scarce supply. The alternative, waiting for certainty, risks having no capacity when demand arrives. Every major buyer has made the same bet, which is itself part of why supply is tight.

What should I watch next?

Whether these commitments keep growing quarter over quarter, and whether AI revenue growth keeps pace with the fixed obligations being signed. Footnote disclosures on purchase and lease commitments are the place to look. This is general information, not investment advice.

We will keep tracking this and bring you the next chapter as it lands. Stay sharp out there.

This newsletter is for general information only and is not investment advice. Always do your own research before making financial decisions.

TECH DAILY // www.techdailynews.org

Opening entities in every country can be slow, expensive, and hard to scale.

That's why more companies are using EOR to hire globally faster.

See how Oyster helps teams hire, pay, and support talent in 180+ countries while staying compliant along the way.

Get all the answers