James here, CEO of Mercury Technology Solutions.
Hong Kong — April 20, 2026
A client recently asked me a very sharp, macro-economic question about the AI boom: "Who is the ultimate buyer?"
They correctly pointed out that for any economic game to survive long-term, the loop must close. For example, the US maintains its massive trade deficit because exporting nations take those dollars and buy US Treasuries and equities, fueling a multi-decade bull market. The money cycles.
When you look at the AI ecosystem through this traditional lens, the math looks terrifyingly broken.
Nvidia sells chips to Big Tech platforms. Big Tech uses those chips to generate compute. They then sell that compute to enterprise software companies. But where is the end consumer? We haven't seen the "Model T" or the "iPhone" of AI yet. Everyday people are not buying raw compute just for the sake of it; they need a tangible product or service that fundamentally improves their lives.
Currently, the primary "application" enterprises are buying compute for is to replace their own human employees. This creates a macroeconomic paradox: If you fire the consumer to pay for your AI server costs, who is left to buy your goods?
The skeptic looks at this and concludes: There is no final buyer. This is a massive bubble, and everyone is about to suffocate on their own CAPEX.
It is a completely logical deduction. But it is built on the wrong premise.
To truly understand the AI economy of 2026, you must stop looking at it as a consumer goods business, and start looking at it as the Arms Trade.
1. The Economics of Fear (The "Lord of War" Protocol)
If you haven't seen the old Nicolas Cage movie Lord of War, you should. It perfectly illustrates the mechanics of the weapons business.
Who is the ultimate consumer of an artillery shell? Is there a normal citizen who buys a howitzer to display in their living room because it brings them joy? No. There is no "consumer happiness" application for a missile. Yet, the defense industry is one of the most ancient, resilient, and profitable sectors on Earth.
Why? Because the arms trade does not sell happiness. It sells the fear of losing your life.
You know you don't need an assault rifle for a happy life. But if your neighbor buys one, do you have to buy one? Yes. The market is instantly created not by utility, but by deterrence.
We forget that capitalism doesn't care whether a transaction is driven by the pursuit of joy or the avoidance of pain. In the United States, the legal industry generates massive GDP. People don't hire corporate lawyers because it makes their lives happier; they hire them because if the other guy has a ruthless lawyer and you don't, you will get crushed. It is defensive spending.
2. AI is Corporate Weaponry
Once you view AI as a form of digital weaponry, the current market behavior makes perfect sense.
Do you think the Big Tech platforms genuinely believe they will generate a direct, positive ROI on the hundreds of billions of dollars they are spending on Nvidia GPUs this year? The depreciation on silicon is brutal; in three years, those chips will be obsolete liabilities on their balance sheets. They will be lucky to make back 20% of their investment in direct compute sales. It is mathematically a massive loss.
But they are buying them anyway. Why? Because it is an arms race.
Think of the US and the USSR in the 1960s. Who actually "used" the thousands of nuclear warheads they built? Nobody. If they had, the planet would be glass. But could either side afford to stop building them? Absolutely not.
The platforms are buying chips because they are terrified. If they stop, their foundational models fall behind, developers migrate to a competitor's ecosystem, and their multi-trillion-dollar core business collapses. They are willing to bleed cash to maintain deterrence.
The enterprises buying AI software are in the exact same position. An enterprise doesn't buy AI agents because it makes the CEO happy. They buy them because they are terrified that their competitor will automate their workflow, slash operational costs by 40%, and initiate a price war that bankrupts them.
You don't buy a bomb to cuddle it. You buy it because your rival just bought one, and you can't sleep until you match their arsenal.
3. Be the Arms Dealer, Not the Philosopher
We are transitioning from a Human-centric internet to an AI-centric (B2A) internet. This is a structural rebuild of global commerce.
Yes, you can sit on the sidelines and philosophize about how this CAPEX is unsustainable in the long run. You can worry about whether the ultimate consumer will ever materialize.
But think about how the arms dealers view the world. They do not care if the war makes geometric sense. They provide the infrastructure, they collect the tolls, and they accumulate capital.
We are living through one of the most violent wealth redistributions in human history. If your business model relies on being the final bag-holder, you should be terrified. But if your business model is providing the picks, shovels, APIs, and data architecture to the corporations fighting this war? You are printing money.
If you sit this out because you are worried about the macroeconomic endgame, fast forward to 2040. Your children will ask you: "Dad, in 2026, when the entire world was re-architecting itself, why didn't we capture any of that wealth?" Do you want to be the person who missed the greatest technological shift of the century because you were too busy worrying about saving the world?
You aren't Superman. You can't save the global economy. But you can build the infrastructure, extract the value, and save your own family.
Mercury Technology Solutions: Accelerate Digitality.
