Real Technology Can Still Be a Bubble
In late June 2026, Apple raised prices on several products. Macs. iPads. HomePod. Apple TV. Vision Pro.
Not because Indians suddenly wanted more gadgets. Not because aluminium became rare. Not because some new luxury tax appeared.
Apple blamed memory and storage costs, driven by the AI data centre buildout.
That is where the AI boom stopped being a Silicon Valley story and quietly walked into the buyer’s wallet.
Apple reportedly said, “We have never seen a component price increase this much, this quickly.” Tim Cook also reportedly described the memory situation as a “hundred-year flood.” The company’s stock fell that day. Other device makers, including Microsoft with Xbox, reportedly moved in the same direction.
Treat all numbers in this post as approximate and verify them from primary sources before using them seriously. The point here is not whether one Apple model rose by $100 or $300. The point is that AI infrastructure spending is no longer abstract. It is entering the price of devices ordinary people buy.
That includes Indian buyers.
A student buying a laptop in Pune. A freelancer replacing a MacBook in Bengaluru. A parent buying an iPad for a child in Mumbai. A gamer looking at console prices. A small business upgrading machines for employees.
You may not own AI stocks. You may not work in Silicon Valley. But when capital floods one sector hard enough, the shockwaves reach your invoice.
Here’s the thing. The question is not simply, “Is AI a bubble?”
That is too lazy.
The sharper question is this: can the technology be real and the financial boom still be dangerous?
Yes.
That is exactly how many bubbles work.
Bubble Does Not Mean Fake
Most people hear the word “bubble” and immediately assume fraud, nonsense, or empty hype.
That is the first mistake.
A bubble does not mean the underlying technology is fake. It means the price, financing, expectations, and timelines may have gone insane.
The internet was real in 2000. It still crashed.
Railways were real in 1840s Britain. Investors still got hurt.
Fibre optic cables were real. Many of the companies laying them still went bankrupt.
AI may absolutely be real. It may reshape software, search, coding, education, customer support, medicine, defence, finance and the way companies operate.
But that does not automatically mean every rupee of AI infrastructure spending earns an adequate return. It does not mean every AI company deserves its valuation. It does not mean the investor buying late into the story gets rewarded.
A technology can change the world and still destroy capital on the way there.
That is the uncomfortable lesson.
The Spending Is Enormous. The Revenue Is Still Catching Up.
The largest technology companies are reportedly preparing to spend somewhere around $660 billion to $690 billion on capital expenditure in 2026, with some estimates going higher. Company guidance and analyst estimates vary, so do not treat this as a settled number. But the direction is clear.
The money being spent is huge.
A large portion of that spending is believed to be AI specific, covering chips, data centres, cooling systems, power, networking, memory, storage, land and long term supply contracts.
This is no longer just software people typing prompts into a chat window. This is steel, electricity, semiconductors and financing.
The problem is that current AI revenue may still be far smaller than the infrastructure being built to support it. Sequoia’s David Cahn famously framed this as AI’s “$600B Question,” asking whether the revenue expectations implied by the infrastructure buildout can realistically be met.
That question has not disappeared.
An MIT study reported in 2025 found that around 95 percent of enterprise generative AI pilots had not produced measurable profit. Separately, consulting research has suggested that only a small minority of companies are seeing substantial value from AI at scale. Exact figures vary by source and should be checked, but the direction is hard to ignore.
Many companies are experimenting. Fewer are earning serious money from it.
That gap matters.
Because capital markets do not price dreams forever. At some point, spending must meet cash flow.
The Capital Cycle Does Not Care About Hype
Manias usually follow a pattern.
First, returns look attractive. A few early players make money. The story spreads.
Then capital floods in. Everyone wants exposure. Nobody wants to miss the future. Banks lend. Investors cheer. Analysts upgrade. Founders raise. Suppliers expand.
Then too much capacity gets built.
Returns fall. Pricing power weakens. Debt becomes heavy. Weak players break.
Then, after the wreckage, survivors buy assets cheaply. Demand eventually catches up. The infrastructure becomes useful. Fortunes are made, but usually not by the people who funded the most expensive part of the buildout.
This is the capital cycle.
It is not complicated. It is just hard to see when everyone around you is drunk on the same story.
The fibre optic boom around 2000 is the cleanest example.
During the dot-com era, telecom companies laid enormous amounts of fibre because internet traffic was expected to explode. The idea was not wrong. The internet did explode. Streaming happened. Cloud computing happened. Remote work happened. Data demand became massive.
But the timing was wrong.
Too much fibre was built too soon. A lot of it sat unused for years as “dark fibre.” Many telecom companies collapsed. Investors lost money. Yet that same infrastructure later became part of the backbone of the modern internet.
The technology won. Most of the companies that built it did not.
That is the line every AI investor, tech employee and ordinary saver should remember.
The future can arrive exactly as promised and still bankrupt the people who overpaid to build it.
Why India Should Not Watch This Like a Foreign Drama
Indian readers often treat these stories like rich world problems.
Apple, Microsoft, Nvidia, OpenAI, Google, Amazon, Silicon Valley and Wall Street may sound far away.
Nice to read. Not directly relevant.
That is false comfort.
India sits downstream of this boom in multiple ways.
First, device prices. If AI data centres absorb memory and storage supply, consumer electronics become more expensive. That reaches India through laptops, phones, tablets, consoles, external drives, servers and office hardware. A global memory crunch does not stop at customs.
Second, IT services. India’s technology and services sector depends on global enterprise spending. If companies keep spending heavily on AI, Indian IT firms may benefit through migration work, integration, automation projects, data engineering and support. But if the spending cycle reverses sharply, budgets can freeze just as quickly.
Third, investor sentiment. Indian investors increasingly own global tech directly or indirectly. Mutual funds, ETFs, international funds, employee stock plans, private portfolios and even domestic market sentiment are tied to global technology expectations. A reset in AI valuations would not politely remain in Nasdaq.
Distance is not safety.
A boom financed in dollars can still raise your laptop bill in rupees.
A correction in America can still hit IT hiring in India.
A chip shortage in Taiwan or Korea can still change the cost of your next device in Delhi.
This is how interconnected money really is.
The Honest Counterargument
Now let’s be fair.
This is not the 2000 dot-com bubble copied and pasted into 2026.
The biggest AI spenders today are not random companies with no revenue and a “.com” in their name. Nvidia, Microsoft, Google, Amazon and other major players are enormously profitable. They have real customers, real cash flows, real balance sheets and real pricing power.
That matters.
Many dot-com companies were burning money they did not have. Today’s giants are spending aggressively, but much of it comes from serious operating cash flow.
Valuations are also not uniformly at the absurd extremes of 1999, though some areas may still be stretched. This makes a total wipeout less likely than a pure hype bust.
So anyone claiming “AI is obviously fake” is not thinking clearly.
The productivity potential is probably real. The adoption curve may still surprise skeptics.
But that does not settle the investment question.
A strong company can overbuild. A real technology can be overfinanced. A profitable sector can still suffer a painful correction if expectations outrun revenue.
The Three Questions Investors Must Separate
This is where most people lose money.
They hear a powerful story and stop thinking.
AI will change everything.
Maybe.
But that statement alone tells you nothing about whether the current spending makes sense.
You need to separate the three questions.
Is the technology real?
Probably yes.
Is the revenue real?
Partly, but still uneven. Some companies are making serious money. Many enterprise users are still experimenting. A lot of pilots have not yet turned into measurable profit.
Is the price real?
That is the hardest question. And anyone pretending to know with certainty is selling confidence, not truth.
This is not financial advice. It is a thinking framework.
Before getting excited by an AI stock, AI mutual fund, AI startup, AI data centre story or AI linked supplier, ask what part of the chain you are actually buying.
Are you buying the user of AI? The seller of chips? The builder of data centres? The owner of power capacity? The software layer? The consulting layer? The hype layer?
Each has a different risk.
Most people just buy the narrative.
That is not ownership. That is crowd participation.
The Fiat to Free View
At Fiat to Free, the question is never just “what is hot?”
The question is: who owns the asset, who funds the buildout, who carries the risk, and who gets paid when the story becomes real?
Ordinary people are trained to chase headlines.
Owners study incentives.
The AI infrastructure boom may build the backbone of the next economy. It may also leave behind excess capacity, broken balance sheets and investors who confused technological inevitability with financial certainty.
That is what history keeps teaching.
So the next time you read a breathless headline about trillion dollar AI spending, do not ask only whether AI is real.
Ask something sharper.
Who is paying for the future, and when will the future pay them back?
Fiat to Free 
Leave a comment