Toward the end of 2025, the question “Is AI a bubble?” has started to come up frequently. Azeem Azhar and Nathan Warren seek an answer to this question in a comprehensive analysis published on Exponential View. But this time not with superficial observations; with historical data, macroeconomic ratios and a five-dimensional 'bubble indicator.'
A Five-Indicator Framework
1. Economic Pressure – Are AI investments bending and distorting the economy?
2. Sectoral Pressure – Are investments balanced with sector revenues?
3. Revenue Growth – Does real revenue growth support the excitement?
4. Valuation Heat – Are share prices excessive relative to history?
5. Financing Quality – Do investments come from solid balance sheets, or are they financed with debt?
The Facts Tell a Different Story Than a Bubble
- The investment volume is enormous: $3 trillion in AI infrastructure investment is expected by 2029.
- Its economic weight is limited: in the US, these investments are less than 1% of GDP – the historical bubble threshold was generally above 2%.
- Revenues are rising fast: GenAI revenues are doubling every year; Citi forecasts 2025 growth at 483%.
- Valuations are moderate: the Nasdaq P/E ratio is ~32, half that of the dot-com era.
- Financing is strong: Microsoft, Amazon, Google, Meta and Nvidia invest with their own cash flows; however, debt financing is growing.
Where Should We Be Careful?
The article says caution is needed if two 'red signals' are seen together:
- If revenue growth slows, or
- If the share of debt-financed investment increases,
it warns that the AI sector could slide from 'boom' to 'bubble.' Short-lived GPUs (about 3 years) make financial discipline essential.
The Engine Is Running but There's No Overheating
What is happening today is not a 'bubble' but, on the contrary, the construction of a new economic infrastructure. AI data centers, just like the railways or electricity grids of the 20th century, are becoming the backbone that will carry the productivity of the future. But like every boom, this cycle is not limitless either. If investments outpace the rate of revenue creation, it's only a matter of time before history repeats itself.
Takeaways for Executives and Investors:
1. Think of AI as an 'infrastructure investment,' not a passing trend.
2. Watch the macro thresholds: 2% of GDP is the risky zone.
3. Don't forget where the profit forms: value accumulates in the ecosystems that use the models, not in the hardware.
4. Look at cash flow rather than liquidity.
5. The 'two red signals' rule: if two of the indicators sound the alarm, it's time to reduce the position.
In short: AI is not a bubble today, but the heat is rising. And if history has taught us anything, it's this: 'Bubbles are the story of excessive optimism.' This time, it's in our hands to write the end of the story with a little more caution.
