A New Era for AI: The Market No Longer Pays for Big Spending, but Demands Proof

For several years, there was one unwritten rule on Wall Street: the more a technology company invested in AI, the greater the market reward.
That narrative is beginning to change.
According to analyst Fu Peng, the logic of the AI market is now shifting from “rewarding cash-burning expansion” toward “punishing capital consumption”. In other words, investors are starting to stop asking “how much AI capacity is the company building?” and are beginning to ask “how much profit is generated from every dollar burned?”
AI is entering a different phase
This shift is emerging as AI infrastructure spending grows larger. JPMorgan estimates that AI capital expenditures rose to around 93% of hyperscalers’ operating cash flow in 2026, compared with just 33% in 2023.
Even in Q2 2026, total hyperscaler capex began to exceed operating cash flow, sending aggregate free cash flow into negative territory.
This does not mean these companies are unprofitable.
On the contrary, the problem is that the scale of investment is becoming too large to be evaluated based solely on revenue growth.
The market is beginning to favor “AI that makes money”
The shift in sentiment is clearly visible in investor responses to companies increasing their AI spending.
Alibaba, for example, recently announced raising around US$10 billion to fund AI investments after its quarterly profit fell 75% and AI capital expenditures surged. The market responded negatively, as investors questioned when those investments would generate returns.
On the other hand, companies with strong balance sheets still have room to invest aggressively.
This is the AI paradox of 2026:
Big spending is no longer automatically bullish. What is bullish is big spending that can be proven to generate returns.
This does not mean the AI bubble will burst
This is the important part.
AI demand remains very strong. Goldman Sachs even estimates that AI will remain one of the drivers of US business investment and could contribute around 0.5 percentage points to consumption through the wealth effect from stocks.

So, the change in the narrative does not mean the market is abandoning AI.
The market is instead becoming more mature in how it evaluates AI.
In the first phase, investors pursued companies with access to GPUs, data centers, and computing capacity.
In the next phase, the market will likely seek companies capable of proving:
AI Revenue → Margin → Free Cash Flow → Return on Invested Capital.
Investors need to change how they read financial statements
Therefore, revenue figures and user growth alone are increasingly insufficient.
Investors now need to pay attention to:
Capex vs operating cash flow
Free cash flow
Return on invested capital
Debt used to finance AI
AI monetization
Data center investment payback period
Companies capable of generating AI growth without compromising their balance sheets could earn a premium valuation.
Conversely, companies that continue increasing capex but are unable to demonstrate a path to monetization may face valuation pressure.
Conclusion
AI is not losing its narrative. Its narrative is moving up a level.
In the past, the market rewarded companies for being willing to spend money pursuing AI.
Now, the market is beginning to demand proof that the money is not merely expansion costs, but capital capable of generating profits.
The next AI battle is no longer about who has the largest data center, but who can turn trillions of dollars in investment into real cash flow.
And for investors, this subtle shift in how the market evaluates capex could become one of the most important themes in determining the winners and losers in AI’s next phase.
$BABA ‌$AMZN ‌$MSFT ‌