AI stocks rewarded believers in 2026, but they didn’t offer a smooth ride.
Now, Franklin Templeton CEO Jenny Johnson is challenging a crucial assumption behind that rally. Johnson, whose firm manages nearly $1.8 trillion, told CNBC that AI has not meaningfully powered today’s productivity gains.
Instead, she credits earlier digital technologies, including cloud computing. The payoff may be real, but adoption could take considerably longer than markets expect.
That shocking take comes at a point when investors and the biggest companies in tech are pouring money into the increasingly capital-intensive AI buildout.
For perspective, Amazon, Microsoft, Alphabet, and Meta Platforms are expected to spend roughly $630 billion on AI infrastructure in 2026 alone, as reported by Reuters.
Johnson remains constructive on earnings and the economy, but her unusually concise AI verdict raises a provocative question: have stock market valuations moved much faster than AI’s measurable impact?
Franklin Templeton CEO says AI payoff is still ahead
“AI is not yet in the system.”
That seven-word verdict from Franklin Templeton CEO Jenny Johnson challenges the market’s biggest assumptions which is that the current AI spending is already transforming economic productivity.
U.S. productivity is running at roughly 2.5%, but Johnson told CNBC, “I don’t think any of that is AI.” She instead attributes much of the improvement to earlier digital advances and cloud computing.
For context, the latest BLS data, shows that nonfarm-business productivity rose 2.2% year over year in Q2, down from 2.9% in Q1 and 2.5% in both Q3 and Q4 of 2025. Quarterly annualized growth also remained modest at 0.8% in Q1 and 1.4% in Q2.
More Wall Street:
- Wall Street’s AI trade faces its biggest valuation test
- The next Wall Street shift is already underway
- Wall Street sends strong 4-word verdict on the stock market
Her point is about timing, not AI’s ultimate potential. Businesses first use new technology to improve existing processes. Only after they understand it do the “leapfrog” applications emerge.
Johnson compared the moment with the iPhone, whose app ecosystem was difficult to imagine at launch. She also noted that electricity took 30 years to penetrate manufacturing. “You can’t possibly imagine it until you start to play with it,” she said.
That lag matters for investors. Johnson worries technology companies have shifted from capital-light models toward capital-intensive ones, while not all spending has reached their bottom-line.
Nevertheless, she feels strong consumer spending, record earnings and tight credit spreads have kept her broadly constructive.
Strong economy is doing AI’s work for now
Johnson’s broader point is that markets do not need an AI productivity miracle to keep climbing, at least not yet.
“As long as the consumer, the economy stays strong [and] corporate earnings stay strong, the market’s going to continue to go up,” she said. Her private-credit team reports that 70% of its middle-market borrowers are beating internal models.
Outside data support that resilience.
FactSet expects S&P 500 earnings to rise 28.9% in Q3, potentially marking an eighth consecutive quarter of double-digit growth. Single-A corporate spreads stood at just 65 basis points on Sept. 22, supporting Johnson’s instruction to “follow the money, follow the bond market.”
But robust earnings should not be confused with proof that AI is driving the economy.
FactSet found AI mentioned during more than 65% of S&P 500 second-quarter earnings calls. Yet a Federal Reserve review showed only 18% of U.S. businesses had adopted AI by year-end 2025.
PwC’s 2026 CEO survey was more sobering.
56% reported no significant financial benefit from AI, while only 12% achieved both revenue and cost gains. Strong earnings support today’s market. AI’s promised productivity windfall must still earn its valuation.
How investors should play AI’s reality gap
Johnson’s argument does not justify abandoning AI stocks but it argues for becoming more selective about what investors are paying for.
The clearest opportunities remain businesses converting AI demand into revenue and cash flow, especially semiconductor, cloud and networking leaders.
However, the rising capital intensity changes the equation. More spending on chips, data centers and power can lift revenue today while depreciation, financing costs and weaker free-cash-flow conversion emerge later.
Investors should therefore track AI revenue separately from cloud growth, while comparing capex growth with operating income and free cash flow. If spending keeps accelerating without corresponding margin expansion, Johnson’s concern becomes harder to dismiss.
The broader economy provides a temporary cushion.
Strong corporate earnings and tight credit spreads suggest that recession and default risks remain contained. A widening in spreads, weakening earnings revisions or deteriorating consumer spending would change that calculus quickly.
The practical move is not to sell AI indiscriminately.
It is to favor profitable, cash-rich leaders with visible monetization, while demanding a larger margin of safety from companies whose valuations depend on productivity gains that have not yet appeared.
