AI hype is outpacing fundamentals, says Parnassus Investments’ Todd Ahlsten

The artificial intelligence boom has dominated markets, but investors need to separate hype from lasting value.

Todd Ahlsten, chief investment officer at Parnassus Investments, has been speaking with InvestmentNews about investing in AI and shared that his approach is to stay disciplined, diversify broadly, and focus on companies that can compound earnings through market cycles.

“The AI investment landscape is broadening rapidly,” says Ahlsten. “We need a massive amount of compute to support the infrastructure. There are also emerging infrastructure needs for inferencing as use cases proliferate for how the AI models can handle new data to make predictions or decisions.”

He sees opportunities expanding well beyond Nvidia and Microsoft.

Nvidia is one of the wealthiest companies in the world. Discover key insights here.

“Because this requires more specialized and efficient hardware, we see a broadening of compute requirements beyond the high-end GPUs Nvidia offers, and custom silicon chips and memory are big components of that opportunity,” he explains. “We favor companies like Broadcom (AVGO), Advanced Micro Devices (AMD) and Micron (MU).”

While infrastructure grabs headlines, Ahlsten believes investors are missing a key part of the value chain.

“Leading software companies have been overlooked as the market questions whether they can monetize AI,” he says. “They have great business models, but the growth rates have come down because their addressable markets have matured.”

He points to enterprise platforms that could reignite growth through AI integration.

“Software companies such as Salesforce (CRM), Microsoft (MSFT) and Workday (WDAY) are well-positioned because their corporate customers use them as central systems of record,” he notes. “They can layer AI agents on top to accomplish tasks more productively, and without the switching costs that could disrupt their customers’ daily business.”

These are high-quality businesses with consistent subscription revenues and durable earnings growth and Ahlsten says “We don’t view them as easily disrupted because they are so entrenched in their customers’ operations, where data and cybersecurity protections are critical.”

Long-term discipline amid AI enthusiasm

Ahlsten’s message to investors is to stay patient and this is informed by wide experience.

“I’ve managed our Core Equity strategy through seven presidential administrations and numerous market cycles,” he says. “These are unprecedented times, and I’ve learned that keeping a long-term perspective and avoiding knee-jerk reactions is key.”

Ahlsten looks for companies that can apply AI to strengthen existing advantages including those enhancing farming experiences, life sciences, and non-bank financials.

“We look for great American companies with durable cash flows and businesses that we think can compound value,” he says.

Despite AI’s potential, Ahlsten warns that “the narrative has reached a fevered pitch.” He says the fast build-out of AI data centers “could face shortages of power or funding or lack of monetization.”

“At the same time, unprecedented revenue ramps at companies like OpenAI will be needed to maintain the AI build-out relative to the high expectations,” he adds. “This has reduced the margin for error for investors.”

That’s why his team maintains “balanced positioning between offensive and defensive holdings.” He’s monitoring inflation, energy demand, tariffs and geopolitical risk — all of which could drive volatility or trigger a market pullback.

Lessons from past tech cycles

Ahlsten sees echoes of earlier market bubbles but notes key differences.

“The dot-com bubble was built by stocks on borrowed money, borrowed time, lack of cash flow and massive investor FOMO,” he says. “Today’s AI investment boom is being driven by a handful of companies with high margins and massive free cash flow from their existing operations.”

Still, he cautions that “the exponential ramp-up of data-center investments is starting to trim the free cash flows of the largest hyperscalers.” Some, he notes, “are starting to issue debt and use more innovative ways to finance these large projects.”

“The excitement feels similar to the dot-com boom, but the position these companies are in is remarkably different,” Ahlsten says. “It was the software applications that ran on the internet’s infrastructure two decades ago that made money in the long run. That’s another reason we favor software companies such as Microsoft, Salesforce and Workday.”

By contrast, “companies that are fragile in terms of having more competitive markets, more debt, more economic sensitivity and higher valuations don’t have the quality standards we seek,” he adds. “There is a lot of hype and animal spirits going into the AI wave — we aim to separate the signal from the noise.”

Ahlsten acknowledges the risks of market dominance by the “Magnificent Seven.”

“Their elevated valuations continue to go higher, dominating the returns for the market index,” he says. “Over-concentration raises vulnerability to a compute-capacity swing, valuation compression or a slower-than-hoped monetization curve. Balancing opportunities across the technology stack is critical.”

Diversification, he adds, is the best safeguard.

“We aim to provide a balanced exposure in our portfolios — capturing market upside and maintaining defensive positions that can help protect on the downside. Our valuation-aware hyperscaler and semiconductor exposure provides upside capture to the AI adoption wave. But we also lean into undervalued areas like life sciences, precision agriculture and business services.”