5 VCs sounds off on the AI question du jour

At Fortune’s IRL Term Sheet Breakfast at Brainstorm AI, five VCs with funds ranging from $5 million to $25 billion debated whether AI is in a bubble, with views spanning cautious to highly bullish: Jenny Xiao said any bubble is “relatively contained” to the infrastructure layer (data centers, GPUs and large language models) while the application layer remains underinvested; Vanessa Larco argued consumer AI may prove more durable than enterprise due to adoption and stickiness; Rob Biederman warned most startups will fail and only a few will become the big winners; Aaron Jacobson predicted short-term overhype and corrections but significant long-term value creation; and Daniel Dart envisaged enormous, still-unseen TAM and potential trillion-dollar companies. The panel expects to deploy tens to hundreds of millions into AI over the next decade, implying near-term valuation and funding cycles—especially in infrastructure—alongside sustained opportunities in application and consumer plays for firms that can systematically create customer value.

Analysis

Fortune’s IRL Term Sheet Breakfast at Brainstorm AI convened five VCs managing funds ranging from $5 million to $25 billion who said they will collectively deploy “tens to hundreds of millions” into AI over the next decade, highlighting active capital flows and divergent views on bubble risk. Jenny Xiao characterized any bubble as “relatively contained” to the infrastructure layer—specifically data centers, GPUs and large language model companies—while arguing the application layer remains underinvested. Vanessa Larco argued consumer AI may prove more durable due to product-led adoption and brand stickiness, while Rob Biederman warned that roughly 99% of companies in a boom can fail without systematic customer value creation. Aaron Jacobson expects near-term overhype and correction cycles but significant long-term winners in 10 years, and Daniel Dart projected vast, still-unrealized TAM including potential trillion-dollar firms by 2030–2034. Ongoing deal activity—examples include Saviynt’s $700 million Series B led by KKR and fal’s $140 million Series D led by Sequoia—supports continued funding momentum, but the signal set is mixed and cautious (sentiment_score 0.05, market_impact_score 0.25), implying concentrated valuation risk in infrastructure and selective opportunity in applications and consumer AI.

Key Decisions for Investors

• Reduce or hedge exposure to pure infrastructure plays tied to data centers, GPUs and LLM training where panelists flagged overinvestment
• Shift or add selective exposure to differentiated application-layer and consumer AI businesses that demonstrate clear, measurable customer value and product stickiness
• Use fundraising cadence and large rounds (for example the $700m Saviynt and $140m fal rounds) as real-time signals to trim frothy areas and to identify high-conviction entrants for diligence
• Maintain long-term optionality with small, concentrated allocations to early-stage/venture opportunities while enforcing strict value-creation and unit-economics criteria given high expected churn