Key Takeaways from the Montgomery Summit

Market Overview

Markets were slightly weaker this week with both the S&P 500/Nasdaq 100 down roughly 1%. To us, tepid performance appeared to be due to a grab bag of (1) uncertainty over a new war with Iran and higher oil prices; (2) slightly weaker than expected US payrolls and (3) rising US treasury yields, with the 10yr +12bp this week to 4.26%. Our NPM Private Market Tracker*, which shows the average price performance of the 50 largest names in our internal Tape D® data, is -2% YTD vs. SP500/Nasdaq 100 -3%/-3%. (Bloomberg; NPM)

Sentiment Suggests the Party’s Not Over Yet

We recently returned from the Montgomery Summit conference for private technology companies and investors held in Santa Monica, CA on March 10-11.   Our top 5 takeaways from our discussions at the event include: 

1. Despite lofty valuations, many VC investors nonetheless see the opportunity set as being robust. We attribute this optimism to one word: “growth.”   Many addressable markets created by AI and other frontier technologies are expanding even faster than valuations.  Breakthroughs in generative AI, robotics, biotech, etc. are creating entirely new industries, allowing startups to scale to billions in value far more quickly than ever before.  Several investors compared the current wave to earlier platform shifts—such as the internet or mobile—where early valuations appeared high but ultimately proved justified by the scale of the companies that emerged.  In short, VC sentiment appears to be less focused on valuation and more on what specific projects or initiatives invested capital will be used to fund.  (NPM)

Venture firms are also increasingly focused on “second generation” companies that build on the infrastructure created by hyperscalers.  As one VC noted, they believe the hyperscalers and AI model providers are subsidizing the next crop of startups by creating low-cost infrastructure into which new application layer companies can interface.  The one area that appears to remain the biggest concern to most of the investors we met is not valuation, but the availability of qualified talent. (NPM)

2. Sentiment on the prospects for the US IPO market in 2026 remain mixed.  With SpaceX ($1.5tn), OpenAI ($830bn) and Anthropic ($350bn) all discussing listings as early as 2026, we see the potential for the IPO market to digest $200+bn of supply.  While we believe this supply is sufficiently telegraphed to the market, the view from many investors is that these three mega deals will largely “suck the oxygen out of the room” for other potential IPOs.  While we note that a number of smaller companies have indicated a desire to go public – including Cerebras, Quantinuum, Hudl, Entrada and Anaplan – the consensus appears to be that 2026 will be a year punctuated by a small number of very large IPOs, with SpaceX likely the first to go.  The view is not that the “floodgates” for IPOs will be open. (NPM; Bloomberg; The Information)

3. The rapid pace of technological change is opening new layers of the “stack”, from infrastructure and chips to applications and vertical software, creating fresh entry points for startups. We spoke with several investors that believe the LLM competitors are already “well-defined” but the world of physical AI (i.e. AI embodied into physical forms such as robots, defense tech and manufacturing equipment) is still wide open.  The belief is that physical AI will be slower to develop due to lower margin for error vs. LLM chatbots where there is generally no consequence for a “hallucination.”  However, many believe the market will grow quickly to significant scale once safety issues are resolved.  Other areas that we now hear mentioned that have been unloved for years include old school physical technologies such as memory, chips, storage and networking.

Lastly, several investors mentioned that Anthropic and, to a lesser extent, OpenAI have “cracked the code” on how to sell AI to enterprises.  While much has already been written on how this is potentially disruptive to the Saas business model, many investors see the opportunity for new and existing software companies to embrace AI successfully through new business models (such as consumption based vs. legacy seat- based Saas models). (NPM)

4. Innovation in power technology may be one of the great advancements from the AI buildout. Today, U.S. data centers consume roughly 4% of total U.S. electricity (Source: Pew Research), but most forecasts suggest this could rise to 8–12% (Source: Goldman Sachs; Gartner) by 2030.  The largest hyperscale AI facilities are already drawing up to 1 GW of power, comparable to the output of a nuclear power plant or the consumption of a medium sized city.  In our view, meeting this demand will require an “any and all” combination of new natural gas generation in the near term, expanded renewable capacity, and, farther out, potentially new nuclear technologies such as small modular reactors. (NPM)

In the face of this demand, the industry is also working on BTM (“behind the meter”) power innovations—such as AI-driven workload scheduling and microgrids combining local generation with battery storage— to help balance and optimize the massive energy demands. At the same time, operators are developing advanced efficiency technologies, such as liquid and immersion cooling systems that significantly reduce energy use.   The portion of the trillions of dollars spent by hyperscalers in the next 5 years earmarked for power technology could dwarf the amount normally spent by the entire domestic utility/IPP industry. (NPM; Bloomberg)

In short, we believe AI driven investment may be driving a broader energy revolution. Currently, hydrocarbons still supply ~86% of global energy (Source: American Gas Association).  We believe that, over time, electrons are poised to gain massive share. (NPM)

5. Following the flood of OpenClaw news out of China, AI agents were naturally a front of mind topic. AI agents are beginning to transform digital commerce by acting as autonomous assistants that can search, compare, negotiate, and complete purchases. While the technology is still early, many analysts expect agent-driven commerce to eventually transform shopping, travel booking, procurement, and financial services.  (NPM)

Today the market remains largely experimental, with, for example, many IT security firewalls not even permitting agent driven commerce.  However, we believe adoption of agents is inevitable.  VC investors speaking on a panel at the conference believe that agents will ultimately “increase the size of the pie” vs. simply cannibalizing existing e-commerce and physical shopping channels.  Points that were made include: (NPM)

• Agents will assist with “procurement” but not replace “shopping.” Shopping for many consumers is an experience with an emotional component.   Agents will remove the human experience, which not every consumer will embrace. (NPM)
• Agents will be better at assessing product quality than humans. Humans shop partly based on brand familiarity.  Agents will be able to query and understand product quality data better and more quickly than humans, enabling better decisions on value. (NPM)
• Security remains a concern. Already there is a high incidence of fraud in the few agent driven transactions that are occurring.  Fraudsters are often first movers in adopting technology, and agentic technology may be no different.  Consumers will need to gain comfort with security and error correction, in our view, before agentic commerce becomes widespread. (NPM)