The AI ROI reckoning

Navigating the AI Gold Rush: VCs Unpack Investment Strategies and Future Returns at Web Summit Vancouver 2026

The Web Summit Vancouver 2026 panel, “The AI ROI Reckoning,” explored venture capitalists’ perspectives on the evolving AI landscape. Moderator Emil Protalinski, with panelists Ms. Aidan Madigan-Curtis (Eclipse Ventures), Mr. Simon Wu (Cathay Innovation), and Ms. Feyza Haskaraman (Felicis Ventures), explored AI stack strategies, startup valuations, and the emerging concept of sovereign AI.

Ms. Madigan-Curtis emphasized Eclipse Ventures’ “picks and shovels” approach, focusing on physical world technology like manufacturing, defense, energy, and supply chain. Investments target foundational layers such as power generation, data centers, and advanced GPUs, with portfolio companies like Cerebros and TenseTorrent developing next-generation chips to meet AI demands.

Mr. Wu highlighted Cathay Innovation’s focus on applied AI at the application layer, viewing AI as augmenting labor. He noted that enterprises seek tangible outcomes and ROI. Cathay invests in verticals where AI can significantly reduce friction and add value, creating a synergistic effect beyond simple automation.

Ms. Haskaraman observed the massive $260 billion invested in AI recently, predominantly in foundational models. She stressed the critical need for robust infrastructure to support the AI-first economy. Felicis Ventures actively invests in hard tech “picks and shovels,” including data centers, photonics, and optical switches, citing their investment in Crusoe as an example.

Regarding ROI, Ms. Haskaraman believed the physical layer, particularly robotics, offers the greatest long-term GDP impact, though a major robotics wave requires more data. Mr. Wu agreed on robotics’ long-term potential for blue-collar augmentation, while acknowledging current significant value in knowledge work like coding and customer service.

Mr. Wu noted a shift in venture capital from “right price” to “right company” and “best teams.” He explained that rapid AI advancements necessitate investing in adaptable teams that can quickly leverage cutting-edge tools to deliver cost-effective solutions, often justifying higher initial valuations.

Ms. Haskaraman attributed high valuations in foundational AI to the scarcity of specialized talent. Deals for “iconic teams” building complex architectures close quickly at market price, driven by the immense potential rewards. She noted a “reckoning” in app layer valuations, with entry prices decreasing for later entrants.

Mr. Wu highlighted the rise of open-source models offering cost alternatives to proprietary labs, making distribution a key differentiator. He also stressed the importance of team cohesion and loyalty in new AI labs, contrasting it with the “mercenary” nature of some talent, which can impact long-term perseverance.

Ms. Madigan-Curtis celebrated the “Revenge of the Nerds” era, where fundamental researchers are finally valued. She pointed to future frontiers like neuromorphic chips, spiked neural networks, and brain-computer interfaces (BCI). For next year’s top ROI, Mr. Wu chose “AI Bio” (healthcare services), and Ms. Madigan-Curtis selected “Behind the meter energy.”