The artificial-intelligence investment story increasingly extends beyond the companies building headline-grabbing generative-AI models. For thematic investors, it also encompasses the physical infrastructure that enables automation and the cybersecurity systems protecting a growing number of digital identities.
Pictet Asset Management seeks exposure to these developments through three long-term investment themes—Robotics, Digital, and Security—rather than through a dedicated AI fund. This approach treats AI as an enabling technology spanning several industries, Anjali Bastianpillai, senior client portfolio manager at Pictet Asset Management, said in an interview on 18 August 2026.
Three routes into the AI economy
The robotics strategy covers what Bastianpillai described as “physical AI,” including semiconductor-factory automation and sensors used in autonomous vehicles. The security strategy focuses on protecting data centres, cloud environments, and digital identities. The digital strategy encompasses cloud adoption and the commercial applications built on that infrastructure.
Together, the strategies illustrate how AI investment can extend beyond model developers and semiconductor designers. They also expose investors to different risks, from industrial cycles to software valuations and cybersecurity competition.
Escaping the trap of crowded trades
Passive technology exposure can be heavily concentrated in a few large companies. At 31 July 2026, Nvidia, Apple, and Microsoft accounted for approximately 48% of the MSCI World Information Technology Index.
Instead of simply buying hardware providers, Bastianpillai argued that investors should seek “pure-play” innovators where revenues are strictly tied to the theme. For example, Tesla is often cited as a robotics play due to its “Optimus” humanoid, but for purists, its car-heavy revenue stream makes it a low-purity investment.
From creative tools to commercial applications
Generative AI entered the mainstream after ChatGPT’s launch in November 2022. Image-generation products such as DALL·E 3 attracted attention in 2023, while OpenAI presented its Sora video model the following year.
Investor attention has since shifted towards coding tools and specialised business applications, according to Bastianpillai. The challenge for technology companies is no longer simply to develop capable models but to turn those capabilities into sustainable revenue.
Growing demand for computing capacity has also increased the strategic importance of custom silicon. Alphabet and Amazon have expanded their use of in-house chips as they seek to control costs, optimise performance, and reduce their reliance on external suppliers.
Securing an agentic workforce
Cybersecurity is also changing as companies deploy AI agents capable of acting on behalf of employees. Unlike the traditional model in which a person might have a relatively limited number of managed accounts, agentic systems can create many additional machine identities, credentials, and access relationships. That proliferation would expand the potential attack surface and increase demand for identity-management providers such as CyberArk and Okta.
Software shares came under pressure during the early-2026 sell-off, sometimes described as the “SaaSpocalypse,” as investors questioned whether generative AI would weaken established subscription-software businesses. Bastianpillai maintained that the underlying businesses of selected identity-security companies remained resilient despite the market decline.
That resilience cannot be taken for granted, however. Security providers must continually update their products as attackers adopt faster and increasingly automated techniques. For investors, the opportunity therefore comes with a central uncertainty: whether today’s cybersecurity leaders can innovate quickly enough to remain tomorrow’s leaders.



