SEB Analysts Say 2026 Marks Shift From Virtual to Physical AI

Why It Matters

Embedding AI in hardware promises a new wave of productivity growth and creates fresh investment opportunities in the AI value chain, reshaping both technology markets and financial services strategies.

Key Takeaways

• AI hardware demand outpaces software returns since late 2025
• Autonomous vehicle and robot deployments remain fraction of global need
• Semiconductor equipment firms see share price surges amid AI bottlenecks
• Banks report early AI ROI but struggle measuring true productivity
• Economic impact expected post‑2026 as AI embeds in capital

Pulse Analysis

The dramatic reduction in AI operating costs during 2025 has democratized access to sophisticated models, yet the productivity lift from pure software remains limited. Analysts at SEB argue that the next frontier lies in marrying AI algorithms with physical assets, turning concepts like autonomous freight trucks or factory‑floor robots into revenue‑generating machines. This hardware‑centric evolution mirrors past industrial revolutions where the real economic boost arrived only after the supporting infrastructure—sensors, processors, and power systems—was widely deployed.

Equity markets have already begun to price this transition. While cloud hyperscalers and pure‑play model developers have seen flat or declining shares since late 2025, firms that produce semiconductors, advanced packaging, and AI‑optimized equipment are enjoying double‑digit gains. The surge reflects a tightening supply chain for AI‑specific chips, prompting investors to reallocate capital toward companies that can alleviate these bottlenecks. Simultaneously, autonomous‑vehicle operators such as Waymo and Chinese robot manufacturers are scaling pilots, but volumes remain a fraction of the projected demand, underscoring a near‑term gap between ambition and execution.

For financial institutions, the shift carries both opportunity and risk. Early adopters like HSBC claim measurable ROI from AI‑driven analytics, yet the lack of standardized metrics hampers broader confidence. Regulators are also tightening oversight, demanding ethical safeguards as AI becomes more autonomous. Looking ahead to 2030, embedded AI in sectors ranging from mining equipment to logistics is expected to become mainstream, potentially igniting a productivity surge reminiscent of past technological inflection points. Investors who position themselves in the hardware side of the AI ecosystem stand to benefit from the long‑term upside as the physical AI economy matures.