The rapid buildout of AI infrastructure is increasing pressure on already complex global supply chains for semiconductors, memory and advanced manufacturing equipment. In an interview published by The Verge, Harvard Business School professor Willy Shih argued that AI demand is consuming computing resources at unusual scale, while trade barriers and shifting industrial policies are adding more uncertainty to how technology products are sourced and priced.

For European businesses, this matters because AI adoption does not depend only on software models and cloud services. It also depends on the availability and cost of the hardware underneath them, from data-centre chips and memory to networking components and devices used across everyday business operations.

What happened

The Verge interview focused on the broader economics behind global supply chains, but one theme stands out for business readers: AI infrastructure demand is creating additional strain on technology manufacturing.

Shih described AI as consuming global computing resources during a major infrastructure expansion. In the interview, he also pointed to memory-chip demand as an example of where capacity can be outpaced by AI-related needs, making some components harder or more expensive to obtain.

The discussion also covered wider supply-chain dynamics, including:

• Heavy reliance on specialised global production networks for chips, components and assembly.
• Trade disruptions and tariffs that can redirect supply chains rather than rebuild them locally.
• Concentration of manufacturing capability in a limited number of countries and suppliers.
• Long investment cycles for fabs and advanced industrial capacity, which make rapid reshoring difficult.

These points were discussed in the context of the US policy debate, but the practical implications are broader. European companies buying cloud services, enterprise devices, networking equipment or AI-enabled tools are exposed to the same global supply conditions.

Why it matters for European businesses

Many companies view AI adoption mainly through the lens of software subscriptions, copilots and automation tools. But the supply side matters too. If AI demand pushes up the cost of memory, accelerators, servers or related infrastructure, those increases can affect pricing across the technology stack.

This can matter to European businesses in several ways:

• Higher IT costs: hardware inflation can feed into more expensive servers, laptops, mobile devices and cloud services.
• Slower AI projects: limited infrastructure availability may delay deployments that depend on GPUs, high-performance hosting or specific cloud regions.
• Vendor concentration risk: businesses relying on a small set of hyperscalers or hardware vendors may have less pricing leverage.
• Budget pressure on SMEs: smaller firms can be priced out of advanced AI use cases if infrastructure costs remain high.
• Procurement uncertainty: lead times for equipment may become less predictable during periods of strong demand or trade disruption.

For companies in Europe, the issue is not only direct chip buying. Most will feel the impact indirectly through SaaS pricing, cloud contracts, digital transformation budgets and the cost of replacing or expanding their IT environment.

Who may be affected

The effects are likely to be strongest for organisations with significant digital infrastructure needs or ambitious AI plans.

• SMEs adopting AI tools that depend on cloud-based model access and may face higher subscription or usage costs over time.
• E-commerce businesses using AI for search, recommendations, content generation or customer support at scale.
• IT teams and digital agencies managing hosting, cloud environments, device refresh cycles and AI-enabled workloads.
• Manufacturers and logistics businesses investing in automation, computer vision or industrial AI systems.
• Founders and finance teams responsible for forecasting technology spend under uncertain supply conditions.

The interview also reinforces a wider point: countries can announce industrial strategies, but actual production capacity depends on years of investment, skilled labour, specialist suppliers and stable economic conditions. That makes quick fixes unlikely, even when governments want more domestic manufacturing.

What companies should consider

European businesses do not need to react by pausing AI adoption, but they should treat infrastructure availability and cost as part of AI planning.

• Review AI business cases carefully: stress-test whether projected returns still hold if cloud or compute costs rise.
• Map supplier dependencies: identify where critical tools rely on a narrow set of cloud or hardware providers.
• Plan IT procurement earlier: for major infrastructure purchases, build in more time for sourcing and delivery risk.
• Prioritise high-value use cases: focus AI investments on workflows where efficiency gains clearly outweigh usage costs.
• Ask vendors about resilience: cloud, SaaS and managed-service providers should be able to explain capacity, hosting regions and pricing assumptions.
• Watch indirect cost exposure: rising component prices can appear later in device renewals, hosting contracts and software fees.

The source is an interview rather than a formal market report, so it does not establish precise price forecasts or timelines. However, it offers a useful reminder that the AI boom is not only a software story. For businesses in Europe, practical AI adoption increasingly depends on resilient access to the global hardware and infrastructure supply chain behind it.