TechCrunch has announced a Disrupt 2026 session featuring Cerebras Systems CEO and co-founder Andrew Feldman on whether AI can keep scaling as demand for compute, energy and infrastructure continues to grow. While this is an event announcement rather than a policy or product launch, the topic is relevant for businesses planning AI adoption because the economics and availability of AI infrastructure affect costs, performance and deployment choices.

What happened

According to TechCrunch, Feldman is scheduled to discuss the growing demand for AI compute, power and infrastructure, how Cerebras says it is approaching those constraints differently, and what may happen if current AI hardware reaches practical limits.

Based on the source, this is a forthcoming conference discussion, not a confirmed change in market rules, pricing or regulation. No new product details, commercial terms or independently verified technical claims were provided in the announcement.

Why it matters for European businesses

For many European companies, AI strategy increasingly depends not only on software models but also on access to underlying infrastructure. Generative AI tools, AI agents, model fine-tuning and large-scale automation workflows can become more expensive or harder to scale when compute capacity is constrained or when energy costs rise.

This matters in several practical areas:

• Cost planning: AI usage costs may remain volatile if infrastructure demand keeps growing faster than supply.
• Deployment choices: Businesses may need to choose between public cloud AI services, managed platforms, smaller models or more targeted use cases.
• Performance expectations: Response speed, throughput and availability can depend on hardware capacity and model size.
• Vendor risk: Heavy reliance on a small number of AI infrastructure providers can create concentration risk for pricing, service continuity and roadmap decisions.
• Sustainability and energy: Energy consumption is becoming a more visible factor in enterprise AI discussions, especially for companies with ESG reporting or operational efficiency targets.

For SMEs in particular, the business question is often not whether AI can scale in theory, but whether it can scale affordably for real workflows such as customer support, content operations, analytics, internal knowledge search or marketing automation.

Who may be affected

• SMEs adopting generative AI: Companies using paid AI platforms may be exposed to pricing and usage limits shaped by infrastructure costs.
• IT and digital teams: Teams evaluating model hosting, API usage or AI integration into business systems need to consider long-term capacity and vendor dependence.
• E-commerce and marketing teams: Businesses using AI for product content, personalization, campaign support or customer service may see infrastructure economics influence ROI.
• Founders and business owners: Leadership teams making AI investment decisions should distinguish between experimental pilots and sustainable production use.
• Digital agencies and software providers: Service firms building AI-enabled offerings may need to reassess architecture, margins and service-level assumptions.

What companies should consider

Even though this source is only an event preview, it highlights issues businesses should already be monitoring.

• Prioritise high-value use cases: Focus AI investment on workflows with measurable business outcomes rather than broad experimentation.
• Review model strategy: In some cases, smaller or more specialised models may offer better cost-performance trade-offs than the largest available models.
• Assess infrastructure dependency: Understand which providers, APIs or platforms your AI workflows rely on and where concentration risk exists.
• Track unit economics: Measure cost per task, per user or per workflow so AI usage can be managed as demand grows.
• Build fallback options: Where possible, avoid architectures that depend entirely on a single model or infrastructure vendor.
• Include energy and compliance considerations: Larger AI deployments may require broader review of operational efficiency, data handling and governance.

For European business leaders, the broader takeaway is that AI adoption is no longer only a software decision. Infrastructure availability, power requirements and hardware economics are becoming part of the commercial reality behind enterprise AI.