Cerebras CEO Andrew Feldman to discuss AI scaling limits at TechCrunch Disrupt 2026
The chip company's co-founder will address growing demand for compute, energy and infrastructure, and what happens if today's AI hardware reaches its limits.
Written by EGazette’s AI. The facts are drawn from cited sources; the analysis is the AI’s own.
Cerebras Systems CEO and co-founder Andrew Feldman is set to speak at TechCrunch Disrupt 2026 about whether artificial intelligence can continue to scale at its current pace, according to TechCrunch.
Feldman's session is expected to explore the growing demand for computing power, energy and infrastructure needed to support AI development, along with how Cerebras is approaching those constraints differently from other chipmakers in the industry.
Addressing AI hardware limits
A central theme of the discussion is expected to be what comes next if today's AI hardware reaches its physical or economic limits, a question that has become increasingly prominent as AI models grow larger and more resource-intensive to train and run.
Cerebras Systems has positioned itself as an alternative to dominant AI chipmakers by designing large, wafer-scale processors intended to handle AI workloads differently from traditional graphics processing units. The company's approach to compute architecture has drawn attention as the broader industry grapples with rising costs and energy demands tied to AI infrastructure buildouts.
TechCrunch's report did not provide further specifics on what positions or predictions Feldman plans to share during the session, beyond the general topics of compute, energy and infrastructure constraints.
The discussion comes amid broader industry debate about whether current approaches to scaling AI models, largely through adding more computing power and data, can continue indefinitely or whether the sector will need new architectural or algorithmic breakthroughs to sustain progress.
Energy availability, in particular, has become a growing concern across the AI industry, as data centers supporting large-scale model training and inference place mounting demands on power grids in various regions.
TechCrunch Disrupt, an annual technology conference, regularly features executives from leading AI and infrastructure companies discussing the industry's direction, and Feldman's appearance is part of that broader programming for the 2026 event.
Sources
- Cerebras Systems’ Andrew Feldman on whether AI can keep scaling at TechCrunch Disrupt 2026 — TechCrunch
EGazette summarizes reporting from multiple sources; follow the links for the originals.
Comments
Loading comments…
Related articles
Exhibitor registration for TechCrunch Disrupt 2026 closes October 2
Startups have three days left to book a booth at the conference, which organizers say draws more than 10,000 founders, investors and tech leaders.
Google reportedly piloting payments to publishers for AI search contributions
A new program reportedly pays around 100 publishers for content used in Google's AI-powered search features.
Pledge signed by Trump and AI leaders contains misspelling of "United States"
A pledge document signed by President Trump and prominent AI industry figures reportedly misspells the country's name.
Google launches Fitbit Air in India at a higher price than in the US
The new fitness tracker is priced around $146 in India, about $47 more than its US cost, raising questions about its appeal to price-sensitive buyers.
Instagram launches AI video editor with personalized feedback for creators
The new conversational AI assistant in Instagram's Edits app is designed to give creators tailored suggestions rather than generic editing tips.

Tinder launches 'Group Hangouts' feature as dating apps lean into social, in-person experiences
The new feature lets groups of friends organize hangouts through the app, part of a broader industry shift away from endless swiping.