Cerebras Systems CEO Andrew Feldman is set to discuss AI scaling at Disrupt 2026, according to Crypto Briefing. The session puts the chipmaker's approach to building out AI compute in front of an audience that increasingly cares about how — and at what cost — data centers get built.
That's a narrow topic with broad reach. Cerebras' pitch on scaling could influence how operators think about data center strategy, and by extension where infrastructure dollars go.
Why Cerebras' scaling talk lands now
AI infrastructure has become one of the heaviest line items in tech budgets, and the companies building the hardware are under pressure to show that scale doesn't just mean more power draw and more racks. Cerebras has built its business around a different architecture than the GPU-heavy approach that dominates the market. Feldman's remarks at Disrupt 2026 are expected to address how that architecture fits into the next phase of scaling.
The stakes aren't confined to one company's roadmap. If Cerebras' scaling innovations can shift how data centers are planned and financed, the effects show up in infrastructure development well beyond the AI sector.
The investor angle
Tech investments tied to AI infrastructure have been sensitive to anything that changes the cost or feasibility of scaling. Cerebras' positioning matters to that calculus. A credible alternative to the prevailing compute model would give buyers more leverage and could redirect capital toward different supply chains.
None of that is guaranteed. Feldman is appearing to make the case, not to announce a completed shift in the market.
What's on the table at Disrupt
Disrupt 2026 gives Cerebras a platform to make its scaling argument directly to the people who allocate infrastructure spending. What Feldman says — and how specific he gets about the economics — will be the thing to watch. The appearance is scheduled; the substance is what the room will be listening for.
