Cerebras Systems will address the future of AI scaling at TechCrunch Disrupt 2026, where CEO Andrew Feldman will explore the challenges of growing compute, energy, and infrastructure demands. The company has been developing its wafer-scale computing approach since 2015, challenging conventional AI chip architectures.
Cerebras reported over 600 megawatts of data center capacity live or under contract by the end of 2027, with manufacturing capacity increasing more than tenfold in 2026. The company also plans to launch its first European data center this year, aiming for 200 megawatts by the end of 2027.
"Scaling AI isn’t solely about designing a faster processor. It requires enough physical infrastructure to put that compute to work," said Feldman. He emphasized that the company is addressing both compute and infrastructure challenges as AI demand accelerates.
The announcement follows Cerebras' August introduction of the CS-4, its latest generation of wafer-scale AI infrastructure. Feldman will discuss how growing demand for compute, energy, and infrastructure may affect AI’s future and what happens if conventional hardware reaches its limits.
Cerebras did not specify how its wafer-scale approach will compare to traditional chip architectures, and remains uncertain about whether current hardware can keep pace with AI’s growth. The company plans to provide further insights at TechCrunch Disrupt 2026.
Source: techcrunch
