Cerebras Systems is redefining AI scalability by questioning the core assumptions of traditional chip architecture. At the upcoming TechCrunch Disrupt 2026 event, CEO and co-founder Andrew Feldman will discuss the surging demand for computing power and the infrastructure that supports it. This topic is particularly timely as AI models evolve, necessitating not just improved algorithms but also substantial increases in compute, energy, and infrastructure. The session titled “[Can AI Keep Scaling?]” aims to illuminate how these challenges are being addressed and what the future holds for AI hardware.
Founded in 2015, Cerebras has established itself in the field of wafer-scale computing. Rather than slicing silicon wafers into individual chips, the company has created a processor that is built directly on the wafer itself. This innovative architecture is tailored to manage the intense workloads that AI demands. With the recent launch of the CS-4, the latest generation of its wafer-scale AI infrastructure, Cerebras is expanding its approach to meet rising demands.
As the company grows, Cerebras has made impressive advancements in its operational capabilities. It raised $5.5 billion during its IPO in May and secured a multiyear agreement with OpenAI to deploy 750 megawatts of its systems from 2026 through 2028. Currently, Cerebras has reported over 600 megawatts of data center capacity that is either operational or under contract for delivery by the end of 2027. To further enhance its infrastructure, the company plans to increase its manufacturing capacity more than tenfold in 2026 and is set to launch its first European data center this year, with an anticipated expansion to 200 megawatts by the end of 2027.
Feldman stresses that merely designing faster processors won’t solve the scalability issue. Effective scaling also requires sufficient physical infrastructure to utilize that compute power. As AI computing expands, the demand for data centers, electricity, cooling, and manufacturing capacity becomes crucial. At Disrupt, Feldman will provide context on these challenges for founders, investors, and technology leaders, offering insights into future compute demands and the limitations of current hardware.
The discussion at Disrupt will be particularly relevant as Cerebras continues to discover innovative solutions to the AI scalability challenge. Attendees will have the chance to learn from an entrepreneur who has dedicated over a decade to this mission. The event will feature more than 200 sessions, with participation from over 10,000 founders, investors, operators, and tech leaders, all eager to delve into the forefront of technology.
As AI continues to progress, one question lingers: how much longer can it keep scaling? Insights from Feldman’s session will be invaluable for anyone involved in developing, funding, or deploying AI technologies. Attendees can gain crucial knowledge about the infrastructure needed to support the next phase of AI advancements and explore what changes are necessary to ensure that AI can continue to grow without hitting a hard wall.
