Cerebras Systems’ Andrew Feldman on whether AI can keep scaling at TechCrunch Disrupt 2026

The landscape of artificial intelligence (AI) is rapidly evolving, with increasing demands for computational power, energy efficiency, and infrastructure capabilities. At TechCrunch Disrupt 2026, Andrew Feldman, the CEO and co-founder of Cerebras Systems, addressed these pressing issues, focusing on how the industry can continue to scale despite potential constraints. Feldman’s insights come at a critical time when many in the tech community are questioning whether current AI hardware can keep pace with the growing needs of advanced AI applications.

Cerebras Systems has made a name for itself by developing cutting-edge hardware specifically designed for AI workloads. The company’s flagship product, the Cerebras Wafer Scale Engine (WSE), is the largest chip ever built and is tailored for deep learning tasks. Feldman emphasized that as AI models become more complex and data-intensive, the traditional computing paradigms may no longer suffice. He proposed that innovative approaches to hardware design and energy consumption are essential for the future of AI.

Key facts

Who's involved

Cerebras Systems is at the forefront of AI hardware innovation, with its leadership team, including Andrew Feldman, driving the company’s vision. Other notable players in the AI hardware space include NVIDIA, which has long dominated the GPU market, and Google, known for its Tensor Processing Units (TPUs). These companies are also grappling with similar challenges related to scaling and efficiency in AI.

The conversation at TechCrunch Disrupt 2026 is not just about Cerebras; it reflects broader industry trends and the collective efforts of various organizations to push the boundaries of what AI can achieve. Feldman’s remarks are particularly relevant as they highlight the competitive landscape where companies are racing to develop the most efficient and powerful AI systems.

As AI continues to permeate various sectors, from healthcare to finance, the demand for robust infrastructure will only increase. This creates a fertile ground for innovation and collaboration among tech companies, researchers, and policymakers.

The challenges facing AI hardware are not new. Historically, the industry has seen rapid advancements in computing power, often referred to as Moore's Law, which predicts that the number of transistors on a microchip doubles approximately every two years. However, as we approach the physical limits of silicon-based technology, the industry must explore alternative materials and architectures to sustain this growth.

Feldman pointed out that while traditional computing methods have served the industry well, they may not be sufficient for the next generation of AI models. The increasing complexity of these models, coupled with the vast amounts of data they require, means that new approaches to hardware design are essential. For instance, the WSE from Cerebras is designed to handle massive parallel processing tasks, which are critical for training large AI models efficiently.

How to read the numbers

While specific performance metrics were not disclosed during the discussion, the implications of Feldman’s insights suggest a need for a new benchmark in AI hardware performance. The following table outlines some potential benchmarks relevant to AI hardware, though these are illustrative rather than definitive:

What you can do with it

For developers and organizations looking to leverage AI effectively, here are some practical takeaways from Feldman’s discussion:

• Invest in Specialized Hardware: Consider adopting hardware like Cerebras’ WSE that is optimized for AI workloads to improve training times and efficiency.
• Focus on Energy Efficiency: As energy costs rise, prioritize solutions that reduce power consumption without sacrificing performance.
• Stay Informed on Innovations: Keep an eye on emerging technologies and architectures that could redefine AI capabilities, such as neuromorphic computing or quantum computing.
• Collaborate Across Sectors: Engage with other tech companies and research institutions to share knowledge and resources in tackling the challenges of AI scaling.

What we're watching

As the AI landscape continues to evolve, the next milestone to watch will be the development of new hardware architectures that can surpass the limitations of current technologies. Companies like Cerebras are leading the charge, but the question remains: will they be able to maintain their competitive edge as more players enter the market? Additionally, the ongoing discussions around energy consumption and sustainability in AI will likely shape future innovations and regulatory frameworks.

Looking ahead, the industry must grapple with the implications of reaching physical limits in traditional computing. As AI models grow in size and complexity, the demand for innovative solutions will only intensify. Companies that can successfully navigate these challenges will not only advance their own technologies but also contribute to the broader evolution of AI as a transformative force across industries. The future of AI hardware is not just about scaling up; it's about rethinking how we approach computation itself.