Summary
Cerebras Systems CEO and co-founder Andrew Feldman has pushed back against growing fears that artificial intelligence could pose an existential threat to humanity. Speaking at TechCrunch Disrupt 2026, Feldman argued that public discussions about AI risks are overshadowing the technology’s potential benefits in areas such as healthcare, scientific research, and productivity.
While acknowledging that powerful technologies can be misused, Feldman called for sensible regulation and better engagement with local communities affected by the rapid expansion of AI data centres. He also said improvements in computing and model efficiency should be viewed as complementary developments rather than threats to the AI hardware industry.
What Happened?
Concerns about the future of artificial intelligence have become increasingly prominent, with researchers and technology leaders debating the possibility of advanced AI systems becoming difficult to control.
At TechCrunch Disrupt 2026, Andrew Feldman challenged what he sees as an overly negative narrative surrounding the technology. He suggested that some prominent AI leaders could communicate their concerns more carefully, particularly when discussing worst-case scenarios.
Feldman argued that the public conversation should also acknowledge what AI might achieve if developed and deployed responsibly. As an example, he pointed to the possibility of medical breakthroughs that could help future generations avoid diseases such as cancer.
His comments come amid a wider debate within the technology industry over how to balance the potential benefits of increasingly capable AI systems against concerns about safety, misuse, and long-term risks.
Why Does the Cerebras CEO Disagree With AI Doomsday Warnings?
Feldman’s central argument is that focusing too heavily on catastrophic scenarios can overshadow the practical benefits AI may deliver.
Artificial intelligence is already being explored across several fields, including healthcare, drug discovery, software development, scientific research, and industrial automation. Continued progress could help researchers analyze complex data, accelerate experiments, and develop new approaches to difficult problems.
Feldman believes technology leaders should communicate these opportunities alongside legitimate safety concerns.
However, his position does not amount to dismissing AI risks entirely. During his interview, he acknowledged that powerful technologies can create serious challenges when used by people with harmful intentions.
His message was that the industry should recognize both sides of the technology: its ability to produce meaningful advances and the need to manage the risks associated with its use.
Feldman Calls for Sensible AI Regulation
Although Feldman criticized the increasingly negative tone of some AI discussions, he also emphasized the importance of responsible development and regulation.
He argued that powerful technologies require careful oversight because they can be misused. In his view, the industry should work toward reasonable rules that address genuine risks without unnecessarily restricting beneficial innovation.
The challenge for regulators is to establish safeguards that keep pace with rapid technological development while allowing researchers and businesses to continue building useful applications.
This debate is particularly important as AI systems become more capable of performing complex tasks with limited human intervention. Questions surrounding accountability, cybersecurity, privacy, and the control of advanced AI systems remain central to discussions about their future.
AI Data Centres Face Growing Community Opposition
Feldman also addressed another issue affecting the AI industry: the expansion of data centres needed to support AI development.
Building and operating these facilities can require substantial electricity, land, cooling infrastructure, and investment. As demand for computing capacity increases, companies are planning new facilities in markets around the world.
According to Feldman, the industry has not always done enough to involve local communities in decisions about these developments, particularly in the United States.
He argued that companies should engage residents earlier and explain the potential effects of data centre projects rather than treating community concerns as an afterthought.
Local opposition can arise from concerns about electricity demand, water consumption, environmental effects, infrastructure pressure, and how the economic benefits of new facilities will be distributed.
These concerns have become an important part of the broader debate about the cost of expanding AI infrastructure.
Why Are Some Countries Attracting More Data Centre Investment?
During the discussion, Feldman contrasted opposition to certain data centre projects in the United States with opportunities he sees in countries including Brazil, Canada, Finland, Iceland, Norway, Mexico, and Sweden.
These markets offer different combinations of energy resources, infrastructure, climate conditions, and investment opportunities that may appeal to data centre operators.
However, the suitability of each location depends on factors such as electricity availability, grid capacity, regulations, connectivity, and community acceptance.
Feldman’s comments highlight a growing challenge for the AI industry: computing capacity is no longer determined solely by access to advanced chips. Access to power, suitable locations, and public support can also influence where companies build their infrastructure.
For AI companies, maintaining constructive relationships with local communities may become increasingly important as demand for data centres grows.
Cerebras CEO Says AI Efficiency Will Not Necessarily Reduce Chip Demand
Feldman also discussed whether improvements in AI efficiency could reduce demand for specialized computing hardware.
As AI models become more efficient, they may require fewer computing resources to complete certain tasks. Similarly, advances in hardware can improve performance while reducing the resources needed for a given workload.
Feldman argued that these developments should work together rather than be viewed as competing trends.
Hardware companies can develop more efficient computing systems, while AI developers can optimize their models to achieve better results using fewer resources.
At the same time, lower costs and improved performance could make AI applications accessible to more users and businesses, potentially increasing overall demand for computing.
For Cerebras and other AI chipmakers, the key question is how the market evolves as models become more capable and AI adoption expands.
What Is Cerebras Systems?
Cerebras Systems is a technology company that develops specialized computing hardware for artificial intelligence workloads.
The company is known for its wafer-scale processor approach, which uses a large silicon wafer as the basis for a processor designed to handle demanding AI computations.
Cerebras operates in a competitive market that includes established chipmakers and other companies developing hardware for AI training and inference.
Its business is closely connected to the growth of AI infrastructure, making developments in computing efficiency, model demand, and data centre investment particularly relevant to its future.
Feldman’s comments also reflect the interests of a company seeking continued growth in the AI market. His perspective is therefore an important contribution to the debate, but it should be considered alongside views from independent researchers, policymakers, and other industry participants.
What Does This Mean for the Future of AI?
Feldman’s remarks reflect a broader disagreement over how the public and technology companies should discuss artificial intelligence.
Some researchers emphasize the possibility of severe long-term risks, including the potential loss of control over increasingly capable systems. Others focus more heavily on the technology’s capacity to improve productivity, accelerate scientific discoveries, and create new products and services.
These perspectives are not necessarily mutually exclusive. AI could produce substantial benefits while also creating risks that require technical safeguards, responsible deployment, and effective regulation.
The industry must also address practical challenges involving energy consumption, data centre construction, cybersecurity, and public trust.
Finding a workable balance between innovation and oversight will be important as AI technology continues to develop.
Conclusion
Cerebras CEO Andrew Feldman believes the AI industry’s public image has become too focused on catastrophic predictions and not focused enough on the technology’s potential benefits.
His comments at TechCrunch Disrupt 2026 highlighted possible advances in healthcare and research, the need for reasonable AI regulation, and the importance of working more closely with communities affected by data centre expansion.
He also argued that improvements in hardware and model efficiency can support the continued development of AI rather than automatically reducing demand for computing resources.
Ultimately, the debate is not simply about whether AI will benefit or harm society. It is about how the technology can be developed responsibly, how its benefits can be realized, and how its risks and wider social costs can be managed.