Brain Cell-Powered Data Centers: Neuron DOOM Company’s Next Leap

Cortical Labs, the Australian startup known for growing neurons on silicon and teaching them to play games, is pushing its biological computing platform beyond headline-grabbing demos and toward infrastructure. The company, which first drew global attention for training lab-grown neurons to play Pong and more recently showed a neuron-based system running Doom, now says its CL1 platform can be deployed locally or through its “Cortical Cloud.” That shift has fueled fresh interest in whether brain cell-powered data centers could become a serious alternative to parts of today’s energy-hungry AI stack.

From Pong to Doom to Compute Infrastructure

Cortical Labs has spent several years developing what it calls biological computing: systems that combine living neurons with silicon hardware. In 2021, the company demonstrated that neurons grown on a chip could learn to play Pong, a milestone that helped establish the commercial and scientific identity of its “DishBrain” work. In late February and early March 2026, the company and outside coverage highlighted a newer demonstration in which neurons on its CL1 biological computer were used to play Doom, extending the concept from a simple arcade task to a more complex game environment.

The more important development for industry is not the game itself. It is the platform behind it. Cortical Labs describes the CL1 as the world’s first code-deployable biological computer and says developers can run workloads either on a local machine or through a distributed cloud offering. That framing moves the company from experimental neuroscience into the language of commercial compute infrastructure.

According to Cortical Labs, the system merges biology with traditional computing so software can interact directly with real neurons. The company argues that such systems may require less energy and less training data than conventional AI approaches for certain learning tasks. Those claims remain early-stage and application-specific, but they help explain why the idea of brain cell-powered data centers is attracting attention at a time when AI power demand is under intense scrutiny.

The Company That Made a Dish of Neurons Play DOOM Is Getting Into Brain Cell-Powered Data Centers

The phrase “brain cell-powered data centers” can sound futuristic, but the near-term concept is narrower than a wholesale replacement for conventional servers. Cortical Labs is positioning racks or clusters of biological computers as specialized compute resources for research and selected machine-learning or bioscience tasks, rather than as a drop-in substitute for every CPU or GPU workload. Coverage of the CL1 points to a shoebox-sized system that sustains living neurons in a controlled environment while interfacing them with a silicon chip that sends and receives electrical signals.

The company’s own site says users can deploy technology “locally on the CL1, or distributed on the Cortical Cloud.” That language suggests a model closer to hosted access, clustered biological compute, or hybrid lab infrastructure than to a traditional hyperscale data center. Even so, the strategic direction is clear: Cortical Labs is trying to turn a laboratory novelty into a service platform.

Recent reporting around the Doom demonstration has also emphasized the CL1 as a commercial product rather than a one-off experiment. Live Science reported in 2025 that the CL1 had been released on March 2, 2025, and described it as a purchasable biological computer designed for applications including disease modeling and drug discovery. That matters because data-center ambitions require repeatable hardware, not just viral demos.

What the CL1 actually is

Public descriptions of the CL1 show a system built around several core elements:

• Lab-grown human neurons
• A silicon chip that exchanges electrical signals with those neurons
• A nutrient and life-support environment to keep the cells viable
• Software tools that let developers deploy code to the platform locally or remotely

In practical terms, this means the “compute” comes from a hybrid of wet biology and electronics. The neurons are not floating freely as a science-fiction brain. They are maintained in a tightly controlled device designed to support experiments and programmable interaction.

Why the Data Center Angle Matters

The timing of Cortical Labs’ push is significant. AI infrastructure costs have risen sharply as companies race to secure more GPUs, more power, and more cooling capacity. Against that backdrop, any computing architecture that promises lower energy use draws immediate interest. Cortical Labs says biological neural systems can be highly energy efficient, and outside coverage has described the CL1 as operating at far lower power than large AI training clusters.

That does not mean neuron-based systems are ready to compete with mainstream AI accelerators on raw throughput. They are not marketed that way. Instead, the appeal lies in the possibility that biological systems may learn differently, adapt with less data, or support new forms of experimentation in neuroscience, drug discovery, and adaptive control systems. If those strengths hold up under broader testing, specialized biological compute clusters could carve out a niche inside research institutions, biotech labs, and eventually commercial facilities.

For stakeholders, the implications vary:

• Researchers may gain a new platform for studying learning and neural behavior.
• Biotech companies may see tools for disease modeling and screening.
• AI infrastructure investors may view biological computing as a long-shot but potentially disruptive category.
• Regulators and ethicists are likely to focus on sourcing, oversight, and the treatment of living neural material.

Scientific Promise and Commercial Limits

The excitement around Cortical Labs is real, but so are the constraints. Publicly available information does not show that brain cell-powered data centers are ready for broad enterprise deployment. The current evidence supports a more modest conclusion: Cortical Labs has built a commercial biological computing platform, demonstrated that it can perform interactive tasks such as game control, and is now framing that platform as deployable infrastructure through local systems and cloud access.

That leaves several open questions. One is scale. Another is reproducibility across many units. A third is economics: even if biological systems use less power for some tasks, they also require specialized maintenance, environmental control, and biological handling. Those operational realities could limit adoption outside high-value use cases. This is one reason the most credible near-term applications remain in research and biomedical fields rather than general-purpose cloud computing.

There is also the issue of public perception. A company that made a dish of neurons play Doom naturally attracts attention, but novelty can obscure the harder question of utility. The commercial test will be whether customers can use the platform to solve problems faster, cheaper, or better than existing tools. So far, Cortical Labs has made the strongest public case in areas such as drug discovery, disease modeling, and experimental computing.

Ethics, Oversight, and the Debate Ahead

Biological computing sits at the intersection of neuroscience, AI, and bioethics, which means scrutiny is inevitable. The use of human neurons in computing systems raises questions about consent, sourcing, governance, and the boundaries of acceptable experimentation. Public reporting on the CL1 describes neurons grown in lab conditions and maintained in a controlled support system, but broader ethical frameworks will likely evolve as the technology matures.

There is also a risk of overstatement in the public conversation. Terms such as “brain-powered” or “living computer” can imply capabilities that current systems do not have. The available evidence does not suggest consciousness or human-like cognition. It suggests a programmable biological substrate that can respond to stimuli, adapt, and perform narrow tasks under engineered conditions. That distinction is essential for both investors and the public.

According to Cortical Labs’ public materials, the company sees biological computing as a path toward more efficient learning systems and new healthcare applications. Whether that vision becomes a durable industry segment will depend on peer-reviewed validation, customer adoption, and regulatory clarity over the next several years.

Conclusion

Cortical Labs is no longer just the company that made a dish of neurons play games. With the CL1 and Cortical Cloud, it is attempting to turn biological computing into a deployable infrastructure category. The leap from Pong and Doom to brain cell-powered data centers is still early, and the technology remains specialized, experimental, and closely watched. But the direction is unmistakable: a startup once known for a striking neuroscience demo is now making a serious bid to shape the future of low-power, biologically informed computing.

Frequently Asked Questions

What company made neurons play Doom?
Cortical Labs, an Australian biological computing startup, is the company associated with neuron-on-chip systems that progressed from Pong demonstrations to a 2026 Doom demo on its CL1 platform.

What is the CL1?
The CL1 is Cortical Labs’ biological computer, which combines lab-grown human neurons with silicon hardware and software tools for local or cloud-based deployment.

Are brain cell-powered data centers real today?
In an early and limited sense, yes. Cortical Labs is offering biological computing hardware and cloud access, but this is not a mainstream replacement for conventional data centers. It is a specialized platform at an early commercial stage.

What are the likely first uses?
The most plausible early uses are in research, disease modeling, drug discovery, and experimental adaptive computing rather than general-purpose enterprise workloads.

Why is this attracting attention now?
The technology is emerging as AI infrastructure costs and energy demand rise. Cortical Labs argues that biological systems may offer energy-efficient learning for some tasks, though those claims still need broader validation.

Does this mean the system is conscious?
Publicly available information does not support that conclusion. The current systems are described as lab-grown neurons interacting with silicon in controlled environments for narrow computational tasks.