Living cells derived from the human brain could soon be the key to addressing the global energy crisis.
Digitalization has led to the rapid expansion of artificial intelligence (AI) and high-density computing.
Now, power grids worldwide are struggling to keep pace with the ever-rising energy needs of traditional data centers.
Will shifting towards a more biological approach help meet electricity demands while ensuring that hyperconnectivity is maintained across all nations?
How the digital age is straining global grids
The world has swiftly transformed in the past decade, after many years of dedication to developing advanced technologies.
Now, society lives in a hyperconnected era driven by AI, which has altered how global data infrastructure is designed and executed.
Cloud-based computing and data transfers, 24/7 internet access, smart devices, digital services, and smart transport have become a given.
To ensure that AI remains up to date, advanced machine learning models are often used for training.
This, along with real-time processing, necessitates highly concentrated computational power, which continues to grow exponentially.
One query on an advanced AI model needs approximately tenfold the amount of electricity required for a traditional web search.
As these models scale to keep up with the developing digital world, the unprecedented pressure on existing IT infrastructure increases.
This means that AI powerhouses are now being perceived as major energy consumers worldwide.
The cost of keeping the world connected
Rapid transformation of the energy sector and data centers is making it difficult to ensure sustainability.
Projections show that global data center consumption will soon exceed 1,050 TWh annually.
Unfortunately, systemic barriers are preventing renewable capacity from meeting modern demands.
Solar and wind power’s intermittency leaves data centers vulnerable to grid disruptions in the absence of large, expensive battery systems.
Furthermore, aging critical infrastructure requires high-power transformers and grid upgrades. Delivery can take several years, causing a severe bottleneck.
Another systemic barrier to overcome is the cooling systems that prevent processors from overheating. These systems consume an equivalent amount of power to keep operations running.
Additionally, cooling facilities in drought-stressed regions require billions of gallons of freshwater annually. This creates conflict with local communities and ecosystems, as the world faces freshwater “bankruptcy.”
Fortunately, Cortical Labs engineers have developed a biological approach to meet demands in a less taxing way.
Synthetic biological computing: The role of living cells
The Australian startup is developing a new computing model called Synthetic Biological Intelligence (SBI).
In this approach, the engineers created a “humanized” device by turning to the human brain. This hybrid unit has intuitive, real-time learning capabilities.
Known as CL1, it is the first commercially available, code-deployable biological computer.
The living cells used for CL1 and its life-support system
Pluripotent stem cells were differentiated into functioning neurons. A living matrix was created by growing nearly 200,000 neurons on top of a silicon chip.
A nutrient-rich liquid modeled after cerebrospinal fluid supplies oxygen and glucose to CL1 while regulating temperature and gas. CL1’s biological components are viable for six months.
A Biological Intelligence Operating System (biOS) creates a simulated environment, relaying information to cells through electrical stimulation. The neurons then react to the environment in a closed-loop system.
CL1 already proved it could learn after playing the 1970s classic game Pong. Shortly after, the system was trained to play the first-person shooter game Doom.
Beyond its proven real-time learning and adapting skills, it also has ultra-low energy consumption. One CL1 unit only uses 30 watts of electricity, which is less than a standard calculator.
The Bio Data Center prototype has been launched in Melbourne, which houses 120 CL1 devices linked through the Cortical Cloud.
Singapore is up next to home 1,000 units with data center operator DayOne. It could soon become the sustainable pathway to power the next generation of AI.
