Two key requirements of today's technological society are uninterrupted electrical power and the certainty of computer data.
And two incidents in the past several months have demonstrated that both have vulnerabilities when the demand exceeds the supply.
The April 28, 2025, blackout that disrupted life for millions in Spain and Portugal and caused widespread chaos was the result of a sudden loss of 15 GW of power; about 60% of Spain's national demand.
The blackout, which severely affected telecommunications, transportation, and emergency services over a 10-hour period, was apparently the result of a "generation disconnection" event.
When Amazon Web Services (AWS) later experienced a significant outage in October, largely due to a DNS issue in its DynamoDB service, popular platforms like Netflix, Snapchat, and Alexa were among the services that experienced significant disruptions.
An analysis of this outage, which lasted 15 hours, highlighted the critical role that Amazon's cloud services play in the functioning of many online platforms and applications.
To the data-dependent community, the AWS incident was a wakeup call about the inherent weakness of the interconnectedness of internet services and the potential for widespread impact from a single point of failure.
This weakness is exacerbated by the fact that just three major companies – AWS, Microsoft’s Azure, and Google – control roughly two-thirds of the world's active cloud services.
The Trump administration is considering reinstating the reprocessing of spent nuclear fuel to address shortfalls in reactor-ready fuel as it rebuilds the U.S. nuclear power industry.
Commercial reprocessing was abandoned by the U.S. half a century ago but continues in France, Russia, China, and Sweden.
The Department of Energy (DOE) was tasked by the President to work with other agencies and file a report by February 2026 on its feasibility and benefits.
Just as recycling of spent nuclear fuel could give a boost to the national grid without any additional mining.
Andrew Sobko, CEO and founder of Argentum AI, is already revitalizing inactive and underutilized graphics processing units (GPUs) to decentralize and democratize cloud computing, providing affordable access for startups and smaller companies and adding up to 25% of available computing power without any new chip fabrication.
Sobko built his reputation primarily in the supply chain logistics technology sector, uniting shippers, carriers, and international operators into a one-stop solution marketplace.
His next company, Batch, acquired NEXT Trucking backed by over $200 million from major investment groups.
That proficiency in AI-enhanced logistics led to the founding of Argentum AI, a pioneering decentralized computing marketplace.
Argentum scours the tech universe for discarded or inactive GPUs that its clients can tap into for fast computing work without waiting periods that have become the norm with the major platforms.
This universe includes computers formerly used for bitcoin mining and open AI that are being retired as even faster computers take their place.
These gently used GPUs are more than adequate for most smaller users, just as a used car works for many American families.
Sobko says the AWS incident showed the dangers of centralized computing for AI and centralized clouds for the web.
When everything is reliant on the functionality of a tiny universe of central systems or companies, one breakdown can create havoc across the entire spectrum.
Make no mistake, he contends — as demand for computer power grows, no company – and no country – can afford to have its progress slowed by a fragile digital foundation.
GPUs, unlike central processing units that handle general tasks, are optimized for parallel processing and thus can perform many calculations simultaneously.
This makes GPUs essential for tasks that require handling large amounts of data – such as gaming, video editing, machine learning, scientific simulations, deep learning, and crypto mining.
Argentum AI is building a global network from the 30% of GPUs that today sit idle at businesses and local network centers. The firm acts as a connector between entities needing additional fast computing power and those otherwise idle GPUs.
It’s a win-win-win scenario.
Argentum is already partnering with thousands of smaller enterprises whose systems would be sitting idle for lengthy stretches.
They market that capacity through their website using a sliding scale pricing system based on each GPU’s unique factors (including cost for electric power) and the amount of computing power needed for the job.
Already Argentum has over 20,000 GPUs at its beck and call without owning any physical infrastructure; Sobko anticipates upping that number to 100,000 by the second quarter of 2026, creating new opportunities for an ever-growing number of users globally.
The focus on giving these still-valuable GPUs a second life also means that companies replacing older units with brand-new technology can still gain revenues that improve their bottom line — while providing much-needed computing capacity for those who cannot afford the latest tech.
Sobko says Argentum is also providing added security layers for clients who heretofore likely had no idea where their data is being processed or how it is being managed.
It's also addressing the fourfold increase in GPU demand just since 2023 - one that has also resulted in waiting times of up to a year just to get access.
One more thing – repurposing these used GPUs is an added hedge against Chinese hegemony.
Duggan Flanakin is a senior policy analyst at the Committee For A Constructive Tomorrow who writes on a wide variety of public policy issues. Read Duggan Flanakin's reports — More Here.
© 2025 Newsmax. All rights reserved.