Amazon Web Services (AWS) and Microsoft have brought themselves back from AI Data Center investments, which suggests that problems with the centralized model. Analysts take this latest development to repeat why decentralized blockchain-based infrastructure could be the solution.
Kai Wawrzinek, co-founder of Impossible Cloud Network, discussed these imminent questions in an exclusive interview with Beincrypto.
AI -Datacenders touch a wall
A few months ago, AI seemed one of the most promising sectors of the worldwide technical industry. However, with companies such as AWS and Microsoft that announce breaks in the construction of AI Datacenter, the photo looks very different. What happened? What does the future of AI look like? Kai Wawrzinek described the situation as it looks nowadays:
“News that AWS has with Microsoft with him when the new data centers are taking off when the demand for AI is growing exponentially, bears witness to the enormous inefficiency that this model presents for scaling the global internet. Microsoft and AWS can realize that centralized infrastructure models.
AWS and Microsoft are not the only companies with which these problems are confronted. Although Meta publicly claimed that it would spend hundreds of billions on AI infrastructure and data centers, the competitors asked for financing less than three months later.
OpenAi is also startled by the enormous costs of operating chatgpt; Sam Altman tacitly admitted that his research may never be profitable.
Wawrzinek sees a clear solution – leave the centralized model completely and focus on Defai. Although these market leaders have collected billions in Capex and developed LLM development, the entire strategy can be self-destructive.
For example, the construction of the US AI data center is the electric engineers of the work with an unprecedented degree flooded with work. With so many professionals who concentrate on the centers themselves, it creates a bottleneck for trained work.
This damages renewable energy projects and the electric grid, which ironically harms the functionality of the data centers.
“The AI era needs infrastructure that can correspond to its speed and scale, and decentralized systems are the only models that are built for that future. In contrast to a decentralized, market-controlled approach, this problem solves: capacity can be used more efficiently where and when it is necessary for years for centralized MegaWzin.
Can Defai take on the challenges?
In comparison with the centralized data center model, Defai has increased the accessibility of AI Compute. Blockchain-compatible economic stimuli can speed up the implementation rate, improve scalability and optimize the allocation of resources without massive capital.
In short, these decentralized systems have more agility than their competitors.
Blockchain-based AI companies have been able to use a considerable calculation capacity without centralized data centers. For example, the Depin company Aethir has taken major steps with its GPU-AS-A-SERVICE model.
Other companies such as 0G Labs have proven that decentralized AI development is not only theoretically feasible; It is profitable and necessary for the ecosystem.
If all this seems far -fetched or utopian, it is important to remember the “Black Swan” event of AI.
The Chinese market-moving Genai model proved the whole world that AI companies can make state-of-the-art LLMs against a fraction of hardware costs. So the AI industry should possibly fully reconsider the data center model if this one developer has proved to be so successful.
Although skeptics have wondered whether decentralized AI can compete with data centers, the reality is that centralization can have its own inefficiencies.
“The future of the AI infrastructure lies in open, permissionless networks, where the supply meets dynamically and worldwide, not due to outdated hyperscaler models that have difficulty keeping up,” ended Wawrinzinek.
Up to now, centralized AI companies have collected billions in investments in risk capital, but their ability to innovate is a brick wall. We may need a better model to create the best possible results.