The competition for computing power is no longer just an internal matter of the technology industry.

Written by: Conflux

On May 31, 2026, the U.S. Department of Commerce issued new export control guidelines: the channels for Chinese companies to procure NVIDIA advanced chips through overseas subsidiaries in places like Malaysia were officially closed.

In the same month, the president of Kenya halted a $1 billion geothermal data center involving Microsoft—because once completed, it would consume one-third of the national electricity. President Ruto's exact words were: "It's like shutting down half of the country."

Meanwhile, Huawei announced last week that its Ascend 950PR chip has entered mass production, with an expected AI chip revenue of $12 billion for the year.

Three events, three continents, three completely different pieces of news. But they point to the same emerging reality: the competition for computing power is no longer just an internal matter of the technology industry.

A new oligarchic era is forming

In the past two years, there has been an often-overlooked reality in the AI industry: while it seems flourishing on the surface, the underlying resources are increasingly centralized.

The current AI industry chain can roughly be divided into four layers: GPU chips, cloud computing platforms, foundational models, and application ecosystems. In each layer, control is concentrating among a few players: in the GPU field, NVIDIA has almost become the only choice; in the cloud computing field, AWS, Microsoft Azure, and Google Cloud dominate; in the model layer, OpenAI and Anthropic now control most of the high-end model market.

In other words: the same group of companies is simultaneously controlling chips, cloud platforms, models, and distribution channels. Eric Posner, a law professor at the University of Chicago, refers to this phenomenon as the "AI Octopus," meaning these companies' tentacles cover the entire AI industry chain.

This is different from the platform monopolies of the internet era—internet platforms control traffic, while AI platforms control intelligence itself. This "oligopolistic monopoly" produces profound systemic risks:

• Concentration of control and pricing hegemony: A few companies control the pricing rights, API access, and content review standards of AI. Developers and enterprises face severe risks of "platform lock-in," as giants can change the rules or cut off access at any time.
• Infrastructure vulnerability: Highly centralized computing power can lead to single points of failure ("a single failure impacting the whole system," such as widespread cloud service outages) and imposes unbearable pressure on the electricity grid and energy sources of single regions.
• Geopolitics and computing power hegemony: Computing power is shifting from a neutral infrastructure to a strategic bargaining chip. Due to export control restrictions, countries without independent computing capacity (especially in the Global South) face the risk of being marginalized and a widening technological gap in this wave of technology.

In the future, more and more enterprises will rely on AI to complete development, operations, customer service, marketing, and even decision-making. Once intelligence becomes a production tool, the importance of its control will far exceed that of search engines and social media.

The deepening "AI Iron Curtain"

In the past two years, U.S. operations regarding chip export controls have become increasingly fragmented. During Biden's term, there was an "AI diffusion rule," categorizing global cooperation into three levels; this rule was revoked after Trump took office, shifting to case-by-case approvals and temporary bans. In the face of this iron curtain, countries' responses have varied widely.

Saudi Arabia has directly designated 2026 as the "Year of Artificial Intelligence": through the sovereign fund’s HUMAIN company, Saudi Arabia invested $3 billion in Musk's xAI, with one condition being the establishment of AI data centers over 500 megawatts in Saudi Arabia; the UAE is building a 5 gigawatt AI park in Abu Dhabi, claimed to be the largest globally outside the U.S.—the first phase goes online this year; in May, the UAE received its first batch of the latest NVIDIA chips exported from the U.S.

The logic of the Gulf countries is straightforward: the previous era relied on selling oil, this era relies on buying computing power.

The EU's anxiety stems from another direction: official data shows that over 80% of digital services in Europe run on non-EU infrastructure. The ongoing proposal for the "Cloud Computing and Artificial Intelligence Development Act" (CADA) aims to triple Europe's computing power by 2030. France's Mistral released a strategic document this April titled "European AI: A Playbook to Own It."

However, the most difficult position is held by economies that are not even eligible to compete: Kenya's $1 billion data center was halted; Malaysia allocated about $490 million to build a sovereign AI cloud. India is subsidizing researchers' GPU usage fees; Indonesia is preparing a domestic large model—these investments are not small within their respective economies.

However, this year alone, the AI capital expenditures of Microsoft, Google, Amazon, and Meta total around $750 billion. This magnitude of difference is itself part of the problem.

The end of the competition for computing power is increasingly pointing to a more fundamental variable: electricity. An AI inference task can consume up to 1000 times the electricity of a traditional web search. To address the anticipated global data center energy consumption reaching 1050 terawatt-hours by 2026, tech companies have even started directly purchasing nuclear power plants.

Is there a possibility of "not taking sides"?

Against this backdrop, decentralized AI (DeAI) has begun to attract attention. It attempts to answer a question: besides handing the future over to a few tech giants or a few countries, is there a third possibility?

If the internet can connect the global network through open protocols, can AI also connect global computing power through an open network? Can idle GPUs, independent developers, research institutions, and enterprise data centers around the world form an open AI infrastructure network?

The core idea of DeAI is not complicated: coordinating independent participants through open protocols to achieve an AI system without a single power center in control. Combined with blockchain technology, cryptoeconomic incentives, and cryptographic verification mechanisms, it addresses the trust issues in anonymous networks, directly responding to the pain points of centralized AI:

• Breaking market concentration: Establishing a distributed network of computing power, data, and model providers to form a free competitive market pricing mechanism.
• Alleviating physical limitations: Dispersing the massive energy demands into power grids around the world.
• Getting rid of geopolitical dependencies: Constructing an infrastructure layer that transcends a single jurisdiction, providing possibilities for "sovereign AI."
• Increasing verification transparency: Replacing blind trust in the reputations of tech giants with provable technological means.

Supporters believe this model could reduce dependence on a single supplier, enhance system resilience, and provide opportunities for small and medium-sized countries and businesses to participate.

Meanwhile, the attitude of institutional investors is shifting from curiosity to substantial investment. Venture capital firms (such as DCG, a16z, etc.) are injecting hundreds of millions of dollars into DeAI protocols; traditional businesses (such as Deutsche Telekom) are starting to participate as verifiers in the network; not only that, but some national governments (like Kazakhstan) are also exploring integrating their idle national supercomputing resources into the decentralized computing market.

Conclusion

As stated in the "State of DeAI 2026" report, the core value proposition of DeAI is not that it can immediately outperform centralized systems in terms of performance today, but rather that it provides a fundamental framework to resist monopolies, reject censorship, and decentralize power.

With the declining costs of dedicated AI hardware (ASICs) and the continued prosperity of open-source models, the time window for DeAI to solve operational challenges has begun to open. The work to establish the DeAI foundation has only just begun.

Of course, DeAI still has a long way to go before becoming mainstream. Whether in terms of performance, stability, or business models, it remains in its early stages. But perhaps its significant importance lies not in immediately challenging OpenAI, but in providing an alternative solution.

Historical experience tells us: when an industry has only one option, the issue is often not whether power will be abused, but rather when it will be abused.

The very existence of competition is a form of checks and balances.