Cover photo

Hyperscalers May Cede Basic Cloud Business to DePIN

The size and unparalleled growth from the tech giants may have set the tipping point for the wider adoption of DePIN.

We are all familiar with the famous cloud services provided by the U.S. tech giants. The likes of Amazon (AWS), Google (Google Cloud), and Microsoft (Azure) account for nearly 70% of the world’s public cloud service market, far outpacing Chinese competitors or any other providers. It is their size and unparalleled growth, however, which may have set the tipping point for the wider adoption of Decentralized Physical Infrastructure Networks (DePIN).

Background on Data Center Capacity

Nearly half of the world’s data center energy capacity is owned by these tech giants (i.e., hyperscalers). Within that figure, approximately half is U.S.-owned and largely controlled by the Magnificent 7. Simply put, close to 25–30% of the world’s total data center capacity is owned by the Mag 7—a very centralized business phenomenon. These data centers support all major SaaS services provided by the Mag 7, from cloud computing and storage to AI development and services. In a typical data center, only about 40% of the energy goes toward the chips; that power is then divided between CPUs (for general cloud computing use) and GPUs (for AI development and services). The remaining 60% goes to cooling, power distribution, and miscellaneous needs like lighting. A point to note is that GPUs generally have stronger cooling consumption than CPUs, that means if hyperscalers want to increase their GPU composition in a data center, they would have to increase the energy usage on cooling as well.

post image

Growth Pressure on the Mag 7 and Hyperscalers

Now, many of these hyperscalers and the Mag 7 face significant pressure for continuous business growth. Especially with their ever-increasing stock prices, many investors expect them to continue investing heavily in fast-growing businesses like AI. However, limited construction capacity and energy availability may constrain their expansion. Many data center projects are being fully leased or "sold" years before construction is complete or even begins. This "pre-leasing" strategy highlights a severe shortage in the ready supply of data centers. Meanwhile, having an enormous amount of energy draining from one location or region can put massive pressure on local power grids. This means electricity is scarce and priority has to be given for data center services.

post image
Source: Gartner

Because resources are limited, this pushes their focus toward AI development and services instead of traditional cloud businesses—either by replacing existing CPUs with GPUs in current data centers or prioritizing specialized AI centers in future constructions. Although many have integrated AI services into their cloud offerings (AI + Cloud), the high opportunity cost of implementing CPUs in their business model means their general cloud service prices may not be competitive.

post image
Amazon first ever AI purpose built data center for the US government

The Possible Tipping Point for DePIN

So what does this have to do with DePIN, you may ask? While the Mag 7 and hyperscalers are finding it challenging to balance cloud services and AI development, the deteriorating global economy may together set the stage for wider DePIN adoption.

1.      Budget Concerns May Create Room for DePIN

In 2025, cloud services could account for up to 60% of IT budgets for a cloud-based service company. It could also represent nearly 12% of total revenue for a sizable enterprise. Meanwhile, according to statistics from Gartner, nearly 50% of the world's businesses use basic cloud services only without AI.

As we approach the danger of an economic downturn in 2025-26, it is possible that some businesses will consider switching to cheaper cloud alternatives, such as alternative cloud providers or DePIN, in order to save costs. While hyperscalers’ price offerings are hard to beat down, DePIN can deliver a much more reasonable price to accommodate basic company needs. In comparison, AWS cloud services could cost up to seven times as much as a DePIN equivalent.

2.      Cloud Service Outages Are Concerning

With a stronger focus being placed on GPUs instead of CPUs, there is a concern about whether the quality of the hyperscalers' general cloud services might deteriorate. With recent outages happening at AWS and Cloudflare, the consideration of a single point of failure has been put on the table once again. The US Fed in particular, highlights the risk of US financial institutional using highly centralized third-party service providers in their systems in November. In contrast, adoption of blockchain technology by businesses has increased over the past few years. Contributing devices, for instance, have increased from fewer than 10 million in 2020–22 to more than 41.8 million in 2025. The DePIN sector has gradually moved past the early stage of VC investing and is moving toward a real-world, demand-driven ecosystem.

post image
A research study taken by the US Fed regarding risk by centralized third party service providers

3.      Wider AI Market Offerings Make Mag 7’s Cloud Less Favorable

Nowadays, with more and more alternative AI services being offered on the market, it is less favorable for businesses to pay higher prices for the Mag 7’s AI addition. Even if their AI cloud services perform better than the competition, the marginal productivity might no longer justify the additional cost. A practical, more economical tool may well satisfy the needs of simple businesses than an expensive one.

Source:

  1. Future data centers are driving up forecasts for energy demand. States want proof they'll get built - ABC News

  2. https://www.constructionbriefing.com/news/data-centre-construction-failing-to-keep-up-with-demand/8083236.article?zephr_sso_ott=jEuDSC#:~:text=Neil%20Gerrard%20Senior%20Editor%2C%20Construction,%E2%80%9D%2C%20according%20to%20DC%20Byte.

  3. Gartner Predicts Power Shortages Will Restrict 40% of AI Data Centers By 2027

  4. Mapped: Data Center Capacity Around the World

  5. The Latest Cloud Computing Statistics (updated October 2025) | AAG IT Support

  6. https://www.channelfutures.com/cloud/aws-azure-gcp-dominate-global-data-center-capacity#:~:text=Hyperscalers%20Hold%20Almost%20Half%20of,the%20world's%20data%20center%20capacity.

  7. Cutting costs in the cloud: six strategies for SaaS companies | EY - US

  8. DePIN: The $3.5 Trillion Infrastructure Revolution That's Already Started - Boosty Labs | Blockchain Development Company

  9. Data centers powering AI could use more electricity than entire cities

  10. The Fed - Cyber Vulnerabilities at Large US Financial Institutions and Their Third-Party Service Providers