Author: Kava Chinese Official Source: medium

The release of ChatGPT 3.5 in November 2022 marked the official beginning of the modern AI era. In the first article of this series on the next great industrial revolution, we introduced this decisive moment. Since then, the popularity of AI and the enthusiasm for investment have continued to rise. Recently, we explored the massive amounts of money that Wall Street and venture capitalists are investing in this field — in 2024 alone, related investments exceeded $100 billion, accounting for 46.4% of the total $209 billion raised for the year.

The massive scale of investment is triggering a fundamental transformation in the power structure of American technology companies. Cloud service providers, which rose to prominence in the Web2 era through aggressive mergers and acquisitions to consolidate their dominant positions and tightly control user data access, now face the crisis of being disrupted due to their inability to meet the growing demands of AI computing. In stark contrast, decentralized infrastructure, with its unique advantages, not only aligns with this new wave of demand but also offers a more attractive and efficient technological framework.

This article first explores how changes in demand scope and shifts in consumer preferences are reshaping the power distribution landscape of traditional cloud service providers. Next, the article will analyze several projects currently operating in the United States, including Bittensor, Kava AI, Render Network, and Ocean Protocol, to highlight the key projects driving this transformation. After focusing on these transformative projects, the article will conclude with predictions about the future development trends of large technology companies.

Surge in Infrastructure Demand

The strong demand for high-end graphics processing units (GPUs) due to AI has put immense pressure on the supply of traditional cloud service providers, leading to inefficiencies and bottlenecks that hinder AI innovation. Leading cloud service providers — AWS, Azure, and Google Cloud Platform have all admitted that their existing capacity is 2.5 times less than the current demand in the AI market, and this gap is expected to continue to widen in the future.

AI projects can try to build their own physical infrastructure by setting up private on-site data centers to train models. However, this requires a large amount of initial startup capital, as well as additional technical knowledge, and the subsequent ongoing maintenance costs are also very high, making this approach too costly and inefficient for most projects. Therefore, most projects have to rely heavily on cloud services from large technology companies.

The reliance on the infrastructure of large technology companies (and their inability to meet new demands) highlights the drawbacks of concentrating critical infrastructure in the hands of a few tech giants. Growing dissatisfaction with large tech companies, combined with a GPU supply shortage and current U.S. government support for cryptocurrency legislation, has created an opportunity for the emergence of decentralized physical infrastructure network (DePIN) projects in the U.S. These emerging DePIN projects leverage the globally underutilized GPU and AI infrastructure networks to gain a competitive advantage. Compared to traditional cloud service providers, they stand out with dynamic pricing models and significant cost savings.

Consumer Preferences

The monopoly of user data by large tech companies, combined with limited compensation mechanisms, has sparked negative public sentiment. This distrust is further exacerbated by users' concerns — nearly 68% of consumers worldwide express worries about online privacy issues. Many are unwilling to see large tech companies continue to dominate in the AI era and are seeking support for alternative solutions.

The integration of blockchain technology and AI positions decentralized physical infrastructure network (DePIN) projects as an ideal alternative to centralized data control. The immutable ledger utilized by blockchain enables users to have direct ownership of data through encryption privacy techniques; while enhanced AI protocols can introduce higher levels of security measures by detecting anomalies and eliminating single points of failure.

American Projects Driving Change

Fortunately, there are currently some decentralized physical infrastructure networks (DePIN) and decentralized AI (DeAI) projects in the United States that, with the help of AI-enhanced blockchain technology, demonstrate advantages of being more efficient alternatives compared to large tech companies.

Bittensor: An Open Decentralized Machine Learning Market

Bittensor was launched a year before the AI boom swept the mainstream market, making it a pioneer in decentralized AI infrastructure platforms in the United States. The company is headquartered in San Francisco and currently has a market capitalization exceeding $3.6 billion.

The platform operates based on the Bittensor network — — an open market for trading AI models and computing resources that rewards users with TAO tokens based on the quality, utility, and usage of their contributions. Its successful practice demonstrates that decentralized applications can provide lower-cost and more accessible AI and computing infrastructure, thereby challenging the monopoly of large tech companies on AI resource supply.

Kava AI: The Evolution of Decentralized AI

Kava AI fully leverages its accumulated experience in providing decentralized financial infrastructure to successfully create the world's largest decentralized AI model. This transformation into the decentralized AI (DeAI) field allows Kava AI to tackle the key challenge of transparency in the AI decision-making process.

By recording AI decision-making transactions on the blockchain, Kava AI provides unprecedented transparency and verifiability to AI systems. This open AI model development approach starkly contrasts with large centralized models, which continue to collect and utilize user data at extremely low compensation standards.

Render Network: GPU Infrastructure

Render Network is a decentralized GPU infrastructure network designed to meet the computational needs of modern artificial intelligence. The Render project was launched in 2009, the RENDER token was issued in 2017, and the complete network was launched in 2020.

The platform was originally designed to allocate idle GPU resources for virtual reality (VR) and 3D content creation rendering, but soon established its position as a leading supplier of GPU for artificial intelligence protocols by the end of 2022.

Render Network presents a fairer and more democratic model, allowing anyone to profit from their excess GPU resources. In contrast to the entry barriers created by large tech companies, Render offers a clear alternative, as these barriers enable large tech companies to artificially restrict supply, raise prices, and force customers into long-term vendor lock-in contracts.

Ocean Protocol: High-Quality Data Marketplace

Ocean Protocol is an open-source protocol that serves as the foundational layer for decentralized AI systems and data infrastructure. The protocol creates an open market where users can earn income through data. It is built on the Ethereum network and provides a trading market that allows businesses and individuals to sell their datasets with permission.

This novel model allows users to maintain control over their own data, effectively addressing a major concern people have about large technology companies. Furthermore, Ocean Protocol creates economic incentives for data sharing while ensuring privacy and security, enabling smaller AI developers to compete with the large tech companies that have historically monopolized access to large-scale datasets.

The Future of Decentralized AI in the United States

Despite the huge potential of these American decentralized projects in disrupting the existing landscape of large tech companies, we still need to pay attention to the responses of these tech giants. In June 2025, Meta acquired the startup Scale AI, which focuses on high-quality training data and annotation services, indicating that large tech companies are acutely aware of the stakes involved. This acquisition highlights that existing enterprises are actively integrating AI infrastructure to consolidate their industry dominance.

Nevertheless, the integration of decentralized infrastructure and blockchain technology with AI is providing a viable and more democratized alternative. Decentralized AI can achieve more efficient scalability and offer dynamic computational responsiveness that surpasses traditional technologies. For these innovative projects, the United States is providing support through deep capital markets and legislation that supports cryptocurrency.

As American crypto projects continue to mature and demonstrate enterprise-level capabilities, they are expected to capture a significant market share in the rapidly expanding AI infrastructure sector, potentially reshaping the entire AI industry around the principles of decentralization, transparency, and universal access. In the coming years, the construction of AI infrastructure will redefine the direction of next-generation technological innovation.