Over the past year, AI has remained one of the most significant themes in global capital markets. From Nvidia’s record-breaking earnings reports to Microsoft, Meta, and Amazon ramping up data center investments, AI has evolved from a mere technology trend into a new cycle of infrastructure development. This shift is now starting to influence the crypto market as well.
Unlike the 2024 market, which focused on AI Agent concepts and AI Meme assets, since 2026, capital has been moving deeper into the industry value chain. More investors are now addressing a practical question: As the number of AI applications grows and user bases expand, who will truly benefit from the long-term growth of the AI industry?
The answer is shifting from the model layer to the infrastructure layer.
Whether it’s AI Agents, AI video generation, AI music creation, or enterprise-level AI services, all rely heavily on GPU resources. As the sector moves from the era of model training to inference, demand for computing power is steadily rising. At the same time, GPU supply remains tight, resource acquisition costs are high, and cloud service prices are climbing—issues that are becoming increasingly prominent. Against this backdrop, decentralized GPU networks are once again drawing attention.
As a key project in the AI infrastructure space, IO has recently released a series of commercialization cases, further strengthening its position as an AI computing platform. In terms of price performance, IO has surged nearly 200% from its April lows. At the industry level, the market is reassessing the long-term value of DePIN computing networks within the AI value chain. So, is IO’s continued price rally merely a temporary rebound for the AI sector, or does it signal that decentralized GPU networks are entering a new growth cycle?
Behind IO’s Price Rally: Shifting Market Focus
Looking at price trends, IO’s current rally is more than just a technical rebound.
Gate market data and recent price charts show that IO dropped to around $0.09 in early April, then remained in a low, sideways range for several weeks. Throughout April, trading volumes were subdued and price volatility was limited, indicating that the market was still cautious about the AI infrastructure sector.
The real shift came in early May.
With trading volumes picking up, IO broke out of its previous consolidation range and quickly crossed the $0.15 mark. Although there was a subsequent pullback, the price stayed above the breakout platform, and trading volumes shrank in tandem. This pattern suggests a rotation of holders during the rally, rather than a withdrawal of capital. From late May into early June, IO saw another surge in volume and price, peaking near $0.27—a nearly 200% gain from its April lows.
What’s even more notable is that IO wasn’t the only project to rebound during this period. Other AI infrastructure projects like Render, Aethir, and Akash also attracted market attention. This indicates that capital is trading not just a single project, but the broader logic of AI infrastructure.
The shift in market focus is clear. Previously, investors were more willing to pay for "AI concepts." Now, more capital is considering which segments of the AI value chain will see sustained demand. As the market moves from narrative-driven speculation to fundamentals, projects with real-world use cases and commercialization stories naturally attract more attention.
From Model Competition to Inference Competition: AI Industry Enters a New Growth Phase
Looking back over the past three years of AI development, it’s clear that the competitive landscape has changed.
In 2023 and 2024, the spotlight was on model capabilities. Who had the largest parameter scale, who performed best in benchmarks, and who had the strongest training capacity—these were the focal points. During that phase, capital poured into model development companies, and GPU resources were mainly seen as production assets for training.
But as models matured, the industry entered a new phase.
More companies realized that while training is expensive, it isn’t the largest long-term cost. The real budget drain is inference. Training typically occurs during model development and upgrades, but inference happens every time a user interacts with an AI product. For AI applications with millions or even tens of millions of users, inference costs often far exceed training expenses.
This is why more tech firms are now emphasizing inference optimization.
For enterprises, reducing inference costs not only saves money, but also enables them to serve more users, improve profit margins, and expand market share. Over the next few years, competition in AI may shift from "who has the strongest model" to "who can deliver AI services at the lowest cost."
GPU resources become even more critical in this context.
The market is now focusing on GPU acquisition costs, resource utilization, and computing power scheduling efficiency. Compared to the model layer, which sees constant new entrants, GPU networks and computing platforms have more stable demand. No matter which model company ultimately wins, they all need to consume vast amounts of computing power. This is why AI infrastructure projects continue to garner market interest.
Why IO’s Recent Commercial Cases Are Attracting Attention
While industry trends set the direction for capital, project-level commercialization progress determines whether the market is willing to assign higher valuations.
One of the biggest doubts facing the DePIN sector over the past year has been the lack of real demand. Many projects can quickly aggregate resources through token incentives, but struggle to prove that these resources are being used by real businesses. As a result, the market has remained cautious about DePIN projects.
IO’s recent disclosures directly address this issue.
The most notable example is Leonardo.AI. According to official data, Leonardo.AI expanded from about 14,000 users to 19 million. During this process, it leveraged IO’s network for GPU resources, cutting overall GPU costs by more than 50% and reducing procurement cycles from weeks or months to just days.
For the market, this case is significant beyond just cost savings. It demonstrates that decentralized GPU networks are now serving real AI platforms with tens of millions of users—not just operating in test environments.
Another high-profile case comes from AI music platform Wondera. Official figures show Wondera used over 550,000 GPU hours for model training and employed 96 high-end GPUs for related tasks. Compared to traditional cloud solutions, Wondera reduced training costs by about 75%, saving approximately $2.48 million.
Together, these cases send a clear signal: decentralized GPU networks are moving from proof-of-concept to commercial validation. As the market sees real businesses using network resources to cut costs and scale operations, the logic behind project valuation naturally evolves.
Why AI Companies Are Seeking Computing Power Beyond Traditional Cloud Services
Decentralized GPU networks are attracting attention not just because of project development, but due to broader industry challenges.
The main reason lies in the real issues facing the AI sector.
In recent years, major cloud providers have almost monopolized the high-performance GPU market. But as AI demand has exploded, traditional cloud models have exposed more problems. On one hand, GPU resources are chronically undersupplied, leaving many companies unable to secure needed resources even with sufficient budgets. On the other, rising cloud service costs are eroding AI company profit margins.
For many small and mid-sized AI firms, this pressure is especially acute.
They lack the ability to pre-secure large GPU allocations like tech giants, and can’t afford ongoing high cloud costs. As a result, the industry is broadly seeking more flexible, lower-cost sources of computing power.
This is where decentralized GPU networks find their opportunity.
By aggregating idle GPU resources worldwide, decentralized networks offer businesses a more flexible way to access resources. When demand spikes, companies can scale up quickly; when demand falls, they avoid long-term resource costs. From an industry perspective, this model resembles an open market rather than a traditional centralized allocation system.
As the number of AI applications continues to grow, the importance of elastic resource scheduling will only increase.
DePIN Computing Networks Are Entering Real Demand Validation
From an industry development standpoint, AI is likely a key catalyst for DePIN commercialization.
Over the past few years, DePIN projects have focused on solving supply issues—using incentives to attract devices and build a global resource market. But the real challenge isn’t supply expansion, it’s demand validation. Without real customers and sustained workloads, even the largest network can’t create lasting value.
AI is now matching supply and demand for the first time.
Previously, GPU networks lacked customers; now, AI companies lack GPUs. Data networks once lacked use cases; now, AI models need more and more data. Open computing networks previously lacked applications; now, AI inference demand is surging.
This shift means DePIN is no longer just about aggregating resources—it now has real industry demand as its foundation.
Recent market performance shows investors are reassessing the sector. Instead of focusing solely on node counts and device scale, the market is looking at enterprise clients, network utilization, and real revenue sources. In other words, DePIN is moving from "storytelling" to "demand validation," with AI as a key driver of this transition.
Why AI Infrastructure Is the Key Beneficiary of This AI Market Cycle
If you look at the current AI sector, you’ll see capital gradually spreading from the application layer to infrastructure.
The reason is straightforward. Competition among applications is highly uncertain, but infrastructure demand is much more predictable. Regardless of which AI company wins market share or which Agent platform becomes mainstream, all will need ongoing computing power, data, and network resources.
This demand won’t disappear even if competition at the application layer ends.
Therefore, instead of betting on a single AI product, more investors are focusing on infrastructure. For the market, the real scarcity in AI isn’t tokens, but the foundational resources that keep the ecosystem running. As the industry enters the inference era, GPU resources will only grow in importance, and related infrastructure projects are likely to benefit from this trend.
From this perspective, IO’s price surge reflects not just recognition of the project itself, but a broader revaluation of AI infrastructure’s long-term value.
Key Metrics to Watch for IO’s Future Price Performance
For IO, factors determining its long-term value go beyond market sentiment.
What truly matters are enterprise client growth, network GPU utilization, real workload scale, and commercialization revenue data. If IO continues to disclose more cases like Leonardo.AI and Wondera, and further demonstrates network operations, the market may increasingly use infrastructure valuation models to assess its worth.
At the same time, the overall pace of AI industry development is crucial. If inference demand keeps rising and enterprises continue to seek more GPU resources, the market space for decentralized GPU networks could expand further.
Thus, IO’s core future logic isn’t just about the AI concept—it’s about whether it can consistently meet real demand driven by AI industry expansion.
Conclusion
IO’s sustained price rally over the past two months is not simply the result of renewed sentiment in the AI sector. It reflects a fundamental shift in how the market values AI infrastructure. As the AI industry moves from the training era into the inference era, GPU resources are becoming increasingly critical, and enterprise demand for low-cost, flexible computing power is rising rapidly.
From Leonardo.AI to Wondera, IO’s recent commercial cases show that decentralized GPU networks are entering real business scenarios. This means the DePIN computing sector is moving from resource aggregation to demand validation. If the AI industry continues to expand, IO’s decentralized computing network could become a vital part of the AI era’s infrastructure ecosystem.
FAQ
Why has IO’s price continued to rally recently?
IO’s price rally has been driven by renewed momentum in the AI infrastructure sector, an increase in commercial use cases, and the market’s renewed focus on GPU network demand.
What are the most noteworthy project developments for IO lately?
The most notable developments for IO are the commercialization cases with Leonardo.AI and Wondera. Leonardo.AI grew its user base to 19 million while cutting GPU costs by over 50%. Wondera saved about $2.48 million in training costs using IO’s network. These cases further validate the commercial viability of decentralized GPU networks.
Why are decentralized GPU networks attracting attention in the AI industry?
Decentralized GPU networks are gaining attention because they integrate globally distributed GPU resources and offer enterprises more flexible, lower-cost computing services. As AI inference demand grows rapidly, this model can help alleviate the cost and resource pressures faced by traditional cloud services.
Why is DePIN becoming a key beneficiary in the AI industry?
DePIN is benefiting because AI companies’ demand for GPU, data, and computing resources is steadily increasing, bringing real external demand to DePIN networks.
What factors should be watched for IO’s future price performance?
Key factors affecting IO’s long-term price include enterprise client growth, GPU utilization, network revenue scale, and the growth of AI inference demand. If IO can continue to expand real workloads and enhance commercialization, its long-term value may see further validation.
