In this article:
1. The Infrastructure Reality: Complexity and Cost
2. Trust, but verify
3. The Strategy of Specialization
4. Innovations at the intersection of Web3 and AI
5. Can Web3 projects change the market?
While tech giants are competing through massive investments in artificial intelligence (AI) infrastructure, Web3 companies are developing an alternative, decentralized model.
According to Dell’Oro Group, Big Tech players — Amazon, Google, Meta, and Microsoft — invested around $180 billion in AI infrastructure and data centers in 2024, and in 2025 their spending is expected to reach $320 billion.
The Incrypted team spoke with her, as well as with Aethir CEO Dan Wang and Messari Protocol Partnerships Lead Brandon Slack at Agents Day during EthCC, about how Web3 companies are building the future of AI and what challenges they face along the way.
Building AI systems is far more expensive and complex than it seems, Tsao noted. According to her, today’s systems require multiple components: a data layer, compute, GPUs, and AI agents.
As soon as a project starts to scale, initial costs skyrocket, he says. That’s why most AI agents never reach the mass market — they simply don’t have the right monetization model, the Aethir CEO pointed out.
As already mentioned, Big Tech companies are prepared to spend hundreds of billions of dollars on AI infrastructure. Moreover, they have other advantages over players in the Web3 segment.
It is worth noting that the decentralized physical infrastructure networks (DePIN) sector, where Filecoin is one of the pioneers, has already reached a market cap of more than $40 billion, according to Messari. According to analysts at the World Economic Forum, this figure could rise to $3.5 trillion by 2028.
In addition to cost, a critically important factor in working with AI and data is their accuracy and quality.
This is where the advantage of a decentralized approach can come into play. In traditional cloud systems, users often do not have access to a verified source of information or cannot verify the authenticity and immutability of data.
This becomes critically important in the era of deepfakes and fabricated information. Today, when there is too much data on the internet, people often take everything at face value — and that is why verification becomes a necessity.
The ability to prove the origin of data is a fundamental advantage of Web3: blockchain allows not only to store information, but also to ensure its authenticity.
Another advantage of decentralization is the focus on narrow specialization, whereas large technology companies try to solve “everything for everyone.” Slack draws an analogy with medicine — generalized AI works like therapists, but there are also narrow-profile specialists. In his opinion, AI development will follow this path:
New technologies are particularly active in the world of finance.
AI agents on the blockchain have already reached impressive scales. According to VanEck estimates, the potential of this segment will lead to the emergence of more than a million new agents by 2025.
Data storage remains a key aspect of this process. Tsao explains:
The AI revolution is just beginning. Web3 companies are betting on decentralization and specialization, contrasting this with centralized monopolies. In addition, blockchain provides transparency of data origin.
History shows that in the battle between Davids and Goliaths, it is often the fast and focused newcomers who win, finding moves that the big players don’t even think of.
Disclaimer: The articles reposted on this site are sourced from public platforms and are provided for informational purposes only. They do not necessarily reflect the views of MEXC. All rights remain with the original authors. If you believe any content infringes on third-party rights, please contact
[email protected] for removal. MEXC makes no guarantees regarding the accuracy, completeness, or timeliness of the content and is not responsible for any actions taken based on the information provided. The content does not constitute financial, legal, or other professional advice, nor should it be considered a recommendation or endorsement by MEXC.
