Currently, the computing and training segments in the AI industry are mainly dominated by centralized Web2 giants. These companies dominate the market with their strong capital strength, state-of-the-art hardware, and vast data resources. While this situation may persist in the development of the most powerful general machine learning (ML) models, the Web3 network may gradually become a more economical and accessible source of computing resources for mid-range or customized models.
Similarly, when inference demands exceed the capabilities of individual edge devices, some consumers may choose the Web3 network for less censorship and more diverse outputs. Instead of trying to completely disrupt the entire AI technology stack, participants in Web3 should focus on these niche scenarios and leverage their unique advantages in anti-censorship, transparency, and social verifiability.
The hardware resources required to train the next generation of foundational models (such as GPT or BERT) are scarce and expensive, with demand for the highest performance chips continuing to exceed supply. This resource scarcity results in hardware being concentrated in the hands of a few well-funded leading companies that utilize this hardware to train and commercialize the most performant and complex foundational models.
However, the pace of hardware updates is extremely fast. So, how will outdated mid-range or low-performance hardware be utilized?
This hardware is likely to be used to train simpler or more targeted models. By matching different categories of models with hardware of varying performance, optimal resource allocation can be achieved. In this case, Web3 protocols can play a key role by coordinating access to diverse, low-cost computing resources. For example, consumers could use simple mid-range models trained on personal datasets, and only choose high-end models trained and hosted by centralized companies when handling more complex tasks, while ensuring that user identities are hidden and prompt data is encrypted.
In addition to efficiency issues, concerns about bias and potential censorship in centralized models are also increasing. The Web3 environment is known for its transparency and verifiability, capable of providing training support for models that have been overlooked or deemed too sensitive by Web2. Although these models may not be competitive in performance and innovation, they still hold significant value for certain groups in society. Therefore, Web3 protocols can carve out a unique market in this area by offering more open, trustworthy, and anti-censorship model training services.
Initially, both centralized and decentralized approaches can coexist, each serving different use cases. However, as Web3 continuously improves developer experience and platform compatibility, and as the network effects of open-source AI gradually become apparent, Web3 may eventually compete in the core areas of centralized companies. Especially as consumers become increasingly aware of the limitations of centralized models, the advantages of Web3 will become more pronounced.
In addition to training mid-range or specialized models, participants in Web3 also have the advantage of providing more transparent and flexible inference solutions. Decentralized inference services can bring various benefits, such as zero downtime, modular combination of models, public model performance evaluations, and more diverse, uncensored outputs. These services can also effectively avoid the 'vendor lock-in' problem that consumers face when relying on a few centralized providers. Similar to model training, the competitive advantage of decentralized inference layers lies not in computing power itself, but in solving long-standing issues such as the transparency of closed-source fine-tuning parameters, lack of verifiability, and high costs.
Dan Olshansky proposed a promising vision of creating more opportunities for AI researchers and engineers through POKT's AI inference routing network, allowing them to put their research into practice and earn additional income through customized machine learning (ML) or artificial intelligence (AI) models. More importantly, this network can promote fairer competition in the inference service market by integrating inference results from various sources, including decentralized and centralized providers.
Although optimistic predictions suggest that the entire AI technology stack may eventually migrate completely on-chain, this goal still faces significant challenges of data and computing resource centralization, as these resources provide existing giants with a substantial competitive advantage. However, decentralized coordination and computing networks demonstrate unique value in providing more personalized, economical, open competitive, and anti-censorship AI services. By focusing on these critical niche markets, Web3 can establish its own competitive barriers, ensuring that the most influential technologies of this era can evolve in multiple directions to benefit a broader range of stakeholders rather than being monopolized by a few traditional giants.
Finally, I would like to extend my special thanks to all members of the Placeholder Investment team, as well as Kyle Samani from Multicoin Capital, Anand Iyer from Canonical VC, Keccak Wong from Nectar AI, Alpin Yukseloglu from Osmosis Labs, and Cameron Dennis from NEAR Foundation, for their reviews and valuable feedback during the writing of this article.

#AI