Opinion by: Michael O’Rourke, founder of the Pocket Network and CEO of Grove
Open data is currently a major contributing towards the development of a global emerging tech economy, with an estimated market of more than $ 350 billion. Open data sources often rely on, however, in centralized infrastructure, contrary to the philosophy of autonomy and censorship resistance.
To realize its potential, open data must switch to decentralized infrastructure. When open data channels start using a decentralized and open infrastructure, many weaknesses for user applications will be resolved.
Open infrastructure has many cases of use, from the pursuit of a decentralized application (DAPP) or a trading bot to sharing research data on training and recognizing large language models (LLM). Looking closely at each other helps us to better understand why the decentralized infrastructure for open data is more utilize than centralized infrastructure.
Affordable Training and Understanding of LLM
The Launching Open-Source AI Deepseek. This is a wake-up call to focus on the new world of the open data economy.
To begin with, the closed-source, centralized AI models have a high cost for training LLMs and developing accurate results.
Not surprisingly, the final phase of training in the Deepseek R1 only costs $ 5.5 million, compared to over $ 100 million for Openai’s GPT-4 GPT-4. However, the emerging AI industry It still depends on centralized infrastructure platforms such as LLM API providers, which are essentially odds with emerging innovative innovations.
Hosting Open-Source LLMs like Llama 2 and Deepseek R1 is simple and inexpensive. Unlike stateful blockchains that require constant continuity, LLMs are pointless and only need to update.
Recently -But: hEre why DeepSeek has crashed your bitcoin and crypto
Despite the simplicity, the computational costs of operating recognition in open-source models are high, as node runners require GPUs. These models can save costs because they do not ask for real-time updates to continue to sync.
Increasing general base models such as GPT-4 enables the development of new products through contextual understanding. Centralized companies such as Openai will not allow any random network support or understanding from their trained model.
Conversely, the decentralized node runner can support the development of open-source LLMs by serving as AI endpoints to provide deterministic data to clients. Decentralized networks have lower entry barriers by empowering operators to launch their gateway to the top of the network.
The decentralized protocols of this infrastructure serve millions of requests with their permissions without open-sourcing the main gateway and service infrastructure. As such, any entrepreneur or operator can deploy their gateway and tap into an emerging market.
For example, a person can train an LLM with decentralized computing resources to the unauthorized protocol Akash, which provides to customize computing services at 85% lower prices than centralized cloud gives.
The AI Training and Inference Market has a huge potential. AI companies have spent approximately $ 1 million days on the infrastructure maintenance to run LLM recognition. It takes services available to the service, or Sam, at about $ 365 million annually.
As data suggests, market conditions indicate a massive potential growth for decentralized infrastructure.
Research data sharing will be accessible
In the domain of scientific and research, the sharing of data combined with the study of the machine and LLMs can accelerate research and improve human life. Access to that data has been walled through the high-cost journal system, choosing to publish research approved by its board and widely inaccessible behind expensive subscriptions.
With increasing blockchain-based zero-knowledge ML models, data can now be shared and calculated without confidence, and privacy can be preserved without revealing sensitive data. Thus, researchers and scientists can share and access research data without de-anonymizing that potentially suppress personal information.
To maintain open research data, researchers need accessing a decentralized infrastructure to reward them for accessing that data, cutting the middleman. An incentive open data network can ensure that the scientific data remains accessible outside the wall with an expensive journal and private corporation wall.
Unstoppable Dapp the DAPP -Host
Centralized data halting platforms -Hosts such as Amazon Web Services, Google Cloud and Microsoft Azure are famous for app developers. Despite their easy access, centralized platforms suffer from a single point of frustration, affecting reliability and leading to rare but may occur.
There are different opportunities in the history of tech when the infrastructura-as-a-service platforms failed to provide incessant services.
For example, in 2022, Metamask temporarily rejected access to users from specific geography regions because Infura blocked them after several US sanctions. Although metamask is decentralized, its default connections and endings depend on centralized tech such as Infura to access Ethereum.
This is not an isolated incident, either. Infura clients also faced a disruption in 2020, while Solana and Polygon experienced excessive loading of centralized remote calls to the procedure (RPC) during the peak traffic.
It is difficult for a company to handle different developer needs in a developed open-source ecosystem. There are thousands of layers 1, rollups, indexes, storage and other middleware protocols with suitable use of the niche.
Most centralized platforms, such as RPC providers, continue to build the same infrastructure, creating friction, slowing down growth metrics, and affecting scalability because protocols focus on rebuilding the foundation rather than adding new features.
In contrast, the massive success of decentralized social network applications such as the Bluesky and the Protocol signal of the search users for decentralized protocols. The transfer of the previous centralized RPC to accessing open data, such protocols remind us of the need to develop and work on decentralized infrastructure.
For example, a decentralized financial protocol may be a source of onchain price data from the chainlink to stop depending on centralized APIs for price feeds and real-time markets.
There are approximately 100 billion serviceable RPC requests in the web3 market, which costs $ 3- $ 6 per million requests. Thus, the total addressable web3 RPC market size is $ 100 million- $ 200 million annually. With the continuous growth of new data existence layers, there may be more than 1 trillion requests on the RPC daily.
It is necessary to pivot towards the decentralized infrastructure to stay in sync with open data transfer and tap into the open-source market.
Open data requires decentralized infrastructure
We will see general blockchain clients who are offloading storage and networking with long -term middleware protocols.
For example, Solana led the decentralization movement when it first began to store its data on chains such as Arweave. No wonder Solana and Phantom are again the main tools for handling Trump’s massive traffic, a major moment in financial and cultural history.
In the future, we will see more data flow through infrastructure protocols, creating dependencies on middleware platforms. As the protocols become more modular and measured, it will create space for open-source, decentralized middleware to be integrated with protocol levels.
It is not possible to have centralized companies that operate as intermediaries for client light headers.
The decentralized infrastructure is unbelievable, distributed, effective and resistant to censorship. As a result, decentralized infrastructure will be a default choice for app developers and companies alike, leading to a mutual -benefit narrative of growth.
Opinion by: Michael O’Rourke, founder of the Pocket Network and CEO of Grove.
This article is for general information purposes and is not intended to be and should not be done as legal or investment advice. The views, attitudes, and opinions expressed here are unique and do not necessarily reflect or represent the views and opinions of the cointelegraph.
