IA could optimize energy when Bitcoin and the selection of templates in the Mempool.
In more programmable networks, such as bittensor, IA have more types of application.
Artificial intelligence (AI) is beginning to transform the Bitcoin ecosystem, although not by changes in its protocol. Rather, through innovative applications that affect work modes in their periphery.
In an exclusive interview for cryptootics, the Engineer José Ríoswho has experience in chips design and was Vice President of Intel, shared his vision on how AI can optimize key processes in the Bitcoin network, with a Special approach to energy savings.
Ríos, currently leader of the Startup BT Labs offered a technical and pragmatic perspective on this technological convergence between Bitcoin and AI, which is also redefining other ecosystems aimed at decentralization, such as bittensor.
AI serves for bitcoin, but “from the outside”
The Bitcoin protocol, based on the SHA-256 algorithm, is immutable by design, which guarantees its safety, but limits its ability to integrate solutions directly. It lacks the programability to do so.
However, José Ríos emphasizes that AI agents – which he defines as “a list of actions that the software executes for you, as a recipe: grabs this, click there, seek this, move it from left to right” – can revolutionize processes external to the protocol, such as the management of mining templates in the mempool and, above all, the energy efficiency in mining.
The most significant advance, according to the interviewee, is The use of predictive to reduce energy consumption in Bitcoin mininga sector frequently questioned about its environmental impact, sometimes unfairly and without data at hand.
As an advisor to Quantum Blockchain Technologies (QBT), a company based in London, Ríos detailed a project led by Italian scientists who use AI algorithms to analyze patterns in the Bitcoin and predict protocol What hashes are more likely to achieve a new block with its respective subsidy in BTC.
As Cryptonoticias reported, the miner who finds the number that coincides with the objective of difficulty proposed by the Bitcoin protocol, gains the right to write transactions in a block, and the right to claim the coinbase or block subsidy. This process, known as function hashIt is preconceived as random, similar to a lottery, and depends on the energy power of Bitcoin ASIC systems.
“They [Quantum Blockchain Technologies] They use AI to study all the hashings that have occurred before, which are publicly in the blockchain, and determine the probability that the next hash will be about 50% likely, ”explained José Ríos.
This approach is unchecked from the traditional idea, and proposes that the hashing process is pseudo -aliacio, which allows to filter unnecessary calculations in the “Pipeline”From Sha-256.
“It’s like breaking a password: in the 10 minutes you have to find the right hash, you are searching Brute Force in a billion space. The AI intercepts some of these requests before the machine does the physical work, saving energy and giving better results, ”said the engineer.
This approach challenges belief that the search for Bitcoin blocks It is purely randomdemonstrating that historical numerical patterns can guide more efficient decisions.
“These scientists have proven that any process in the universe is truly random. There is always an origin. It is pseudo -alleuate, and with simulations they can calculate step by step how to optimize the process of Bitcoin mining,” said the interviewee, underlining the scientific rigor behind this innovation.
The preliminary results, indicates rivers, show that This technique could reduce electrical consumption by up to 30%a savings that would translate millionaire benefits for industrial miners.
The project, still in development, He has shown advances In events such as the Bitcoin Conference 2025, happened in Las Vegas, and Ríos estimates that it could be ready for commercial implementation at the end of 2025 or beginning of 2026.
Other AI applications on the periphery of Bitcoin
In addition to energy savings, José Ríos confirmed the idea that AI agents could decentralize the management of mining templates, reducing the dependence of the great pools that concentrate most of the hashrat.
“The danger with the Pools is that they are in the hands of seven or eight companies. They could be coordinated to control more than 51% of the hashrate and manipulate the network, but that would require a perfect multibillionaire crime that would damage confidence in Bitcoin. It is not feasible because you would be damaging the system you benefit,” he argued.
AI agents, allowing users to manage templates autonomously, could reinforce decentralization, making the most objective and less dependent process on human judgment.
Other uses of AI in Bitcoin covers Data analysis for trading and fraud detection.
According to Ríos, AI agents can process large volumes of data, especially transactions, in the Bitcoin network to identify money laundering patterns or feeling analysis on social networks such as Reddit or X. “You can put all the comments of the networks in an AI blender and take out a sentimental analysislike the fear and greed index, to see if people are scared or positive about the price, ”said the engineer.
These applications, although external to the protocol, show how AI can work around Bitcoin and produce utility at technological and market levels without altering its fundamental structure.
AI and programmable networks: the case of bittensor
Beyond Bitcoin, José Ríos, as leader of BT Labs, is exploring the integration of AI with more programmable protocols, such as Bittensor, which occupies much of his current work.
Inspired by the Bitcoin principles – a limited supply of 21 million currencies, one halving every four years and a work test mechanism – Bittensor uses its network and protocol to connect distributed computational resources and promote global collaboration in the development of AI.
“Bittensor is an open system where engineers and scientists can contribute to GPUS, algorithms or data, and be rewarded with tokens. You do not need to be in Silicon Valley or work for OpenAi. You can be in the Dominican Republic, Rio de Janeiro or Vietnam, and monetize your skills,” Ríos explained.
This model contrasts with the centralization in the development of AI that currently exists, dominated by five major American companies – Openai, Amazon, Microsoft, Google and Meta— that control 90% of the global budget in this field. “That is an alarm. There is a strong bias, and they can censor points of view or innovate only when it suits them,” the interviewee warned.
Bittensor, then, in the opinion of Ríos, embodies a role similar to that of Bitcoin, but not within the monetary system, but for the distribution of the computational power that feeds the AI. This promotes a democratization of access to AI, allowing communities in Latin America, Africa or Southeast Asia to participate actively.
In addition, he stressed that The protocol already supports 116 applicationsfrom detection tools for Deepfakes even inference systems that exceed centralized solutions such as Perplexityoffering responses to 400 tokens per second compared to 200-250 of its competitors. “It is a distributed system, with computers everywhere, provided by people like you and I. That makes it faster and more efficient,” he said.
The engineer, who has presented these ideas in international conferences and before legislators in Washington, advocates open systems as a solution to the risks of centralization. “Open systems win because they are like the Internet or Linux: nobody controls them, everyone can collaborate. Innovation does not occur in closed systems,” said José Ríos.
A decentralized future for technology
The convergence between AI and the “blockchain”, according to José Ríos, transcends the technical to address social and economic issues. Projects such as Bittensor not only solve computational problems, but also make marginalized communities of technological development.
“In our countries [latinoamericanos]the level of intelligence is the same or better, but we do not have the resources. With bittensor, an engineer can work from anywhere and be rewarded without asking permission, ”he said.
For Bitcoin, AI is opening New possibilities in optimization of peripheral processeswith energy savings as an especially promising advance.
In networks such as bittensor, decentralization extends to the development of AI, creating a model where innovation has a more global reach.
Ríos concludes with an optimistic vision: “Open technology always finds its path. Bitcoin demonstrated it, and now the decentralized AI is opening a new chapter so that everyone, not only a few, we are part of this revolution.”
