Qtum Foundation powers AI Web3 initiatives with 10,000 Nvidia GPUs

Images generated by Qtum Qurator.

Image Credit: Qtum Foundation

Qtum Foundation is bringing online 10,000 GPUs to power its new blockchain AI ecosystem.

If that sounds like an achievement, then you’re right about that because demand for graphics processing units (GPUs) has gone through the roof and made GPU chip designer Nvidia into a $2.1 trillion company.

It’s not easy to get your hands on modern GPUs, thanks to demand for AI chips as well as gaming and cryptocurrency mining.

It might sound like a scam to some, but having the GPUs brings cred. Qtum Foundation has these 3000 Series Nvidia GPUs because it converted itself from a cryptocurrency mining operator to an AI chatbot and image operator, said Miguel Palencia, COO and cofounder of Qtum, in an interview with GamesBeat.

Of course, the challenge isn’t just about getting your hands on GPUs. It’s about knowing what to do with them. It won’t be easy to compete with the likes of OpenAI when it comes to free AI applications.

Qtum Foundation’s GPUs will power Quantum AI initiatives that include Qtum Solstice, a conversational chatbot akin to ChatGPT, based on open-source models. This alpha deployment will engage users in helpful and intelligent dialogue.

It is also working on Qtum Qurator, a text-to-image generation model similar to Midjourney, also based on open-source models. This utility will allow users to generate images that normally require tedious work in image manipulation software applications, Palencia said. Later on, video generators will come. In the long run, Qtum wants to generate an AI assistant, Palencia said.

Qtum is a blockchain platform that combines Bitcoin’s security with Ethereum’s smart contract abilities. It focuses on decentralized governance, speed, and energy efficiency. Its governance model lets the community adjust important features like gas costs and block size through smart contracts.

As for the capabilities now, Palencia said, “Right now, you can talk to it, you can create many different images from several different open-source models that are already in existence. We’re working on a new model that uses video capabilities. And this is all being powered by our own GPUs.”

Qtum’s network can handle around 1,100 transactions per second, with the potential to process millions to billions of transactions per second by using layer-1 and layer-2 solutions. Qtum employs a proof-of-stake system, which lets people stake with low-powered devices or even offline. All of this tech used to power mining operations.

AI as the next step

For Qtum, artificial intelligence is the logical next step in light of its foray into Web3.

“We have ambitious goals for AI, which will require a 3-stage rollout plan,” Palencia said. “Today, we are unveiling stage 1, which is the chatbot and image generator. The next two stages will include the modeling layer and finally the decentralized economy layer, which will blend everything together on the Qtum blockchain.”

Immediately following the first stage of deployment of Solace and Qurator, the Qtum Foundation will begin work to improve the underlying technology of the product offerings. Qtum developers have followed AI closely since even before the debut of ChatGBT, and have acquired technical expertise in four areas.

These include intelligent dialogue, which includes knowledge base question-and-answer systems, text-based games, etc. It also includes speech generation, with text-to-speech, speech-to-text, cross-language conversations, and other emerging fields.

It also includes image generation, with text-to-image conversion, image super-resolution, 3D cartoon image generation, AI-based face swapping, facial/expression/pose detection, and more.

And it includes video generation, which encompasses video re-rendering/face swapping, animated illustrations, and other resource demanding video manipulation technologies.

Based on these technologies, Qtum developers can develop and offer multiple AI features, such as keyword-based image/video generation. This enables users to conveniently create image and video assets based on keywords.

Developers can also create emotional AI voice-based conversational bots. These are AI bots that understand users’ intentions and can address their questions through voice or text. This “Natural Language Understanding” technology allows the AI bot to detect tone and respond with emotion.

They can also generate intelligent image recognition. By utilizing computer vision techniques, this feature helps users identify image content. Especially useful for categorizing large amounts of images or searching for specific traits in existing graphics.

Developers can also build AI-based filters and photo effects. Users can experience various makeup, photo filters, and other effects. A more powerful experience can be obtained with AI compared to static offerings found on social media platforms due to the fact that AI can generate whatever you ask.

Leveraging the advantages of Qtum’s decentralized nodes, it is easy to open-source AI models and harness the power of the community to develop more interesting and practical AI products.

The Qtum Foundation plans to offer up to 10 more AI-related products in the coming months to further utilize the 10,000 GPU’s. This abundance of computing power will aid the Qtum Foundation in seeking partners for collaboration and joint development.

Basic access to Qtum AI services will be free, but premium features like renting large portions of computing power or blockchain-based intellectual property protection with Qtum tokens is planned.

Big numbers

Based on recent data from 2023 and according to Grand View Research’s most recent report, the global market value of AI is $196.63 billion, which represents an increase of around $60 billion from the previous year. This rise can be attributed to the growing number of practical applications for AI technology.

Predictions state that the global AI market will be worth $1.81 trillion by 2030, and as many as 97 million people could be working in the AI industry by 2025.

Qtum is entering the AI field to pursue cutting-edge technology that will drive industries into the future. Qtum’s AI will be closely linked to the QTUM token, making it one of the ecosystem’s most prominent features. The QTUM token has a market value of $380.6 million.

As for people who are suspicious of scams, Palencia said this has persisted for as long as he has been in crypto. He’s aware of pump-and-dump schemes and he warns people to stay away from them and pay attention to serious projects like the biggest blockchains.

Origins

Qtum started out as a Bitcoin mining company around 2017. That’s how it started collecting GPUs and now it has 3090 RTX Series GPUs, an older generation of Nvidia gear.

“It’s wild trying to get some of the new ones,” Palencia said.

The Qtum Foundation has a history of delivering products that incorporate solutions. For example, in 2017, the Qtum Blockchain featured an Ethereum Virtual Machine-capable smart contract platform with native UTXO support.

It also featured staking and the ability to adjust gas fees and block sizes to enable high throughput without the need for a hard fork. None of these features sacrifice decentralization or leave behind a heavy carbon footprint.

“I’ve been a fan of technology since I was a little kid,” he said. “I used to live in Shanghai about seven years ago. And there were quite a few AI solutions that were being built internally in China for different purposes, like mostly governmental applications. But you could see there was quite quite an interest in machine learning, and it was just getting stronger over time. The first few projects that were viable, that were economically feasible, that were actually sustainable,” got him excited.

He learned how to do mining and started buying his own GPUs. He started building mining pools with collections of servers in the 2013 to 2016 time period. He served as the CTO for VeChain, a blockchain started in 2015, as well as the project architect for the Qtum Foundation. Last year, the cofounders shifted the focus of Qtum to pivot to AI.

“Our GPUs weren’t being used anymore for mining so we decided that we could use this for something new, something better. And we we came up with the idea last year about building this new AI solution. And if we have the resources, why not? Why not train yourselves?” he said. “Most of the others in this space have to rent their GPUs. That increases their costs quite a lot.”

Looking at image generators, Palencia was amazed.

“It was the start, and it was a very powerful start,” he said. “The reaction of people to it told you it was a really disruptive technology. It created fear and excitement. It’s exciting to see what can be done, and the possibilities that can happen in the future. But at the same time, of course, our fear is many jobs will disappear.”

To get ahead of that, Palencia decided to get smart and move fast. The company is making its chatbots and other AI products available for free “to get people to fall in love with it.” Over time, the company expects to charge a subscription fee.

To deal with the fear, Qtum decided to lean on the community, not only for funding but also guidance on how to pursue AI. Palencia wants to upgrade the GPUs to the most modern chips that Nvidia is making.

“The idea for this is to be a fully open-source AI technology, similar to what Twitter is doing with Grok,” he said.

Roadmap

Qtum has about 47 people now. The company is based in Dubai, where regulations on blockchain businesses are pretty clear.

Qtum is still working on its roadmap, but it expects to have its visual generator soon as well as an assistant.

“We’re building our own LLM with the team and part of the reason is we started with open source. We are overcoming the different issues around security, openness and performance,” he said. “Performance is very important. This is not cheap. And if you’re not optimized, you’re just going to be wasting resources.”

Palencia views his role as a kind of caretaker as well as an opportunistic entrepreneur.

“AI is probably the third most disruptive technology I’ve seen in my life,” Palencia said. I’m 42. I was able to see the dawn of the internet. I was able to see the dawn of Bitcoin in the crypto era. And this is about as disruptive or perhaps even more than those two. It’s going to combine everything that exists.”

Yet he knows AI can be a dangerous weapon if not handled correctly. And that’s why he doesn’t want AI to be a closed source technology. He wants it to be open and embrace the foundations of crypto such as decentralization.

“This is probably the most dangerous weapon. So I believe that it’s not just for profit we’re in this for right now. It’s mostly a tool that can help shape the future of humankind. And if possible, it better be a positive tool instead of a negative tool for people,” he said.

AI has to be open, he said.

“This just can’t be handled by a single entity that might or might not have their own agenda,” he said.