The artificial intelligence sector has exploded and seems to show no signs of slowing down. However, most companies and projects utilize centralized services such as Amazon Web Services (AWS) for cloud computing power, which can be costly and prone to outages.

Decentralization and blockchain are solutions to these issues, according to Theta Labs head of strategy Wes Levitt, who told The Agenda podcast that decentralized cloud computing saves customers money and offers greater reliability and flexibility.

Levitt spoke to hosts Jonathan DeYoung and Ray Salmond for episode 51 of The Agenda, where he broke down how decentralization helps democratize access to computing power, why academia has become Theta Labs’ top customer and more.

Decentralized cloud services expand access to AI

Theta Labs is the company behind Theta Network, which describes itself as a decentralized cloud for AI, media and entertainment. According to Levitt, Theta is focused on decentralizing the cloud in order “to get away from having only a few points of failure and from having near-monopolies in the case of the major cloud providers and what they can provide in pricing and other terms.”

Levitt shared that while Theta started out focused primarily on media and entertainment, it has seen a massive influx in AI-related customers since the current AI boom began. One particular sector that has found decentralized cloud computing especially useful is academia. “The strongest inroads we’ve made in the last six to 12 months has been in academia, actually,” he said. 

“We caught on with KAIST, a university in Korea, and pretty quickly we were able to expand to four out of the five top Korean universities,” he explained, adding that Theta also “recently signed on University of Oregon and [is] talking now to a few of the US universities.”

“It’s not a model that really fits for them to say, ‘We’re going to sign a three-year contract with heavy commitments on it.’ They want to have flexibility and be able to use a lot when they want, and not pay for times when they’re not.”

Theta Labs vs. Amazon Web Services

The decentralized AI niche within the blockchain space has exploded over the past year. According to data from CoinMarketCap, the market capitalization of AI and big data crypto projects has grown from $16.17 billion in December 2023 to peak at over $70 billion on Dec. 6, 2024.

                   Market capitalization of AI and big data tokens. Source: CoinMarketCap

Japanese tokenization platform Progmat is exploring the potential for tokenizing overseas real estate. It’s the most prolific tokenization platform in Japan, founded by MUFG with backing from the other big three banks, SMBC and Mizuho, alongside others. So far most of the security tokens issued have been linked to single large real estate properties in Japan, with one or two digital bonds. Now they’re considering how to expand the types of security tokens.

There’s an affiliated Digital Asset Co-creation Consortium (DCC) with 282 members. A much smaller DCC working group explored which types of tokenized assets to consider. Out of sixteen options, overseas real estate is the top of a shortlist of five. The others are real estate construction, wind power infrastructure, solar infrastructure and funds.

The decision making process was quite informative. Each token type was assessed according to four criteria. On the commercial front, does the asset serve a need or is it in demand? And how profitable would it be? Then there’s the question of feasibility. Is there an existing scheme or structure that can be used? And what are the risks? Each of the 20 working group members voted on each criterion for every asset class.

Overall the conclusion was there would be strong demand for all five of the shortlisted assets and all were considered profitable, with less confidence about real estate construction.

The real estate tokenization choices

Regarding the scheme or structure, most Japanese real estate security tokens use a trust structure and this could be used as a template for overseas real estate. Except it’s a two tier trust structure with a domestic trust and an overseas real estate trust. Alternatively, the investment could be via a limited liability company (LLC). Regarding risks, the one big caveat was exchange rate risks. Additionally, the asset managers would need to watch out for local tax laws, especially the potential for double tax.

The working group also considered different types of exposures to overseas real estate – single property versus portfolios or real estate loans. Portfolios could either be a physical portfolio, a real estate fund or a REIT.

When they compared the trust model investing in a specific property versus an LLC investing in REITs, the LLC/REIT route was the lowest cost and risk, and the easiest to implement by far. However, the working group preferred single properties held via a trust. That’s in part because it’s the same model that has worked for security tokens so far in Japan. But the final decision appeared to be about marketing. It’s easier to sell single properties to individual investors because they are more tangible compared to a faceless REIT.

Meanwhile, Progmat Coin is progressing its stablecoin initiatives. The big three banks launched Project Pax, which will use stablecoins for cross border payments. Companies will instruct their banks in the usual way, and on the backend it will still use Swift messages. However, instead of routing payments via correspondent banks they will go via blockchain.

Link