
Pasadena Startup Compute Labs Launches Pilot Program to Tokenize AI Infrastructure
A Pasadena startup is looking to capitalize on the AI boom by investing in the infrastructure that powers data centers – the GPUs, or graphic processing units, that make up the data centers supporting large language models.
'Our goal is to democratize access to the infrastructure layer,' said Albert Zhang, chief executive of Compute Labs. 'There is a significant demand over supply situation. Many GPUs are running at 100% utilization. The yield for these assets are really high at the moment.'
That moment could have long tail winds as companies such as OpenAI, Google and Meta plan to invest billions of dollars in data center development. For example, Meta reportedly was in discussions with private credit investors including Apollo Global Management KKR, Brookfield, Carlyle and PIMCO to raise $26 billion in debt to finance data centers.
Compute Labs raised $3 million in a pre-seed round last year led by Protocol Labs. The company purchases equipment on behalf of accredited investors and then leases it to data centers which pay on a revenue sharing model. The assets are sold to investors through a digital token that is collateralized against the physical asset. These tokens pay regular distributions and can be traded on digital currency exchanges. In this model, the data centers are able to offload a capital expenditure and turn it into a regular operating expense. Otherwise, operators would typically rely on private lenders.
It launched its first data center investment in June with $1 million that has all been invested and distributed as tokenized GPUs. It plans to raise $10 million following the pilot deal and has over $100 million in GPUs in its pipeline ready to match with investors.
Zhang's background includes working at a Y Combinator company and at a financial technology company. He pivoted to AI in 2022 when OpenAI was released. A discussion with an angel investor from the semiconductor industry told Zhang that if he could have started over, he would have invested in the infrastructure of the business. Companies had started selling assets such as U.S. Treasuries as tokenized digital assets. Plus, Jensen Huang hosted a GPU conference in March 2024 where he said that the computer will be the currency of the future.
'After that, we closed within two weeks,' said Zhang. 'It's like syndicating real estate deals, but the asset class is new. We have a lot of challenges as a super young company without a track record, and investors don't realize that the GPU can yield at such a high rate.'
There are additional regulatory challenges, but some of those were addressed with the recently passed Genius Act, which includes a clearer framework around stable coin and real-world assets. Compute Labs is also looking into a SPAC merger, which would make it a public company and give it broader access to capital.

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


Bloomberg
15 minutes ago
- Bloomberg
OpenAI Staffers to Sell $6 Billion Worth of Shares
Bloomberg's Matt Miller discusses plans by current and former OpenAI employees to sell $6 billion worth of shares to an investor group that includes SoftBank. Plus, investors react to US government plans to take a stake in Intel. And global competition in the electric vehicle space picks up. (Source: Bloomberg)


Entrepreneur
44 minutes ago
- Entrepreneur
Google AI Pioneer Employee Says to Stay Away From AI PhDs
Jad Tarifi, who founded Google's first generative AI team and has a Ph.D. from the University of Florida, wouldn't recommend starting a program now. AI researchers are in high demand, with some offered billion-dollar compensation packages from Meta amid the ongoing AI talent wars. However, one AI pioneer, Jad Tarifi, who founded Google's first generative AI team after obtaining a Ph.D. in AI, would not recommend higher study to break into the field. In a new interview with Business Insider, Tarifi, 42, predicted that within the five to seven years it takes to obtain a Ph.D., most of AI's problems will be solved. "Even things like applying AI to robotics will be solved by then," Tarifi told BI. Related: AI Is Going to 'Replace Everybody' in Several Fields, According to the 'Godfather of AI.' Here's Who He Says Should Be 'Terrified.' Tarifi explained that obtaining a Ph.D. was only for "weird people" who were "obsessed" with a certain field because higher education required "a lot of pain" and at least five years of their lives. He recommended staying away from the Ph.D. route altogether or choosing to specialize in a subfield of AI that is still in its early stages, like AI for biology. Tarifi received a Ph.D. in 2012 from the University of Florida, where he worked on an AI theory that combined principles from neuroscience, geometry, and machine learning, according to his LinkedIn. He then joined Google, where he became a tech lead and manager for nearly a decade, working on models for Google's generative AI projects. Tarifi is now the founder and CEO of Integral AI, a startup that focuses on creating AI agents to act autonomously on behalf of users. Related: These 3 Professions Are Most Likely to Vanish in the Next 20 Years Due to AI, According to a New Report In the BI interview, Tarifi also warned prospective students from completing degrees in law and medicine, arguing that the information in these programs was "outdated" and memorization-based. Tarifi isn't the first person to warn students away from higher degrees. Venture capitalist Victor Lazarte said earlier this year that AI is "fully replacing people" in the legal profession. He predicted that AI would take over entry-level legal positions usually filled by recent law school graduates within the next three years. Join top CEOs, founders and operators at the Level Up conference to unlock strategies for scaling your business, boosting revenue and building sustainable success.


CNET
44 minutes ago
- CNET
OpenAI CEO Sam Altman Believes We're in an AI Bubble
OpenAI CEO Sam Altman believes that we're currently in an AI bubble, given all the AI hype from investors and capital expenditures. Altman made the statement during a conversation with The Verge and a handful of other reporters on Thursday. "Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes," said Altman. "Is AI the most important thing to happen in a very long time? My opinion is also yes." Other revelations include his regrets over the sudden rollout of GPT-5, worries about the parasocial relationship some people have with his AI chatbot and how he doesn't want ChatGPT to become a sex robot. "I think we totally screwed up some things on the rollout, " Altman said of the launch of GPT-5, the companiy's latest AI model. When it launched earlier this year, it replaced all previous models, including GPT-4o. This led to protests by fans who preferred the more conversational nature of 4o. In response, OpenAI once again gave Plus users access to 4o. 700 million users and growing Altman also revealed that ChatGPT has 700 million weekly users, making it the fifth most popular website in the world. He predicts that ChatGPT will jump up to the third spot soon, beating Instagram and Facebook but behind Google and YouTube. The popularity means OpenAI's servers are at capacity. The load is so great that Altman admits OpenAI can't release better models it has already developed because there isn't enough server capacity to keep up. Altman says OpenAI will spend a trillion dollars on data centers in the "not very distant future." Altman also took a slight dig at Elon Musk's Grok, which released an AI companion that leaned more into risqué territory. "You will definitely see some companies go make Japanese anime sex bots," said Altman, arguing that OpenAI wants to make useful apps and to not exploit those in fragile mental states. Investment into AI development is at an all-time high. Capital expenditure in AI added more to the US GDP in the last two quarters than all consumer spending, according to Renaissance Macro Research. Considering that US consumers like spending money, it's a staggering statistic, and the firs time such a stat has ever been recorded. Google, Amazon, Meta and Microsoft plan on spending $364 billion in AI in 2025 alone. The problem is that AI capital expenditure is boosting the economy overall, and any changes to it could have major external effects. The ongoing global tariffs placed by the Trump administration means investors are looking to software companies as a bit of a safe haven, because they deal less with importing and exporting of goods. Analysts worry that AI is creating a massive economic bubble, and if it were to pop, the reverberations could be massive, potentially crashing the economy. Altman also said that OpenAI will make a brain-computer interface to take on Elon Musk's Neuralink, more apps beyond ChatGPT are on their way and that OpenAI is interested in buying Chrome if the government forces Google to sell the popular web browser.