Kalshi CEO: AI Compute Market Needs Price Discovery

The burgeoning AI sector's voracious appetite for computing power is creating a new frontier for financial markets. Kalshi, a prediction market platform known for its event contracts, is stepping into this space with the launch of markets focused on AI computing power. Udesh Jha, Chief Risk Officer at Kalshi, who previously spent 16 years at CME Group, believes these new offerings will provide essential price discovery for a rapidly evolving and critical commodity.

From Derivatives to Compute Markets

Jha, who joined Kalshi earlier this year, drew a parallel between the current excitement surrounding AI and the financial industry's transformation following the Dodd-Frank Act, where he was instrumental at CME Group. "Looking ahead 16 years, I think I see the same excitement in the financial industry with prediction markets, with the migration to tokens, with cloud adoption, and so many other things," Jha stated. "From that standpoint, I think Kalshi, at the center of a lot of that innovation, is a really exciting place."

While Kalshi is widely recognized for its event contracts, such as betting on the outcome of sporting events, the platform is actively diversifying. Jha highlighted the company's recent initiatives, including the expansion into areas like cryptocurrency markets and KPI markets, and now, the compute market.

The full discussion can be found on Bloomberg Podcast's YouTube channel.

Addressing the Compute Commodity Challenge

The introduction of AI compute markets is particularly timely, given the global demand for GPUs and the increasing commoditization of computing power. Jha acknowledged the complexity of this market, noting that unlike standardized commodities like oil, compute power varies significantly based on factors like location, chip type, grade, and intended use.

"The compute or the GPU market is not standardized at this point in time," Jha explained. "There are many different factors that play into that index. It could be locations, the chips, the different kinds of chips, the grades, the different uses." He added that traditional futures contracts might struggle to capture this nuance. "Rather than focusing on a particular index or a particular contract, the prediction market allows us to list several of these contracts, which is what we have done."

Kalshi's market-driven approach, where prices are discovered through the collective wisdom of participants, is seen as a key differentiator. "Ours is a truly market-driven forward curve," Jha emphasized. "And that is where the power of the wisdom of the crowd, of the marketplace, and the accuracy that comes out of it, comes to full form."

Market Potential and User Base

The total addressable market for compute power is estimated to be in the trillions, with hyperscalers alone committing hundreds of billions to infrastructure by 2026. Jha expressed confidence that this burgeoning market will attract significant institutional interest.

"We do expect large institutions to take advantage of this innovation by looking at the forward curve and using it to price deals, to do OTC contracts, to do other sorts of derivatives for their purposes," he said. The platform anticipates a diverse user base, including hyperscalers, cloud providers, AI inference companies, medical research firms with high computing needs, and speculative traders.

Competition and Regulatory Scrutiny

Kalshi is entering a space where established players like CME Group are also exploring compute futures. However, Jha believes Kalshi's prediction market model offers a distinct advantage in capturing the market's inherent variability.

The company's commitment to regulatory compliance, including waiting four years for regulatory approval for perpetual futures, was highlighted as a testament to its risk-first approach. "Risk first, and adopting of technology, is one of the reasons why I feel very confident that we have the right to build in this marketplace," Jha stated.

Addressing concerns about the concentration of activity in sports gambling on prediction markets, Jha noted that while sports events remain popular, the growth in true risk management products like KPI markets and the new compute markets is significant. "Overall, I think the whole pie is growing quite comfortably," he concluded.