In this episode of Investor Mindset, Anton Saburov speaks with trader and investor Abhinesh Daas about a new financial idea at the intersection of artificial intelligence and derivatives: compute futures. They discuss why the cost of AI infrastructure may need its own hedging market, how professional traders could provide early liquidity, why Nvidia can be compared structurally with OPEC, and what a more transparent market for AI compute could mean for investors and companies.
What we discuss in this episode
Abhinesh sees compute futures as a response to a familiar commodity problem: when an essential input has volatile supply and demand, businesses need a way to make future costs more predictable rather than carrying the full price risk themselves.
For AI companies, the practical use case is hedging. A startup or enterprise that expects large compute or token expenses could use futures or options to create a target cost range, reducing the risk that an unexpected price spike destroys the operating budget.
He expects professional traders and market makers to arrive before most operating companies. In his view, Wall Street liquidity is likely to come first, followed by broader institutional adoption as pricing, regulation and trust improve.
The bigger financial change could be the forward curve. A transparent market price for future compute would give companies, investors and lenders a reference point for budgeting AI costs instead of relying only on internal assumptions.
Abhinesh compares AI compute with oil because both can become critical inputs whose price affects entire industries. He also argues that AI may eventually have an even wider range of applications because programmable intelligence can be deployed across software, healthcare, cybersecurity and other sectors.
The Nvidia-OPEC comparison is about concentration and pricing power rather than intent. Nvidia currently sits at the center of a dominant GPU ecosystem, while AMD, Google TPUs, open-source models and future technologies could play a role similar to new supply sources that weaken a concentrated market over time.
The supply chain adds another layer of risk. TSMC in Taiwan and ASML are important parts of advanced chip production, so capacity constraints, geopolitical tension or operational disruptions could quickly affect the economics of AI infrastructure.
A serious compute derivatives market still needs credibility, transparent benchmarks and multiple sources of pricing data. Abhinesh argues that regulation and participation by established financial institutions would make the instrument easier for companies and investors to trust.
For his own fund, Abhinesh describes a more defensive posture: less time fully exposed to the market and more cash held in reserve while waiting for a possible reset in AI-related pricing and infrastructure expectations.
His long-term thesis is not that one company must win. He is interested in gaining exposure to the broader demand for AI compute without having to choose whether Nvidia, AMD, Google or another technology platform ultimately dominates.