# Carmen Li: The compute market shifts to forward contracts, fungibility challenges complicate trading, and revenue stability is crucial for GPU providers | Odd Lots > Source: > Published: 2026-06-15 09:09:42+00:00 # Carmen Li: The compute market shifts to forward contracts, fungibility challenges complicate trading, and revenue stability is crucial for GPU providers | Odd Lots Forward contracts emerge as a key tool for stabilizing GPU market amid rising price volatility. ## Key takeaways - The compute market is experiencing a shift from on-demand to reserve and forward contracts due to price volatility. - AI startups and traditional enterprises are major players in the compute market, often requiring short-term resources. - Fungibility poses a significant challenge in creating a market for compute power, as different chips may not be equal in value. - Forward contracts are essential for stabilizing revenue for GPU providers by hedging against price volatility. - There is a notable 38% performance variance for the same GPU chip, influenced by various factors. - The market for compute futures is still uncertain, but there is significant interest in the concept. - Transparency in GPU performance metrics is crucial for users to make informed decisions. - The development of index models for GPUs involves complex data analysis rather than simple averaging. - Normalization of prices aids in understanding basis risk and volatility in GPU markets. - Daily volatility for a 100 H100 GPU is around 20 to 30, indicating a healthy range for commodities. - Performance variability in GPUs can affect trading and valuation in commodity markets. - The compute market dynamics are influenced by the types of buyers and their specific needs. - Understanding how forward contracts work is crucial for managing financial risk in GPU procurement. - The complexities of GPU performance metrics significantly influence pricing in the market. - Price normalization and basis risk management are key to handling volatility in GPU markets. ## Guest intro Carmen Li is the chief executive officer of Silicon Data and Compute Exchange, where she works alongside Don Wilson to build markets for GPU and compute procurement. She has focused on creating a GPU pricing index and a spot marketplace for GPUs, helping standardize compute in a volatile market. ## The uncertainty of compute futures - The market for compute futures remains uncertain, with significant interest but no clear path forward. - Whether there’s interest in compute futures it’s tradable instruments it’s very tbd — Carmen Li - The concept of compute futures is intriguing but lacks a standardized approach. - Interest in compute futures suggests potential growth in this emerging market. - The uncertainty reflects the nascent stage of compute futures as a tradable commodity. - Understanding the current state of the compute market is essential for assessing future opportunities. - The potential for futures trading in compute power is still being explored. - Market participants are keenly observing developments in compute futures. ## Challenges of fungibility in compute power - Fungibility is a major challenge in creating a market for compute power. - Fungibility is a big one like one chip might not necessarily be equal to another chip — Carmen Li - Different chips may not have equal value, complicating trading and valuation. - The lack of fungibility poses a barrier to establishing a standardized compute market. - Variability in chip performance affects their fungibility and market value. - Understanding fungibility is crucial for navigating the complexities of the compute market. - The challenge of fungibility highlights the need for standardized metrics in compute trading. - Fungibility issues underscore the complexities of trading in the compute market. ## The shift to reserve and forward contracts - The compute market is shifting from on-demand contracts to reserve and forward contracts. - You see a lot of people shifting from on demand to reserve even forward contracts right — Carmen Li - Price volatility and resource availability are driving this shift in contract types. - Reserve and forward contracts offer more stability in the volatile compute market. - The shift reflects a strategic response to market dynamics and uncertainties. - Understanding the factors influencing contract types is crucial for market participants. - The trend towards reserve and forward contracts highlights changing market needs. - This shift is significant for both buyers and sellers in the compute market. ## Key buyers in the compute market - AI startups and traditional enterprises are significant buyers in the compute market. - Enterprises who are traditional businesses but they are needing a node two nodes a few servers here and there for their inferencing — Carmen Li - These buyers often require short-term compute resources for their operations. - The diversity of buyers reflects the broad demand for compute power across industries. - Understanding buyer behavior is crucial for navigating the compute market landscape. - The needs of AI startups and enterprises drive demand for flexible compute solutions. - The compute market caters to a wide range of buyers with varying requirements. - Buyer behavior significantly influences market dynamics and pricing strategies. ## Stabilizing revenue with forward contracts - Forward contracts help stabilize revenue for GPU providers. - The producers they need to hedge their revenue volatility by shorting futures or port options — Carmen Li - These contracts allow providers to hedge against price volatility. - Stabilizing revenue is crucial for managing financial risk in the GPU market. - Understanding how forward contracts work is key to navigating market dynamics. - Forward contracts provide a mechanism for managing uncertainty in GPU pricing. - Revenue stability is a critical concern for GPU providers in a volatile market. - The use of forward contracts highlights the importance of risk management strategies. ## Performance variance in GPUs - There is a 38% performance variance for the same GPU chip. - We proved there’s 38% performance variance for the same chip — Carmen Li - Various factors contribute to this performance variability. - Understanding GPU performance metrics is crucial for market participants. - Performance variance affects trading and valuation in the GPU market. - The finding highlights the complexity of GPU technology and its implications. - Performance variability is a significant consideration for buyers and sellers. - The 38% variance underscores the need for transparency in performance metrics. ## Developing complex index models - The development of index models involves complex data analysis. - The way we develop our index model is not simple math it’s not hey you have two h 100 to simple average — Carmen Li - Accurate data analysis is crucial for pricing differentiation in the GPU market. - The methodology behind index models reflects the complexity of GPU metrics. - Understanding index models is essential for navigating the compute market. - The sophistication of index models highlights the importance of accurate data. - Index models play a critical role in determining market pricing strategies. - Complex data analysis is key to developing effective index models. ## Importance of transparency in GPU metrics - Transparency in GPU performance metrics is crucial for informed decision-making. - We believe gave people the option and transparency is the most important thing — Carmen Li - Transparency builds trust and confidence among market participants. - Understanding the significance of transparency is vital for the tech industry. - Transparent metrics enable users to make informed choices in the compute market. - The emphasis on transparency reflects a commitment to user empowerment. - Transparency is a key factor in fostering a fair and competitive market. - The importance of transparency underscores the need for clear performance metrics. ## Normalizing prices in GPU markets - Normalization of prices helps in understanding basis risk and volatility. - We normalize the traded prices based on different characteristics of the model itself — Carmen Li - Price normalization is crucial for managing basis risk in the GPU market. - Understanding price normalization is key to handling volatility in commodity trading. - Normalization aids in calculating settlement prices and managing risk. - The process of normalization reflects the complexity of GPU market dynamics. - Price normalization is a critical tool for traders and investors in the GPU market. - Managing basis risk is essential for navigating the volatile GPU market. ## Healthy volatility in GPU commodities - Daily volatility for a 100 H100 GPU is around 20 to 30. - The daily volatility for a 100 h 100 is around 20 to 30 it’s a very healthy commodity volatility range — Carmen Li - This indicates a healthy range for commodity volatility in the GPU market. - Understanding what constitutes healthy volatility is crucial for market participants. - The specific metric for volatility provides insight into market dynamics. - Healthy volatility is essential for a stable and predictable market environment. - The volatility range reflects the dynamic nature of the GPU market. - Traders and investors rely on volatility metrics to make informed decisions. **Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our [Editorial Policy](https://cryptobriefing.com/editorial-policy/).