The hardest question for AI computing networks isn’t “How many machines have connected?” but who is willing to keep paying for the service. We can start with the prices: Rolling 24 hours, $TAO -0.40%. Rolling 24 hours, $RENDER -0.52%. Rolling 24 hours, $FET +0.28%. These are only changes in spot quotes. They don’t show how strong customer demand is, much less serve as a scorecard for computing orders or product acceptance.
To assess product demand, we need to know whether customers are paying for a real service—and ideally, whether they buy it repeatedly. To assess computing supply and demand, we need to know whether resources can be matched to workloads and whether delivery meets requirements. Right now, there is no reliable information to confirm the customers, revenue, utilization, or orders of these networks, so I won’t fill in the blanks for the projects. Even if there are paid tasks, we still need to clarify exactly what customers are buying and whether demand can last.
Network incentives address why providers are willing to participate. Who funds the rewards, and whether they can be sustained over time, are separate questions. Rewards can attract computing supply, but they can’t replace customer payments; stable supply doesn’t mean anyone wants the service. If participation depends on subsidies, we’ll need to see whether supply remains when rewards change.
Token value needs to be assessed separately too: Where do fees go? Are tokens actually used for access or settlement? Do holders have clearly defined rights? A network having users doesn’t necessarily mean its token captures value. Nor does growth in a company’s AI business automatically imply new orders for the networks associated with these tokens. Each link between product demand, computing supply and demand, network incentives, and token value needs evidence.
So these price moves don’t explain how the networks are performing as businesses. I’d rather see who is paying consistently, whether services are delivered to standard, who is funding rewards, and whether there’s a clear mechanism connecting revenue to the tokens. AI products are worth watching, but simply adding an “AI” label isn’t enough to prove that anyone is paying—or that the tokens will benefit.
To assess product demand, we need to know whether customers are paying for a real service—and ideally, whether they buy it repeatedly. To assess computing supply and demand, we need to know whether resources can be matched to workloads and whether delivery meets requirements. Right now, there is no reliable information to confirm the customers, revenue, utilization, or orders of these networks, so I won’t fill in the blanks for the projects. Even if there are paid tasks, we still need to clarify exactly what customers are buying and whether demand can last.
Network incentives address why providers are willing to participate. Who funds the rewards, and whether they can be sustained over time, are separate questions. Rewards can attract computing supply, but they can’t replace customer payments; stable supply doesn’t mean anyone wants the service. If participation depends on subsidies, we’ll need to see whether supply remains when rewards change.
Token value needs to be assessed separately too: Where do fees go? Are tokens actually used for access or settlement? Do holders have clearly defined rights? A network having users doesn’t necessarily mean its token captures value. Nor does growth in a company’s AI business automatically imply new orders for the networks associated with these tokens. Each link between product demand, computing supply and demand, network incentives, and token value needs evidence.
So these price moves don’t explain how the networks are performing as businesses. I’d rather see who is paying consistently, whether services are delivered to standard, who is funding rewards, and whether there’s a clear mechanism connecting revenue to the tokens. AI products are worth watching, but simply adding an “AI” label isn’t enough to prove that anyone is paying—or that the tokens will benefit.
