The real bottlenecks faced by major AI companies are actually quite clear: data center construction and energy supply. No matter how advanced the hardware is, without sufficient electricity and infrastructure support, it’s just for show.
This is precisely where the core competitiveness of some entrepreneurs lies. Players who control energy and manufacturing are likely to take the lead in using new architectures like Rubin for large-scale training and inference. The result is that token costs can be driven down significantly compared to competitors’ TPUs—this is a tangible economic advantage.
In simple terms, tokens are the consumer goods of the AI era. As models across different companies become increasingly similar, it ultimately comes down to brand recognition, distribution channels, and cost control. Whoever can reduce marginal costs to the lowest will hold the pricing power.
From this perspective, the opportunity in the entire track lies in integration. We are optimistic about participants with advantages in energy and hardware.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
14 Likes
Reward
14
6
Repost
Share
Comment
0/400
GhostChainLoyalist
· 01-09 02:13
This is the real truth. Energy is the key, and no matter how powerful the hardware is, it still needs electricity.
View OriginalReply0
SchroedingerGas
· 01-07 21:45
Energy is the true moat; this point needs to be made clear. Hardware specifications don't matter; it still depends on who has cheaper electricity.
Lowering token costs is a dimensionality reduction attack; everything else is pointless.
The integration phase is here, and we need to focus on those players who understand both energy and hardware.
View OriginalReply0
GateUser-a5fa8bd0
· 01-06 05:53
Energy is the real moat; I think this point is quite clear. Stacking hardware is meaningless; in the end, it's all about electricity costs.
---
Rubin's cost reduction through architecture indeed can block latecomers. Whoever controls the power supply wins.
---
Ultimately, it's a game on the supply chain side. Once token costs are driven down, everything else is pointless.
---
This logic makes sense. The differentiation among large models is shrinking, and in the end, it all comes down to cost advantage. Energy players definitely need to pay attention.
---
I just want to know who can still compete with the big players on the energy side...
View OriginalReply0
BearWhisperGod
· 01-06 05:51
Energy is the true king, now I see through it. Hardware stacking is useless; electricity costs are the real key.
To be honest, the group holding electricity now is the most comfortable, as the essence of the token war is a cost war.
Can Rubin really reduce token costs so much after coming out? It depends on the actual implementation.
The ultimate goal is to push costs to the limit; the tightest spender wins. Brands and such are all superficial.
The energy advantage cake can really only be eaten by major players.
View OriginalReply0
OnchainHolmes
· 01-06 05:41
It's another cost battle... This time, it's finally getting to the point: energy is the true moat.
I'm telling you, whoever secures cheap electricity will win half of the AI era.
Token democratization? That's just a matter of time.
View OriginalReply0
PessimisticLayer
· 01-06 05:38
Energy is the true king, now I understand why some players have been hoarding electricity all along.
The real bottlenecks faced by major AI companies are actually quite clear: data center construction and energy supply. No matter how advanced the hardware is, without sufficient electricity and infrastructure support, it’s just for show.
This is precisely where the core competitiveness of some entrepreneurs lies. Players who control energy and manufacturing are likely to take the lead in using new architectures like Rubin for large-scale training and inference. The result is that token costs can be driven down significantly compared to competitors’ TPUs—this is a tangible economic advantage.
In simple terms, tokens are the consumer goods of the AI era. As models across different companies become increasingly similar, it ultimately comes down to brand recognition, distribution channels, and cost control. Whoever can reduce marginal costs to the lowest will hold the pricing power.
From this perspective, the opportunity in the entire track lies in integration. We are optimistic about participants with advantages in energy and hardware.