IO

그냥 꼬맹19d ago

1. https://www.artemis.ai/tim0x/article/305506378247231184 2. https://x.com/Grayscale/status/2102098573827428852 3. https://x.com/BlackRock/status/2102409739175141458?s=20 The convergence point of recent posts and flows ultimately comes down to computing, energy → DePIN, inference. Using the article's framing... token costs are falling and AI demand is naturally increasing. Cheaper intelligence generates more prompts, more prompts lead to more inference, more inference drives computing demand increases. As access costs drop, compute demand grows faster than infrastructure can scale. Thus computing becomes a potential bottleneck. When computing becomes a bottleneck, markets seek efficiency—this is where DePIN becomes important. Therefore DePIN demand could emerge first from computing and expand into physical infrastructure like energy and data centers. I focused on the structure itself from energy → inference, and organized my research around tokens from projects that have actual demand and value capture, even if minimal. 1. Recurring inference supply / inference marketplace: $VVV, $DOT, $ORBIO, $POD, $SERV 2. GPU collateral/insurance, marketplace, rental: $CHIP, $AKT, $IO, $ATH 3. Computing, energy, network: $PRL, $NOCK, $GENSYN, $TAO 4. DePIN: $GEOD + 2nd category Teams I think are the best at this. Personally I wonder if focusing on DePIN right now would be good? Physical infrastructure itself is becoming a bottleneck, so I can see the flow shifting. Some tokens even haven't launched yet.

4.4K3.5× usual viewsSpread to 4 channelsVVVGEODTAOShow Korean original