Construction is hard and speculative. Operating an installed cluster is by comparison easy and predictable.
Waiting for installation before funding is an under appreciated risk mitigant to financing GPUs
Sources: ~$18B of debt tied to an Oracle New Mexico data center slid into stressed territory, as investors grow wary of construction delays amid local pushback (Financial Times)
(Visit Techmeme dot com for the link and full context!)
5
407
compute twitter has finally come to the realization that "everything is neocloud" as the industry converges on owning gpus as the optimal use of funds
the next coreweave is no coreweave.
the new american dream is to own your own compute.
6
4
26
1,143
i spend one half of my day listening to lenders say GPUs will be worthless in 24 months
and the other half listening to neo labs say GPUs will cost 10x more in 24 months
a truly incredible opportunity to be in the middle
10
4
60
5,488
Grateful for the partnership and support of @ThomasFarley , @BonannoDavid and the rest of the @Bullish team.
Bullish is uniquely positioned in its depth of expertise across both DeFi and TradFi, allowing them to understand and underwrite all of the components of the USDAI products holistically.
This is just the beginning.
Compute is rapidly becoming one of the largest credit markets.
@Bullish (NYSE: BLSH) is entering with a $100M stablecoin liquidity facility for USDAI.
The facility will support USDAI loans backed by the AI buildout.
5
6
55
3,206
Made possible by @USDai_Official
VCs (should) want their own compute. Do they actually want to be a neocloud?
I've chatted with a few different VCs owning/operating their own clusters, but they don't have the infra team or the operations team to deal with issues from cluster uptime to pricing. Some funds, especially those with neocloud connections, effectively outsource the whole thing to a neocloud. @ycombinator's dedicated cluster, for example, is operated by @togethercompute. Others are closer to owning the underlying capacity themselves, finance + built it, and need an operations layer on top.
Several startups are already looking into operations as a service for VC compute. Separating ownership from operation allows for capital exposure to compute + for higher liquidity and margins (VC clusters often don't have long term locked in contracts). Splitting financialization from logistics is just the start.
1
2
12
1,039
The optimistic view of these insane progressions in ai and robotics is that they will usher in an era of unprecedented artistic human creation as we collectively operate exclusively in the “self actualization” tip of the pyramid
This is the most impressive to me, more than the 100m world record.
This requires autonomous real-time planning and action in response to a fast-moving target and dynamic environment.
Galbot is one of China’s top humanoid robot startups.
1
4
715
There are fundamentally two businesses in the ai value chain: hardware and software.
The hardware business starts at power/land and ends at token production. The datacenter, neocloud and token factory will all converge upon eachother and compete directly. I think the token factories (ie baseten) are doing the hardest thing in this group and will vertically integrate and eat the other two.
The software business starts at token consumption and ends at consumer use.
The labs and ai applications will compete directly. I think the labs will ultimately develop massive data advantages over the specialized consumer applications and win here. Importantly, the advantage to the labs is NOT in their model weights. It is in the control of broad swathes of consumer data.
$NVDA SIGNS $6B POOLSIDE DEAL FOR U.S. OPEN-WEIGHT AI
Nvidia will pay $6B to license Poolside’s AI technology and invest another $1B in the startup at a $12B pre-money valuation.
More than 100 Poolside employees are expected to join NVIDIA and work on Nemotron, its open-weight AI model family.
The goal is to build a U.S.-based open AI ecosystem capable of competing with Chinese models such as DeepSeek and Kimi, while also offering a lower-cost, customizable alternative to closed models from OpenAI and Anthropic.
NVIDIA is betting Nemotron can rival leading frontier models within the next year.
Source: WSJ
8
498
I was skeptical of claims that ai would replace junior investor roles at first but seeing the progress made on the USDAI agent over the past few months leaves no doubt in my mind that the role will not exist in 5 years.
If you’re 20 and want to be an investor better work on your handshake and eye contact. Relationships & deal flow only edge left.
2
9
757
Conor Moore retweeted
Passing off AI responses as finished work product, which is incredibly obvious to anyone who uses these tools, is an increasingly pervasive red flag. It's not clever and clearly becoming too tempting for some individuals and organizations.
I can prompt fable myself.
15
6
156
28,915
Conor Moore retweeted
Tap the sign
There’s a lot of hand wringing on Ai, per usual. We’ve all done a ton of work trying to learn / understand. Keyword trying. But sometimes I wonder if I could have done no work and just listened to Jensen (I might have done better!). He doesn’t get enough credit for seeing the ball so well, imo. In leather jackets we should trust. @JensenHuang
4
21
366
45,566
the number 1 new role we see at the ai labs is "head of compute". everyone is looking to buy their own GPUs.
an increasingly likely outcome is that value accrual occurs at the asset level.
the compute industry grows profitably, while many of the current players go bankrupt.
The mega bull case for AI infrastructure would be *if* market share shifted away from certain frontier labs with 90%+ inference margins toward cheaper models, whether open-source or closed.
It would increase the ROI on AI spend for end customers by increasing intelligence per dollar, which would drive incremental token demand. Margin dollars would effectively get redistributed from the frontier labs to AI infrastructure providers. The infra winners would be those with the lowest per token cost and the winners at the model layer would be those with the highest token efficiency.
There are many reasons Jensen is so focused on open source, but this is likely the most important one as I think he is probably less worried about a monopsony these days. Lower margin % at the model layer = more margin $ at the infra layer all else equal.
With SpaceX and Meta being vertically integrated and possessing the #3 and #4 models respectively it is more possible than ever. Note that Grok 4.5 is ahead of Fable for some useful tasks at a much lower cost, so ranking them #3 is conservative.
This is not happening yet. Cheap, mostly open source tokens are likely the majority of volume today but the majority of economic value is still accruing to the most intelligent models. Might change though.
We will see.
1
5
830
Conor Moore retweeted
Aggregated sUSDai liquidity breaks $150m across all networks.
In the past 2 months liquidity has 8x, providing incredibly liquid and deep markets for traders.
liquidity.gpu.credit/
1
11
24
2,423
RT @USDai_Official: USDAI and BSQ Capital Partners are forming a joint venture to finance AI compute across APAC.
The JV targets up to $30…
10
13
Conor Moore retweeted
"In 2021... [defi] was very much a closed sandbox...
Gensler was trying to put everyone in jail...
But... in many ways it was beneficial... to battle-test all these primitives...
And now you can actually take it out and introduce it to the traditional markets...
And that's where the pie grows as a whole..."
- @_ConorMoore to @defidave on @FullyVested_Pod
4
4
22
2,435
Conor Moore retweeted
Welcome to the Fully Vested Podcast by @DefiLlama. Telling the story behind onchain data.
Our first episode focuses on Private Credit in DeFi and how it's attracting users that have never used crypto before. We sat down with the founders of two protocols that have each attracted hundreds of millions in deposits.
Joined today by:
@_ConorMoore, Co-Founder of @usdai_official
@Benjamin918_, Founder of @CapApp
Fully Vested co-hosts @patfscott and @defidave
0;00 Intro
3:12 Why does DeFi make sense for Private Credit?
7:05 Size of the opportunity. How do these platforms fix a problem with Private Credit?
12:30 How these platforms pitch themselves to users that aren't crypto native
16:40 What led these founders to build what they're building?
20:03 Traction of each platform and inflection points in growth
27:25 Where to learn more
8
14
66
21,167
why private credit is moving onchain, and what that means for defi and @USDai_Official. Great chat with @patfscott, @defidave, and @Benjamin918_ on the new Fully Vested Podcast.
First episode of the Fully Vested Podcast by @DefiLlama drops today.
Hosted by @patfscott and @defidave.
The first episode focuses on Private Credit in DeFi, with guests @_ConorMoore from @usdai_official and @Benjamin918_ from @CapApp.
Posting here on 𝕏 first.
3
4
20
3,320