English
Liquid Compute
Report

Liquid Compute✓

@liquidcompute

12 Following1.3K Followers50 Posts
  1. Liquid Compute✓@liquidcompute
    We had a great time last night in San Francisco hosting Compute & Cocktails with the General Compute team! It was a fun night talking compute and financing with some of the sharpest minds in the space over a few drinks. Thanks to everyone who came out—we loved getting to catch up with familiar faces and meet new ones. We’re planning another event in NYC soon. Stay tuned for details! @general_compute @mdshelton_
    Download
    019315
  2. Liquid Compute✓@liquidcompute
    Just last week at the @nvidia GTC conference, @JensenHuang put his finger on the problem we've been working on: "What is missing in the middle is the working capital to fuel that growth… because there's just so much demand across so many sectors of AI, there's a scarcity of capital." Today, at Liquid Compute, we're excited to share a milestone on that front: arranging a $19M bridge facility for a neocloud with signed customers and GPUs on order. The supplier wanted to be paid when the order was placed, but term lenders only fund once the hardware is installed, months later. Closing that gap is how capacity that otherwise wouldn't get built comes online, and it's why we're building Liquid Compute as a venue to buy, sell and finance compute in one place. As Jensen said, most neoclouds run into this gap, and it's one of several financing problems we work on. If you're dealing with any of this, reach us at
    Download
    02275.8K
  3. Liquid Compute✓@liquidcompute
    Anthropic filed its S-1 on Oct 26. Most coverage will focus on the revenue. We read it for the compute contracts: who they’re buying from, on what terms, and how some of it is being financed. Notes below.
    224110.8K
  4. Liquid Compute✓@liquidcompute
    Posts @liquidcompute
    4/5 The Broadcom TPU deal is a lease: about $161B of rent, paid quarterly, usually once the gear is ready for service. Broadcom also agreed to lend up to $42B, if needed, to cover that rent, at 9% PIK. The agreements are in the exhibits.
    103529
  5. Liquid Compute✓@liquidcompute
    Posts @liquidcompute
    5/5 Anthropic says it books compute 12 to 18 months out and pays up for short-term capacity when demand runs ahead. Most of its revenue is pay-as-you-go. Fixed multi-year costs against usage-based revenue is the exact mismatch we work on.
    003495
  6. Liquid Compute✓@liquidcompute
    Another week down for the neoclouds. CoreWeave closed at a new 2026 low. B300 rental prices rose again, their 34th straight daily gain. Rent up, landlords down.
    Download
    11261.5K
  7. Liquid Compute✓@liquidcompute
    This is the underpinning of our thesis at Liquid Compute. There's plenty of demand for compute but not enough capital to fund it. Oil, power and grain all have mature financing and hedging markets. Compute has none, and we're building them to help close that financing gap.
    First Squawk @FirstSquawk
    NVIDIA CEO HUANG: NEOCLOUD DEMAND IS ‘OFF THE CHARTS’ BUT CAPITAL IS SCARCE - CNBC
    11242K
  8. First Squawk✓@FirstSquawk
    NVIDIA CEO HUANG: NEOCLOUD DEMAND IS ‘OFF THE CHARTS’ BUT CAPITAL IS SCARCE - CNBC
    3240584111K
  9. Liquid Compute✓@liquidcompute
    Compute & Cocktails 🍸 How do you finance AI infra? @stanleylee21 (@liquidcompute ) + Mark-Daniel Shelton (@general_compute ) on underwriting compute, getting past prepay and the right debt/equity mix. 📍 Dawn Club, SF · Wed, Oct 28 Request to join →
    02161.5K
  13. Liquid Compute✓@liquidcompute
    Liquid Compute
    RT @stanleylee21: Good post from @prashantrad . The question that matters for AI debt is who actually bears the adoption risk. Take-or-pay from a strong counterparty and the lender barely cares about adoption speed. Leave utilization or pricing exposed and it's a different credit. That's exactly what our @liquidcompute Offtaker Credit Map scores. All publications can be found here
    P_Rad @prashantrad
    There's understandably a lot of consternation about AI related debt out there. I read a NYT interview with Peter Orszag, tied to a paper he co-wrote on the new supply side era, where he compares the scale of AI capex debt to the fiber buildout of the 1990s. The scale of capex investment is enormous, but what is going to matter the most is the speed of adoption of AI, particularly at 'traditional' firms that are not VC or tech related. Whether real adoption happens over the next few years or the next decade, changes the return on capital completely. For the debt financing this capex, there is another question: who actually bears the adoption risk? If the contracts are long-term, take-or-pay and backed by strong counterparties, the pace of end-user adoption may matter much less to the lender. If the contracts leave utilization, pricing or renewals exposed to the market, that is a very different credit. Compute pricing is showing a similar pattern. Rent it now and it's expensive, because near term demand is the one thing everyone can see clearly. Go out 6 months, a year or two years and the compute pricing curve inverts, meaningfully, because nobody has real visibility into usage, adoption or which GPUs will be standard by then. Thanks @stanleylee21 for sharing this info. So if you're pricing any AI related financing, the pace of adoption is a key underwriting input. And just as important is determining who has contractually agreed to bear the risk if adoption takes longer than expected.
    01025
  14. P_Rad✓@prashantrad
    There's understandably a lot of consternation about AI related debt out there. I read a NYT interview with Peter Orszag, tied to a paper he co-wrote on the new supply side era, where he compares the scale of AI capex debt to the fiber buildout of the 1990s. The scale of capex investment is enormous, but what is going to matter the most is the speed of adoption of AI, particularly at 'traditional' firms that are not VC or tech related. Whether real adoption happens over the next few years or the next decade, changes the return on capital completely. For the debt financing this capex, there is another question: who actually bears the adoption risk? If the contracts are long-term, take-or-pay and backed by strong counterparties, the pace of end-user adoption may matter much less to the lender. If the contracts leave utilization, pricing or renewals exposed to the market, that is a very different credit. Compute pricing is showing a similar pattern. Rent it now and it's expensive, because near term demand is the one thing everyone can see clearly. Go out 6 months, a year or two years and the compute pricing curve inverts, meaningfully, because nobody has real visibility into usage, adoption or which GPUs will be standard by then. Thanks @stanleylee21 for sharing this info. So if you're pricing any AI related financing, the pace of adoption is a key underwriting input. And just as important is determining who has contractually agreed to bear the risk if adoption takes longer than expected.
    0031.1K
  15. Liquid Compute✓@liquidcompute
    Part II of the GPU Offtaker Credit series today. You can find it along with our other research on our website: Stay tuned as the series continues!
    Stanley Lee @stanleylee21
    025677
  17. Liquid Compute✓@liquidcompute
    Liquid Compute
    RT @ArpitGuptaji19: big and huge, financing is the one true problem in infra sourcing for majority of the market
    Liquid Compute @liquidcompute
    Liquid Compute and K8 announce a $250M prepayment facility to help compute buyers fund prepayment capacity commitments. This facility makes compute financeable the way every other commodity already is, and we are already finalizing our first deployments with a few large inference clouds. We're proud to build this with K8 as we bring the discipline of commodity finance to compute.
    01028
  18. Liquid Compute✓@liquidcompute
    Liquid Compute and K8 announce a $250M prepayment facility to help compute buyers fund prepayment capacity commitments. This facility makes compute financeable the way every other commodity already is, and we are already finalizing our first deployments with a few large inference clouds. We're proud to build this with K8 as we bring the discipline of commodity finance to compute.
    Download
    5167627.5K
  19. Show more posts