Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)H

humanspiral

@ humanspiral @lemmy.ca

Posts
104
Comments
2696
Joined
2 yr. ago

  • Dog shoots woman with shotgun at Nebraska convenience store

    Jump

  • it's important that dogs be armed. Only a good dog with a gun can stop Kristi Noem.

  • California moves to exempt Linux from its upcoming age-verification law after backlash over forcing operating systems to collect users’ ages

    Jump

  • I’d love for these politicians to explain to me how age verification of Linux web servers should work for auto-scaling environments where instances are spun up and terminated automatically based on traffic volume.

    Come on, can't be that tedious. What could it be 200-300 instances tops per day? My kid sister does that many selfies.

  • Surplus value

    Jump

  • There are massive economic plotholes in the movie. Namely that other currencies for coffee or busfare and the like would be normal.

    it seems arbitrarily unnecessary to give everyone 25 years at birth, even if technology to stop aging makes that a "good age". For normal education progression, there is a massive penalty to goof off at college for 4 years, if you only get 1 year to start earning income from degree. There is an infinite money loop of having babies (killing them for their 25 years), though maybe they're unable to spend until age 25.

    What is certain, is the system promotes stronger oligarchy/aristocracy than current one, because everyone is forced to stay in a system with cops protecting the artistocracy. You can't just go live in the forest off berries and squirrels, or move to Africa and buy yourself a goat and build a hut on an empty spot, and survive on $3/year. Mandated systemic participation ensures the increased inflationary gouging that is part of movie's world/plot.

  • Surplus value

    Jump

  • Movie "Out of Time" did this. Oligarchy was definitely not solved by this.

  • Desensitization to guns

    Jump

  • Desk guns are for desk pops. When was the last time you did a desk pop?

  • Nivida keeps selling more GPUs than the datacenter capacity that came online. estimate of 560mw for Q1 2026. Up significantly from 1270mw surplus over entire 2025.

    Jump

  • More precise pricing trends from premium tier 2 networks, show demand has drastically fallen over the quarter. H200 very close to its bare runcosts. Theory is that Anthropic's overpayment for Collosus 1 (xAI) capacity has drastically reduced utilization at cloud rental service.

    NVIDIA B200 Blackwell (192GB HBM3e) 

    • Tier-1 Spot Mechanic: AWS introduced B200 spot availability (p6 family baseline). However, because Tier-1 spot pools are subject to extreme automated reclamation, they command a rigid premium.
    • The Rebound: B200 spot experienced a localized surge around March 13th due to massive API volume peaks. 

    Week Ending (2026)Tier-1 Spot (AWS/Azure)Tier-2 Spot (CoreWeave/Nebius)Tier-2 On-Demand (Nebius Menu)The True Market State
    Jan 2$6.40 / hr$4.50 / hr$7.80 / hrInitial launch window; hardware access highly constrained.
    Jan 16$6.40 / hr$4.20 / hr$7.50 / hrData center pipelines face massive backlog queues.
    Jan 30$6.40 / hr$4.00 / hr$7.50 / hrTier-2 unallocated floor space begins opening up.
    Feb 13$6.12 / hr$3.85 / hr$7.20 / hrMulti-agent frameworks consume near-term supply.
    Feb 27$6.12 / hr$3.50 / hr$6.80 / hrInflux of new Blackwell nodes flattens spot markup.
    Mar 13$6.12 / hr$4.10 / hr$6.50 / hrThe Agentic Peak: Spot surges via automated bidding.
    Mar 27$5.90 / hr$2.95 / hr$6.00 / hrHigh supply volumes trigger a localized margin correction.
    Apr 10$5.90 / hr$2.40 / hr$5.50 / hrNebius and Lambda drop public on-demand baseline rates.
    Apr 24$5.56 / hr$2.06 / hr$5.50 / hrThe System Bottom: Spot drops to its absolute low.
    May 8$5.56 / hr$2.40 / hr$5.50 / hrShort-term enterprise fine-tuning contracts absorb space.
    May 22 (Current)$5.56 / hr$2.90 / hr$5.50 / hrCurrent Rebound: Spot firms ahead of June 1 price hikes.

    NVIDIA H200 Hopper (141GB HBM3e) 

    • The Clearance Reality: Unlike the B200, the H200 has failed to rebound. Because its FP8 processing layout lacks Blackwell's native execution upgrades, developers are abandoning the H200 for production serving.
    • Below Cost: Tier-2 spot has flatlined at $1.45/hr, forcing providers to eat losses relative to raw facility overhead just to prevent expensive liquid-cooled spaces from sitting entirely dark.

     

    Week Ending (2026)Tier-1 Spot (AWS p5e baseline)Tier-2 Spot (CoreWeave/Nebius)Tier-2 On-Demand (Lambda/Nebius Menu)The True Market State
    Jan 2$4.20 / hr$2.50 / hr$4.40 / hrHighly stable; utilized as the core long-context architecture.
    Jan 16$4.20 / hr$2.30 / hr$4.25 / hrAWS implements dynamic Capacity Block adjustments.
    Jan 30$4.20 / hr$2.10 / hr$4.00 / hrEarly enterprise teams migration toward Blackwell blocks.
    Feb 13$3.90 / hr$1.95 / hr$3.95 / hrMinor spot stabilization as alternative backends fill up.
    Feb 27$3.90 / hr$1.80 / hr$3.80 / hrShift to newer precision matrices devalues older stock.
    Mar 13$3.90 / hr$1.95 / hr$3.80 / hrMinor agent-driven peak provides short-term support.
    Mar 27$3.85 / hr$1.70 / hr$3.65 / hrMassive bulk capacity deployments flood European hubs.
    Apr 10$3.85 / hr$1.55 / hr$3.50 / hrMarket signals show severe oversupply on legacy nodes.
    Apr 24$3.83 / hr$1.45 / hr$3.50 / hrThe Floor: Prices slide below break-even run costs.
    May 8$3.83 / hr$1.45 / hr$3.50 / hrCapacity remains completely unallocated across major nodes.
    May 22 (Current)$3.83 / hr$1.45 / hr$3.50 / hrCurrent Stagnation: Zero rebound; structural value tier.

  • Chinese robot tries to dance like Michael Jackson

    Jump

  • they do have more successful dance shows too.

  • Trump’s Endgame Is Surrender. He seems to hope to slip away without Americans noticing the magnitude of this defeat.

    Jump

  • Atlantic doesn't want what's best for you anymore than Netanyahu. This is best outcome, and it's already taking too long.

  • Nivida keeps selling more GPUs than the datacenter capacity that came online. estimate of 560mw for Q1 2026. Up significantly from 1270mw surplus over entire 2025.

    Jump

  • A weird point that Nvidia CFO made to say "Nvidia is awesome" is a claim that GPU rental rates are up year to date. There was a crash at end of 2025. The low for the quarter was Jan 1st. The high was March 10th at peak of openclaw frenzy (validated by openrouter charts). Current rates are lower than that peak. But also comparison to 2025 Q1 (what I thought CFO meant, rates are down significantly) For single GPUs.

    1. NVIDIA A100 (Ampere — 80GB SXM) 

    • Q1 2025 Baseline: High: $2.40 | Low: $1.60 | Close: $1.85
    • Q1 2026 Window: High: $1.65 | Low: $0.80 | Close: $1.15
    • Current Normalized Rate: ~$1.07 / hr (Stable floor; primary use shifts to entry-level fine-tuning and quantized serving). 

    2.[NVIDIA H100 (Hopper — 80GB SXM) 

    • Q1 2025 Baseline: High: $7.00 | Low: $5.50 | Close: $5.80 (Supply constraints started easing, down from the absolute peak $10/hr overcharges of late 2024).
    • Q1 2026 Window: High: $3.45 | Low: $1.70 | Close: $2.35 (Hit an absolute low floor of $1.70 in late 2025 before a 38% contract rebound in March due to an influx of video-generation workloads).
    • Current Normalized Rate: ~$2.49 / hr (The standard baseline workhorse for mainstream API serving). 

    3. [NVIDIA H200 (Hopper — 141GB HBM3e) 

    • Q1 2025 Baseline: High: $5.20 | Low: $4.50 | Close: $4.80 (Extremely scarce; reserved exclusively for elite labs running early frontier training).
    • Q1 2026 Window: High: $4.40 | Low: $3.50 | Close: $3.80 (Inventory stabilized as neoclouds widely deployed HGX baseboards).
    • Current Normalized Rate: ~$3.39 / hr (The most cost-effective tier for high-concurrency FP8 deployment). 

    4. [NVIDIA B200 (Blackwell — 192GB HBM3e) 

    • Q1 2025 Baseline: N/A (Sampling/Testing phase; unreleased to the public marketplace).
    • Q1 2026 Window: High: $6.11 | Low: $3.05 | Close: $4.95 _(Initial public availability; premium pric

    5. NVIDIA B300 (Blackwell Ultra — 288GB HBM3e) 

    • Q1 2025 Baseline: N/A (In architectural development; unavailable for rental).
    • Q1 2026 Window: High: $8.50 | Low: $5.50 | Close: $7.25 (Early access provisioning; highly volatile due to constrained data center site capacity).
    • Current Normalized Rate: ~$6.10 / hr (Neocloud standard rate; pricing reflects the premium for its 288GB memory pool). 

    for clusters, google AI mode simply can't provide accurate info. Some providers have fixed premiums, others 0 premium. Many never change prices but mass email promotional discounts. For all I know, this entire analysis could have been a halucination meant to drive my narrative. I have not verified most data claims made as it would be too much work. I imagine most of the specific ones are accurate, and single GPU rental rates are the dominant market in the US, and that data should be solid, but FIIK.

  • Ontario Police Deploy Secret Spyware That Lets Them Remotely Take Over Your Phone

    Jump

  • They need access to your phone to sideload their spyware?

  • With Trump's approval ratings so low, how does he keep ousting GOP rivals?

    Jump

  • I would doubt this. I'd especially doubt that a winning campaign would be to promise to give all KY electricitity to a datacenter. AIPAC did make a victory tweet. I didn't see campaigns, but I doubt it was focused on who loves AI the most.

  • Wayland Comes To Minecraft

    Jump

  • As the target audience of people who have a feeling they may be missing out on something they either need or want, but just don't get it. Sell me!!!!

  • Xbox makes more leadership changes, hiring analyst who said games were losing the attention battle with gambling, crypto and porn as chief strategy officer

    Jump

  • Even if you like all 3 of the other things, it doesn't mean you need it in gaming. It's not as though porn thinks it needs to blend in other attention competitions. A lot of games have math overlap. Gambling is for conning people bad at math.

    In summary, this all seems like bad directions.

  • With Trump's approval ratings so low, how does he keep ousting GOP rivals?

    Jump

  • for Massey, this was mostly AIPAC money even more than they spent on previous primaries to defeat Cori Bush. The playbook for successful APAIC operation is to never mention Israel. Just smear opponent. Not sure how Massey primary was conducted, though.

  • He said "any"

    Jump

  • taps ctrl and shift repeatedly.

  • Congressman Dan Goldman: Trump’s nearly $2 billion fund for MAGA allies is an ‘impeachable offense’ CNN Politics

    Jump

  • Impeaching the head of DOJ, and seizing the fund, should be lower hanging fruit than impeaching Trump, and should proceed today.

  • Nivida keeps selling more GPUs than the datacenter capacity that came online. estimate of 560mw for Q1 2026. Up significantly from 1270mw surplus over entire 2025.

    Jump

  • Yet another big problem for Nvidia is that the H200 is their better product for FP8 mainstream LLM service. Vera-Rubin only has 30% more performance per watt, gb200/300 is lower performance/watt at fp8. But the big expense of all its later generations is liquid cooling, and the extreme weight of liquid cooled racks/NVL72 (3000lbs) that require ultra strong floors with embedded pipes inside them. In yet another F'd up supply chain crisis driven by AI is a 2 year backlog for liquid cooling equipment.

  • Nivida keeps selling more GPUs than the datacenter capacity that came online. estimate of 560mw for Q1 2026. Up significantly from 1270mw surplus over entire 2025.

    Jump

  • Only 2.15gw (out of 5gw) of global datacenters under active construction with hope for 2026 completion is for Nvidia hardware. If there is already high excess inventory (not guaranteed as result of hand me down GPU replacement) then sales/growth must hit a wall eventually. Next 9 months of optimistic deployments is more than next quarters sales forecast.

  • Nivida keeps selling more GPUs than the datacenter capacity that came online. estimate of 560mw for Q1 2026. Up significantly from 1270mw surplus over entire 2025.

    Jump

  • The surplus sales are actually heavily underestimated because the datacenter capacity additions for 2025/26 include non Nvidia hardware. It "appears" that under half of their sales actually make it into datacenter capacity additions.

    1. Stripping Non-NVIDIA Slices from the Available GW Grid 

    To see the true depth of the backlog, we have to look at how much of that newly brought-online data center capacity was immediately consumed by alternative architectures during the 2025 calendar year (4.10 GW total online) and Q1 2026 (1.55 GW total online). 

    A. The Hyperscaler Internal Custom Silicon Tax (ASICs) 

    The largest tech giants do not deploy NVIDIA exclusively. They heavily prioritized their own lower-cost, custom-tailored accelerator chips to handle their native workloads: 

    • Google TPUs (v5p & v6e): Google directed a massive portion of its internal data center buildouts to its proprietary Tensor Processing Units. Throughout 2025 and early 2026, TPU deployments swallowed roughly 450 Megawatts (MW) of Google's net-new global capacity.
    • Meta MTIA & Amazon Trainium/Inferentia: Meta scaled its internal MTIA silicon, while AWS aggressively expanded its Trainium2 clusters. Combined, these internal hyperscaler projects consumed an estimated 300 MW of online grid space across the two periods.

     

    B. The AMD Alternative Squeeze 

    AMD's MI300X and MI325X series secured massive enterprise and cloud traction, specifically anchoring flagship clusters inside Microsoft Azure and Oracle Cloud Infrastructure (OCI). AMD's total shipment footprint accounted for roughly 400 MW of power demand globally over this timeframe. 

    C. Specialized Wafer-Scale Architectures (Cerebras) 

    While smaller in pure megawatt terms compared to hyperscalers, Cerebras built massive high-density footprints. Their multi-million dollar wins—such as the massive 750 MW master deployment framework with OpenAI—began systematically occupying high-density colocation space. Across 2025 and Q1 2026, Cerebras deployments locked down roughly 100 MW of specialized, high-cooling capacity. 

    2. Recalculating the True NVIDIA "Space Deficit" 

    When we subtract these non-NVIDIA hardware deployments from the total physical data center capacity brought online, we find the Net Grid Space Actually Available for NVIDIA: 

    Time HorizonTotal New Global Capacity OnlineMinus Non-NVIDIA Hardware (TPUs, AMD, etc.)Net Grid Space Left For NVIDIA
    Full Year 20254.10 GW\- 1.10 GW3.00 GW
    Q1 20261.55 GW\- 0.25 GW1.30 GW

    Now, let's remap this accurate "Available Space" baseline against the True Grid Power Shipped by NVIDIA (GW Sold) that we calculated using our refined financial models: 

    The Compounding Backlog Realities 

    • The Refined 2025 Gap: NVIDIA shipped 5.37 GW of compute power. If only 3.00 GW of real-world grid space was actually left over for them after accounting for Google TPUs and AMD chips, the real 2025 data center deficit jumps from 1.27 GW to a staggering 2.37 GW.
    • The Refined Q1 2026 Gap: NVIDIA shipped 2.11 GW of compute power. With only 1.30 GW of net data center capacity available to absorb it, the quarterly deficit widens from 560 MW to 810 Megawatts.
  • Technology @lemmy.ml
    humanspiral @lemmy.ca

    Nivida keeps selling more GPUs than the datacenter capacity that came online. estimate of 560mw for Q1 2026. Up significantly from 1270mw surplus over entire 2025.

  • Technology @lemmy.ml
    humanspiral @lemmy.ca

    Open AI shenanigans in Cerberas IPO. They have a deal to pay a minimum of $10.53/m tokens (under 100% 24/7 capacity utilization) that is more like $40-$50/m in real world. They charge $14/m

  • Technology @lemmy.ml
    humanspiral @lemmy.ca

    The key hurdle this year facing AI frenzy/bubble: Not much datacenters are actually being completed.

  • Green Energy @slrpnk.net
    humanspiral @lemmy.ca

    update: solar/wind hydrogen in Nebraska that can power a local datacenter, plus another 50km away with 36% ROI

  • Green Energy @slrpnk.net
    humanspiral @lemmy.ca

    Green Ammonia systems in a container now $500/kw. At least in China. This offers simpler storage/shipment than H2 systems. Link is to Canadian company that is competitive.

    fuelpositive.com
  • Music @lemmy.world
    humanspiral @lemmy.ca

    Iran (Queens) of the stone age -- Paper Machete

  • Green Energy @slrpnk.net
    humanspiral @lemmy.ca

    Trump forces higher natural gas and electricity prices on Americans: Gives TotalEnergy (French FFS) $1B to cancel wind to reinvest in US LNG

    totalenergies.com /news/press-releases/united-states-totalenergies-signs-agreements-us-department-interior-end-its-us
  • World News @lemmy.ml
    humanspiral @lemmy.ca

    Someone tell Trump to walk down 5th Avenue, like Iran's President just did.

  • Music @lemmy.world
    humanspiral @lemmy.ca

    Fireworks over Israel

  • LocalLLaMA @sh.itjust.works
    humanspiral @lemmy.ca

    A possible hardware solution for ultra speed (73x faster than H200) self hosted small models that is not dependent on RAM

    dev.to /mothasa/a-25-person-startup-built-a-chip-that-only-runs-one-ai-model-its-73-times-faster-than-nvidia-1b0d
  • Technology @lemmy.ml
    humanspiral @lemmy.ca

    Nvidia results show path towards AI bubble pop. $45B+ increase ($95B total) in supply commitments at all time high RAM prices

    stansberryresearch.com /whitney-tilsons-daily/a-look-at-nvidias-latest-earnings-report-michael-burry-says-nvidia-is-ratcheting-up-the-risk-with-its-ai-spending
  • LocalLLaMA @sh.itjust.works
    humanspiral @lemmy.ca

    Artificial Analysis Intelligence Index and cost benchmarks are useful decision/guidance determinants for which models to use. Analysis for top models.

  • Electric Vehicles @slrpnk.net
    humanspiral @lemmy.ca

    Chinese electric trucks pay for themselves in five years. They may be half the market in three.

  • LocalLLaMA @sh.itjust.works
    humanspiral @lemmy.ca

    openrouter rankings for programming tokens show sharp rise in open models and stagnation of US frontier models

    openrouter.ai /rankings
  • Green Energy @slrpnk.net
    humanspiral @lemmy.ca

    Datacenters can be powered by renewables for 0 cost, if Hydrogen can be sold at $2/kg.

  • Technology @lemmy.ml
    humanspiral @lemmy.ca

    Anthropic CEO important, but evil, essay: The adolescence of technology.

    www.darioamodei.com /essay/the-adolescence-of-technology
  • Green Energy @slrpnk.net
    humanspiral @lemmy.ca

    super critical CO2 generator in commercial deployments in China

  • LocalLLaMA @sh.itjust.works
    humanspiral @lemmy.ca

    Minimax M2.1 : Reasonably small open model that is breakthrough in multi language coding.

    huggingface.co /MiniMaxAI/MiniMax-M2.1
  • Economics @lemmy.ml
    humanspiral @lemmy.ca

    Malaysia made a poor deal with US to provide it with rare earths. It will either ruin its rare earth processing industry (needs Chinese technology/investment, or ruin their mining industry...

  • Green Energy @slrpnk.net
    humanspiral @lemmy.ca

    Solar on water works even in Northern Norway. Good match with aquaculture.