AI Compute RadarGPU rental prices per hour
Rent the spike.
Live GPU rental prices per hour for eight classes across Vast.ai, RunPod and Clore.ai, the list prices of AWS, Azure and Oracle Cloud beside them, and the Rent Index that says what they did. A heavy job once a month does not justify a permanent purchase: rent an accelerator for the afternoon, hand it back, pay for the hours.
Eight classes, ten sources
What an hour costs today
One card per GPU class. The big number is what most offers cost right now on the Vast marketplace; below it the lowest price we read anywhere and what the big clouds list. Nothing is blended; every price was read from a public page.
- RTX 309024 GB
$0.27an hour, typical
- Lowest price
- $0.20 · Clore.ai
- Behind the typical price
- 63 offers
- Big clouds
- not offered
▼ Down from last week
- RTX 409024 GB
$0.51an hour, typical
- Lowest price
- $0.34 · RunPod
- Behind the typical price
- 64 offers
- Big clouds
- not offered
▼ Down from last week
- A10040 GB
$0.73an hour, typical
- Lowest price
- $0.73 · Vast.ai
- Behind the typical price
- 33 offers
- Big clouds
- from $1.35 · Hyperstack
● About the same as last week
- RTX 509032 GB
$0.74an hour, typical
- Lowest price
- $0.68 · Clore.ai
- Behind the typical price
- 64 offers
- Big clouds
- not offered
▼ Down from last week
- L40S45 GB
$0.80an hour, typical
- Lowest price
- $0.79 · RunPod
- Behind the typical price
- 13 offers
- Big clouds
- from $1.55 · Nebius
● About the same as last week
- H10080 GB
$3.08an hour, typical
- Lowest price
- $2.50 · Hyperstack
- Behind the typical price
- 7 offers
- Big clouds
- from $2.50 · Hyperstack
▲ Up from last week
- H200140 GB
$5.88an hour, typical
- Lowest price
- $3.59 · RunPod
- Behind the typical price
- only 4 offers
- Big clouds
- from $3.99 · Hyperstack
● About the same as last week
- B200179 GB
$9.99an hour, typical
- Lowest price
- $5.98 · RunPod
- Behind the typical price
- only 2 offers
- Big clouds
- from $6.00 · Hyperstack
▼ Down from last week
"Typical" is the middle of all offers listed on the Vast marketplace at the top of the hour; "lowest" is the cheapest reading across every source we read; "this week" compares this week's middle with last week's, within five percent counts as the same. Fewer than five offers are marked "only". Every number here also stands in the table below.
More detail
Every source in one table, the price per gigabyte-hour, what the clouds list, the index over 7 to 90 days, break-even, and when to rent or buy. Unchanged, one scroll down.
Price board
Every source, one table
What an hour of each GPU class costs right now at every source we read, side by side. Three kinds of number stand here and are never blended: a marketplace median, RunPod's posted floor, and a cloud's list price. The blocks below carry the detail behind each cell.
- Lime marks the cheapest reading in a row. A marketplace median over fewer than 5 offers is shown but does not compete for it: it is one host's asking price, not a market price.
- medianlist price median = the middle of the offers a renter could take at the top of the hour (Vast.ai, Clore.ai). list price = a posted price per GPU-hour: RunPod's lowest on-demand type, or a cloud's instance price divided by its GPU count.
- Marketplaces and RunPod are read every hour, the clouds four times a day. A class opens its own page, a source its own.
| GPU | Vast.aimedian · hourly | Clore.aimedian · hourly | RunPodlist price · hourly | AWSlist price · 4×/day | Azurelist price · 4×/day | Oracle Cloudlist price · 4×/day | Lambdalist price · 4×/day | Nebiuslist price · 4×/day | CoreWeavelist price · 4×/day | Hyperstacklist price · 4×/day | Cheapest | 30-day swingVast.ai low–high |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RTX 3090 | $0.27median 63 offers | $0.20 Cheapest competing reading in this rowmedian 108 hosts | $0.22list priceRTX 3090 | — | — | — | — | — | — | — | $0.20Clore.ai | 49 %$0.21–$0.34 |
| RTX 4090 | $0.51median 64 offers | $0.53median 125 hosts | $0.34 Cheapest competing reading in this rowlist priceRTX 4090 | — | — | — | — | — | — | — | $0.34RunPod | 91 %$0.45–$0.96 |
| A100 | $0.73 Cheapest competing reading in this rowmedian 33 offers | $0.79median 4 hostsfew offers: Fewer than 5 offers stood behind this price, so it is one host's asking price rather than a market price. See "How many offers stand behind this price?" below. | $1.39list priceA100 SXM | $2.74list pricep4d.24xlarge · 8× | $3.40list priceND96asr v4 · 8× | $3.05list priceBM.GPU4.8 · 8× | $1.99list priceA100 SXM | — | $2.70list priceA100 · 8× | $1.35list priceA100 | $0.73Vast.ai | 39 %$0.68–$1.00 |
| RTX 5090 | $0.74median 64 offers | $0.68 Cheapest competing reading in this rowmedian 142 hosts | $0.69list priceRTX 5090 | — | — | — | — | — | — | — | $0.68Clore.ai | 60 %$0.60–$1.07 |
| L40S | $0.80median 13 offers | $0.50median 1 hostsfew offers: Fewer than 5 offers stood behind this price, so it is one host's asking price rather than a market price. See "How many offers stand behind this price?" below. | $0.79 Cheapest competing reading in this rowlist priceL40S | $1.86list priceg6e.xlarge | — | $3.50list priceBM.GPU.L40S.4 · 4× | — | $1.55list priceL40S with Intel CPU | $2.25list priceL40S · 8× | — | $0.79RunPod | 8 %$0.73–$0.80 |
| H100 | $3.08median 7 offers | $0.94median 1 hostsfew offers: Fewer than 5 offers stood behind this price, so it is one host's asking price rather than a market price. See "How many offers stand behind this price?" below. | $2.69list priceH100 SXM | $6.88list pricep5.4xlarge | $6.98list priceNC40ads H100 v5 | $10.00list priceBM.GPU.H100.8 · 8× | $3.29list priceH100 PCIe | $4.50list priceHGX H100 | $6.16list priceHGX H100 · 8× | $2.50 Cheapest competing reading in this rowlist priceH100 | $2.50Hyperstack | 52 %$2.16–$3.60 |
| H200 | $5.88median 4 offersfew offers: Fewer than 5 offers stood behind this price, so it is one host's asking price rather than a market price. See "How many offers stand behind this price?" below. | — | $3.59 Cheapest competing reading in this rowlist priceH200 SXM | $7.91list pricep5en.48xlarge · 8× | $10.60list priceND96isr H200 v5 · 8× | $10.00list priceBM.GPU.H200.8 · 8× | — | $5.40list priceHGX H200 | $6.30list priceHGX H200 · 8× | $3.99list priceH200 SXM | $3.59RunPod | 39 %$4.34–$6.30 |
| B200 | $9.99median 2 offersfew offers: Fewer than 5 offers stood behind this price, so it is one host's asking price rather than a market price. See "How many offers stand behind this price?" below. | — | $5.98 Cheapest competing reading in this rowlist priceB200 | $14.24list pricep6-b200.48xlarge · 8× | — | $14.00list priceBM.GPU.B200.8 · 8× | $6.69list priceB200 SXM6 · 8× | $8.50list priceHGX B200 | $8.60list priceHGX B200 · 8× | $6.00list priceB200 | $5.98RunPod | 36 %$7.50–$10.63 |
Prices per GPU-hour in USD, read from public listings; a list price buys a whole instance with its own terms, a marketplace median buys one GPU from one host. Prices move during the day: check the live offer before renting. Nothing here is a forecast.
Every number on this board was listed: read from the public page of the named source at the stated time, with the offer count where there is one, and never blended into a basket. What a renter actually paid, or what a settlement index pays out, is a different measurement made from data this site does not have; it is not estimated here.
The daily series behind the board, every class and source, one row per day: 30 days as CSV · every day since 26 Aug 2026 as CSV · JSON · how to read it. CC BY 4.0, name the source.
Memory per dollar
What does a gigabyte of GPU memory cost per hour?
The cheapest competing reading of each class divided by the memory of the card behind it. This is the figure a renter with a memory requirement compares across generations: a model that needs 60 GB can live on one card of one class or two of another, and the per-gigabyte price says which is cheaper before speed enters the picture.
In plain words: on November 2, 2026 the cheapest gigabyte-hour was the RTX 3090's at $0.009, the dearest the B200's at $0.033, 3.9 times as much.
How to read this
- What it is: the row's cheapest competing reading from the price board above, divided by the memory in GB of the card that reading refers to; for the A100 class that is the memory of the cheapest variant, 40 or 80 GB, and the row names it.
- Where the numbers come from: the same readings as the board, so a marketplace median and a list price can both set a row's figure; the kind is marked next to the source.
- What it says: what one gigabyte of GPU memory cost per hour at the cheapest source, which is the right first comparison when memory decides whether a model runs at all.
- What it does not say: nothing about bandwidth, compute or interconnect, which differ by a factor of several between the classes, and nothing about availability. A cheap gigabyte on a card too slow for the job is not a bargain.
| GPU | Memory | Cheapest $/h | Source | $ per GB-hour computed — Computed: arithmetic over published inputs, which are shown with it. |
|---|---|---|---|---|
| RTX 3090 | 24 GB | $0.20 | Clore.ai median | $0.009 |
| RTX 4090 | 24 GB | $0.34 | RunPod list price | $0.014 |
| A100 | 40 GB | $0.73 | Vast.ai median | $0.018 |
| RTX 5090 | 32 GB | $0.68 | Clore.ai median | $0.021 |
| L40S | 48 GB | $0.79 | RunPod list price | $0.016 |
| H100 | 80 GB | $2.50 | Hyperstack list price | $0.031 |
| H200 | 141 GB | $3.59 | RunPod list price | $0.025 |
| B200 | 180 GB | $5.98 | RunPod list price | $0.033 |
Memory only. The classes differ in memory bandwidth by up to a factor of five and in compute by more, and a workload that needs several cards pays for interconnect and for the whole machine; none of that is in this table. Prices move during the day: check the live offer before renting.
Cloud list prices
What the clouds list per GPU-hour
AWS, Azure and Oracle Cloud, and the GPU clouds Lambda, Nebius, CoreWeave and Hyperstack, publish their prices without an account, so they are read here four times a day. A pay-as-you-go list price per GPU-hour: not a market median, and no statement about free capacity. The hyperscalers sell these GPUs as whole instances and Lambda in clusters of one to eight, so each row names the unit the price sits in.
| AWS from | $2.74p4d.24xlarge8 × 40 GB per instance$21.96/h for the whole instanceUS East (N. Virginia) +2 regions |
|---|---|
| Azure from | $3.40ND96asr v48 × 40 GB per instance$27.20/h for the whole instanceUS East +1 region |
| Oracle Cloud from | $3.05BM.GPU4.88 × 40 GB per instance$24.40/h for the whole instanceone list, no regions |
| Lambda from | $1.99A100 SXM1 × 40 GB per instanceone list, no regions |
| Nebius from | — not listed |
| CoreWeave from | $2.70A1008 × 80 GB per instance$21.60/h for the whole instanceNorth America +1 region |
| Hyperstack from | $1.35A1001 × 80 GB per instanceone list, no regions |
| vs. Vast.ai median | 3.3×AWS · 2 Nov |
| AWS from | $1.86g6e.xlarge1 × 48 GB per instanceUS East (N. Virginia) +2 regions |
|---|---|
| Azure from | — not listed |
| Oracle Cloud from | $3.50BM.GPU.L40S.44 × 48 GB per instance$14.00/h for the whole instanceone list, no regions |
| Lambda from | — not listed |
| Nebius from | $1.55L40S with Intel CPU1 × 48 GB per instanceone list, no regions |
| CoreWeave from | $2.25L40S8 × 48 GB per instance$18.00/h for the whole instanceNorth America +1 region |
| Hyperstack from | — not listed |
| vs. Vast.ai median | 2.3×AWS · 2 Nov |
| AWS from | $6.88p5.4xlarge1 × 80 GB per instanceUS East (N. Virginia) +2 regions |
|---|---|
| Azure from | $6.98NC40ads H100 v51 × 94 GB per instanceUS East +3 regions |
| Oracle Cloud from | $10.00BM.GPU.H100.88 × 80 GB per instance$80.00/h for the whole instanceone list, no regions |
| Lambda from | $3.29H100 PCIe1 × 80 GB per instanceone list, no regions |
| Nebius from | $4.50HGX H1001 × 80 GB per instanceone list, no regions |
| CoreWeave from | $6.16HGX H1008 × 80 GB per instance$49.24/h for the whole instanceNorth America +1 region |
| Hyperstack from | $2.50H1001 × 80 GB per instanceone list, no regions |
| vs. Vast.ai median | 2.4×AWS · 2 Nov |
| AWS from | $7.91p5en.48xlarge8 × 141 GB per instance$63.30/h for the whole instanceUS East (N. Virginia) +2 regions |
|---|---|
| Azure from | $10.60ND96isr H200 v58 × 141 GB per instance$84.80/h for the whole instanceUS East 2 +2 regions |
| Oracle Cloud from | $10.00BM.GPU.H200.88 × 141 GB per instance$80.00/h for the whole instanceone list, no regions |
| Lambda from | — not listed |
| Nebius from | $5.40HGX H2001 × 141 GB per instanceone list, no regions |
| CoreWeave from | $6.30HGX H2008 × 141 GB per instance$50.44/h for the whole instanceNorth America +1 region |
| Hyperstack from | $3.99H200 SXM1 × 141 GB per instanceone list, no regions |
| vs. Vast.ai median | 1.5×AWS · 2 Nov |
| AWS from | $14.24p6-b200.48xlarge8 × 179 GB per instance$113.93/h for the whole instanceUS East (N. Virginia) +2 regions |
|---|---|
| Azure from | — not listed |
| Oracle Cloud from | $14.00BM.GPU.B200.88 × 180 GB per instance$112.00/h for the whole instanceone list, no regions |
| Lambda from | $6.69B200 SXM68 × 180 GB per instance$53.52/h for the whole instanceone list, no regions |
| Nebius from | $8.50HGX B2001 × 180 GB per instanceone list, no regions |
| CoreWeave from | $8.60HGX B2008 × 180 GB per instance$68.80/h for the whole instanceNorth America +1 region |
| Hyperstack from | $6.00B2001 × 192 GB per instanceone list, no regions |
| vs. Vast.ai median | 1.6×Oracle Cloud · 2 Nov |
In plain words: the factor says how many times the Vast.ai price the cheapest big cloud listed for the same card on the same day.
How to read this
- We call this the cloud premium.
- What it is: the lowest list price per GPU-hour among AWS, Azure and Oracle Cloud, divided by the same day's Vast.ai median for the same card.
- Where the numbers come from: both sides are read by this site, the clouds four times a day from their public price lists, the marketplace every hour; the day is a UTC day.
- What it says: how much more a reserved, supported cloud instance cost than one GPU from a marketplace host on that day. The two are different products: the cloud comes with capacity guarantees, networking and support; the marketplace host is one machine somewhere.
- What it does not say: nothing about speed, reliability or availability, and nothing about tomorrow. The factor can rise because the marketplace got cheaper, not because the cloud got dearer.
The cloud premium divides the cheapest cloud list price per GPU-hour by the same UTC day's Vast.ai median, both measured here; a day without both readings has no factor. A list price buys a whole instance with reserved capacity and support, a marketplace median buys one GPU from one host, so the factor states what that difference cost on the day, not why, and never where it is going.
AWS: the lowest Linux on-demand price per GPU across the instance types in AWS's accelerator documentation and its public regions — an instance's hourly price divided by its documented GPU count. Read from the price files behind AWS's own on-demand pricing page, which carry the same publication as its bulk price list. GovCloud and the European Sovereign Cloud are left out.
Azure: the lowest Linux pay-as-you-go price per GPU across Microsoft's documented VM sizes and its public regions — a VM's hourly price divided by its GPU count from Microsoft's size documentation. Spot and low-priority prices and the US government regions are left out. The region shown is the first of those that list this price.
Oracle Cloud: the lowest pay-as-you-go price per GPU-hour among the class's bare-metal shapes on Oracle's public price list, which has no regions. The whole machine costs that price times its GPU count.
Azure and Oracle Cloud collected since 2 October 2026, AWS since 5 October, Lambda and Nebius since 19 October, CoreWeave since 24 October, Hyperstack since 27 October. A list price is its own kind of number: it stands beside the marketplace medians and is never blended with them. The cloud premium compares the three hyperscalers only.
AWS EC2 on-demand pricing ↗ (External link)Azure pricing ↗ (External link)Oracle Cloud price list ↗ (External link)Lambda GPU Cloud pricing ↗ (External link)Nebius pricing ↗ (External link)CoreWeave pricing ↗ (External link)Hyperstack GPU pricing ↗ (External link)
Radar Rent Index
What GPU rental prices did
One line per GPU class from the same hourly marketplace snapshots as the table: this week's median against last week's, and against the first week we collected. Described, never forecast.
| GPU | This week | vs. last week | vs. 30 days ago | Since start | Offers | Day band $/h |
|---|---|---|---|---|---|---|
| RTX 3090latest day Nov 2 | $0.25median of 7 days | −10 %down | — | +5 %since Oct 6 · 28 days on record | 64more offers than last week | $0.15–$0.30min–p75 of the latest day |
| RTX 4090latest day Nov 2 | $0.49median of 7 days | −36 %down sharply | +7 % | −15 %since Sep 18 · 46 days on record | 64supply steady | $0.32–$0.54min–p75 of the latest day |
| A100latest day Nov 2 | $0.86median of 7 days | 0 %flat | +12 % | +7 %since Sep 18 · 46 days on record | 40supply steady | $0.31–$1.01min–p75 of the latest day |
| RTX 5090latest day Nov 2 | $0.74median of 7 days | −23 %down sharply | +22 % | +10 %since Sep 18 · 46 days on record | 64supply steady | $0.47–$0.87min–p75 of the latest day |
| L40Slatest day Nov 2 | $0.80median of 7 days | 0 %flat | — | 0 %since Oct 6 · 28 days on record | 16more offers than last week | $0.44–$0.80min–p75 of the latest day |
| H100latest day Nov 2 | $3.08median of 7 days | +28 %up sharply | +19 % | −2 %since Sep 18 · 46 days on record | 6supply steady | $1.74–$3.40min–p75 of the latest day |
| H200latest day Nov 2 | $5.13median of 7 days | +5 %flat | +3 % | +11 %since Sep 18 · 46 days on record | 8more offers than last week | $3.76–$5.95min–p75 of the latest day |
| B200latest day Nov 2 | $8.75median of 7 days | −12 %down | +17 % | +8 %since Sep 18 · 46 days on record | 7supply steady | $6.25–$10.31min–p75 of the latest day |
No day was excluded.
A day is the median of its hourly medians, counting only hours with at least five offers (verified hosts on Vast.ai, free hosts on Clore.ai). A week is the median of its days and needs at least four of them. Words follow fixed thresholds: up or down from ±8 %, sharply from ±20 %; offers up or down from ±15 %. How the index is made →
Price history
Robust daily medians per GPU class: the median of each day's hourly medians, counting only hours with at least five verified offers. Days the Rent Index excludes as marketplace glitches are left out. Lines appear once a class has three days of readings.
Clore.ai marketplace
What Clore.ai prices did
The same rule applied to Clore.ai's hourly snapshots — hosts free to rent, priced per GPU-hour from their daily host price: this week against last week, and against the first week collected. A second market beside Vast.ai, never blended into it.
| GPU | This week | vs. last week | vs. 30 days ago | Since start | Offers | Day band $/h |
|---|---|---|---|---|---|---|
| RTX 3090latest day Nov 2 | $0.21median of 7 days | −23 %down sharply | — | 0 %since Oct 6 · 28 days on record | 115more offers than last week | $0.06–$0.29min–p75 of the latest day |
| RTX 4090latest day Nov 2 | $0.55median of 7 days | −11 %down | +8 % | −11 %since Sep 24 · 40 days on record | 126supply steady | $0.22–$0.75min–p75 of the latest day |
| A100latest day Nov 1 | $0.46median of 6 days | −35 %down sharply | +38 % | +38 %since Sep 29 · 35 days on record | 3supply steady | $0.32–$0.79min–p75 of the latest day |
| RTX 5090latest day Nov 2 | $0.71median of 7 days | −15 %down | −11 % | −12 %since Sep 24 · 40 days on record | 156more offers than last week | $0.21–$0.83min–p75 of the latest day |
| L40Scollecting | —collecting | —collecting | — | —collecting | —no comparison yet | — |
| H100latest day Oct 5 | —collecting | —collecting | — | —collecting | 4no comparison yet | $0.94–$2.50min–p75 of the latest day |
| H200collecting | —collecting | —collecting | — | —collecting | —no comparison yet | — |
| B200collecting | —collecting | —collecting | — | —collecting | —no comparison yet | — |
No day was excluded.
A day is the median of its hourly medians, counting only hours with at least five offers (verified hosts on Vast.ai, free hosts on Clore.ai). A week is the median of its days and needs at least four of them. Words follow fixed thresholds: up or down from ±8 %, sharply from ±20 %; offers up or down from ±15 %. How the index is made →
RunPod list price
What the RunPod price did
The same rule applied to RunPod's lowest posted on-demand price per class: this week against last week, and against the first week collected. A list price has no offers and no band, so those columns stay empty.
| GPU | This week | vs. last week | vs. 30 days ago | Since start | Offers | Day band $/h |
|---|---|---|---|---|---|---|
| RTX 3090latest day Nov 2 | $0.22median of 7 days | 0 %flat | — | 0 %since Oct 6 · 28 days on record | —no comparison yet | —min–p75 of the latest day |
| RTX 4090latest day Nov 2 | $0.34median of 7 days | 0 %flat | 0 % | 0 %since Sep 21 · 43 days on record | —no comparison yet | —min–p75 of the latest day |
| A100latest day Nov 2 | $1.00median of 7 days | 0 %flat | 0 % | −9 %since Sep 21 · 43 days on record | —no comparison yet | —min–p75 of the latest day |
| RTX 5090latest day Nov 2 | $0.69median of 7 days | 0 %flat | 0 % | 0 %since Sep 21 · 43 days on record | —no comparison yet | —min–p75 of the latest day |
| L40Slatest day Nov 2 | $0.79median of 7 days | 0 %flat | — | 0 %since Oct 6 · 28 days on record | —no comparison yet | —min–p75 of the latest day |
| H100latest day Nov 2 | $2.59median of 7 days | 0 %flat | 0 % | 0 %since Sep 21 · 43 days on record | —no comparison yet | —min–p75 of the latest day |
| H200latest day Nov 2 | $3.59median of 7 days | 0 %flat | 0 % | 0 %since Sep 21 · 43 days on record | —no comparison yet | —min–p75 of the latest day |
| B200latest day Nov 2 | $5.98median of 7 days | 0 %flat | 0 % | 0 %since Sep 21 · 43 days on record | —no comparison yet | —min–p75 of the latest day |
No day was excluded.
A day is the median of its hourly medians, counting only hours with at least five offers (verified hosts on Vast.ai, free hosts on Clore.ai). A week is the median of its days and needs at least four of them. Words follow fixed thresholds: up or down from ±8 %, sharply from ±20 %; offers up or down from ±15 %. How the index is made →
Break-even
How many rented hours equal the price of buying a machine outright. Rented accelerators are not like-for-like substitutes for a personal AI system — they are faster but temporary, and your data leaves your building — so read this as a budget comparison, not a spec comparison. Reference purchase: NVIDIA DGX Spark at its published MSRP of $4,699.
| GPU | Median $/h | Rented hours to match a purchase |
|---|---|---|
| RTX 3090 | $0.27 | 17,572 h≈ 42 years at 8 h/week |
| RTX 4090 | $0.51 | 9,293 h≈ 22 years at 8 h/week |
| A100 | $0.73 | 6,405 h≈ 15 years at 8 h/week |
| RTX 5090 | $0.74 | 6,382 h≈ 15 years at 8 h/week |
| L40S | $0.80 | 5,869 h≈ 14 years at 8 h/week |
| H100 | $3.08 | 1,528 h≈ 4 years at 8 h/week |
| H200 | $5.88 | 799 h≈ 23 months at 8 h/week |
| B200 | $9.99 | 470 h≈ 14 months at 8 h/week |
The purchase price is NVIDIA's published MSRP. It excludes electricity and, on the other side of the ledger, any resale value you would recover — owning is therefore somewhat cheaper than this table implies, renting somewhat less attractive at high hours. Hourly rates come from the collected price bands above.
Rent when
- + The job is occasional — a fine-tune, a batch run, one big evaluation.
- + You need far more memory than you own, briefly.
- + You want to try a model class before committing to hardware.
Buy when
- + You run models daily — owned hardware has no marginal cost.
- + Your data must not leave your machine.
- + Latency matters and you do not want to upload anything.
The Tower Report · every Monday
Monday's board, in your inbox.
Prices, the pick, what newly fits your card. Measured, never forecast. One mail a week, one click to leave.
Name your card and every issue opens with what it runs this week. You can change it by subscribing again with the same address.
Double opt-in; you confirm by one click. Details in the privacy policy · what is in an issue