Market snapshotB200 $7.21 −2.6%
AI COMPUTE
TRACKER

Market benchmarks / compute economics / market structure

The price of compute, tracked.

GPU rental markets, market-implied forward curves, chip economics, token prices, and the gigawatt buildout — one independent desk for the machines that make intelligence.

Market tape
H100 FLOOR$1.74/hrdecreased·B200 MEDIAN$6.69/hrincreased·H200 MEDIAN$4.34/hrincreased·B300from $6.25/hr·A100from $1.35/hrincreased·MI355Xfrom $2.59/hr·NVDA DC REV$75.2Bincreased·BIG-4 CAPEX ’26E$710Bincreased·GPT-4-CLASS TOKENS$0.18/1Mdecreased·FRONTIER PIPELINE~32 GW·

01Forward curves

What the market thinks compute will cost

The first market-implied term structure for GPU rental: spot today, monthly tenors through Jul ’27, and the shape of supply pressure ahead.

As of JUL 17 2026 · 22:59 UTCHow curves are derived

● Market-implied term structure

Implied forward curve

First point is implied spot; remaining points are monthly-average contracts — shown in $/GPU·hr or % vs spot

Kalshi monthly-average event ladders, settled to the Ornn hourly index. Implied level is the strike where bid/ask-mid probability crosses 50%; hollow points are thin-ladder estimates.

How to read this curve

The shape is the story. B200 and H200 trade backwardated; A100 sits in contango; H100 and RTX 5090 are flat, priced to hold through Jul ’27.

Observed

Provider list quotes and settled index values are printed market observations.

Inferred

Forward points are reconstructed from contract ladders and are not executable quotes.

Curve time machine

Replay the archived market-implied term structure.

Computed signals

The market in four numbers

Derived from the current datasets on every build — never hand-entered.

B200 specialist median
$6.69/GPU·hr
+16% y/yBlackwell demand outran supply
NVIDIA data-center revenue · FY27 Q1
$75.2B
+92% y/yNext-quarter guide $91B total
Tracked AI capex · 2026E
$780B
+66% vs ’25MSFT · GOOG · META · AMZN · ORCL
GPT-4-class token floor
$0.18/1M
−99.5% since Mar ’23DeepSeek V4 Flash · blended 3:1

Today’s read · market snapshot JUL 17 2026 · 22:59 UTC

Compute is becoming a commodity — the fourth factor of production — and for the first time it trades like one.

B200 repriced +45% since Jul 2. Its curve is backwardated: buyers pay a premium for delivery now, and the market prices the squeeze to ease. A100 is the anchor of the board: ~$1.05 through Jul ’27 on a contango curve.

01
B200 sets the tone.

Spot is $7.21; the Jul ’27 tenor settles near $5.65.

02
H100 rental prices are still easing.

The marketplace floor moved −3.3% over the latest three-quarter window to $1.74/GPU·hr.

03
Intelligence keeps repricing faster than hardware.

The GPT-4-class blended floor is now about 210× cheaper than launch at $0.18 per million tokens.

02GPU cloud

Renting an H100, 2023 → today

US$ per GPU-hour, on-demand. From $12 at peak scarcity to a stable ~$4 — and a marketplace floor near $1.74.

As of 2026-07-06All GPUs & providers

H100 SXM · on-demand price

Quarterly midpoints by provider class

Price-history chart awaiting interactive loadStatic chart frame · exact values in table view

Provider list prices, launch announcements, and archived snapshots; quarterly midpoints ±15%. Current quotes verified 2026-07-06.

03The macro picture

Compute demand is compounding

The frontier training-compute trend runs at 4.4× a year; tracked 2026E capex is +66% versus 2025.

Training compute of notable models

Total FLOP, log scale — estimates marked

Training-compute chart awaiting interactive loadStatic chart frame · exact values in table view

Epoch AI estimates and lab disclosures; estimated points carry wide error bars.

Big-tech capital expenditure

US$B per calendar year (2026 = guidance)

Capital-expenditure chart awaiting interactive loadStatic chart frame · exact values in table view

Company earnings releases and SEC filings; Microsoft includes finance leases; Oracle fiscal year mapped.

04Inference

Intelligence, repriced monthly

GPT-4-level capability became roughly 210× cheaper since launch. Then 2026 flagships reintroduced a frontier premium.

Cheapest GPT-4-class · blended

$0.18/1M

−99.5% since Mar ’23 · DeepSeek V4 Flash

OpenAI flagship · GPT-5.5

$11.25/1M

+227% vs GPT-5 · Flagship list price rose in Apr 2026

Batch & cached reality

−50 / −90%

Batch APIs ~50% off list; cached input ~90% off — production pays below list

05Quick answers

The methodology, without the footnotes

The three definitions readers need before comparing the curves, provider classes, or token-price series.

How are the GPU forward curves derived?

From Kalshi event-contract ladders on GPU rental prices, settled to Ornn's hourly index. The implied level for each tenor is the strike where the ladder's yes-probability crosses 50% — the market's median expectation, not an executable quote. Thin back months are estimates.

What does the blended token price mean?

Blended price = (3 × input + output) ÷ 4 per million tokens, reflecting a typical 3:1 read/write mix at standard list rates.

What are the provider classes?

Hyperscaler is undiscounted big-cloud list price; specialist is the AI-native cloud median; marketplace is the open-market floor on community hardware.

06Go deeper

Explore the full desk