DEV · local workspace

Skip to content

Company

The product

Megatron Markets provides infrastructure that catalogues the inference capacity providers publish, standardise it into instruments both sides can evaluate, and supports matching through an order book when the selected instrument and desk are enabled for new exposure, with account-scoped operating controls.

The product catalogues AI-model capacity as reference data and defines connected order-book, RFQ and enquiry workflows. It records terms, orders, fills and participant-provided capacity evidence.

414
Models observed
102
Inference providers observed
1,505
Provider endpoints observed
28
Providers on the most observed model

Reference observation of provider-published rates. Read at 17 August 2026, 17:19 UTC.

Instrument

Decisions behind the instrument

Workload-shaped bundles
A workload consumes prompt, cache-read, cache-write and output tokens in its own ratio. A bundle names those weights explicitly and gives both sides a common unit to evaluate.
Two durations on a forward
The tenor states how far ahead a named window opens; the window states the future period represented by the planned instrument. Separating the two lets desks compare forward-dated prices without hiding their time shape.
Provider-level price discovery
The reference catalogue keeps model identity separate from provider endpoint. A desk can see prompt, output and cache rates, quantisation and spread before deciding what to investigate.
Several connected product paths
The product connects a central order book, requests for quote and desk enquiries. Orders, fills and cancellations share one attributed workflow.
Electricity as research context
Electricity markets offer useful ideas about capital-intensive, perishable capacity and time-shaped demand. The comparison is structural only: Megatron Markets is preparing AI-model capacity infrastructure, not electricity, energy contracts, grid rights or power delivery.
Evidence with a timestamp
Public catalogue figures are labelled with the moment they were read. They preserve the provider-published rate, endpoint identity and observation time together.

Research and market references

Available data, explicit context

Product data

A signal-derived demand gauge

The terminal exposes an AI demand gauge computed in timestamped buckets from classified source items. It reports the contributing signal count and category components. It is informational research, alongside its source context. Read the signal methodology and limitations.

Market references

Provider references and governed indices

External product references are kept separately identified from Megatron Markets instruments. The documentation describes the source, identity and comparison method for each reference.