Targon is a decentralized compute network of trusted execution environments.
SN 4. Rank 7 of 32 by emission
Targon
Emission4.7%of the block reward · rank 7 of 32
Participants5 miners · 7 validators
Registration0.0005 TAO
Verified2026-09-09links checked and scores merged
The programcompute · github.com@511128a · read 2026-09-19
The work
Attested GPU cards. A miner installs TargonOS, a signed operating system that wipes the machine and attests it to the network on every boot, and keeps confidential virtual machines available for rent. What is paid for is a live, verified card inside an auction class, whether or not a customer is using it. There is no request or response the miner answers.
code · raw.githubusercontent.com · docs/miner/miner.md
Miner emissions are paid per card based on each auction's target price and the amount of compute live on the network
How it is scored
No quality score. Each auction, fetched from the Tower service, names target cards, a target price and a cap in cents per card hour. The pool is target price times target cards; each verified card earns the lesser of pool divided by live cards and the cap, times 1.2, as a fraction of the dollar value of the miner emission pool, itself computed from chain emission and the TAO price.
code · raw.githubusercontent.com · internal/validator/callbacks/weights.go
perMiner := min(pool/nodes, aucInfo.MaxPrice); minerIncentive := (float64(perMiner) * 1.2 * float64(bid.Count)) / (*c.EmissionPool * 100)
How the pay splits
Proportional to verified cards in each auction, capped per card. Weight no auction covers is burned: split among the burn keys Tower names, remainder to uid 28. If payouts exceed the pool, every weight scales down pro rata. A node reporting fewer cards than the auction's min_cluster_size, 8 for the H200 and B200 classes in the docs example, earns nothing from that auction.
code · raw.githubusercontent.com · internal/validator/callbacks/weights.go
forBurn := max(setup.U16MAX-sumScores, 0); finalUids = append(finalUids, uint16(28))
Who judges
Seven validators were active on the 2026-09-09 chain snapshot, against five miners. Commit-reveal is on, weights at most once per 100 blocks. A validator re-attests each miner node during the interval, counts its GPU cards against the auction's cluster minimum, prices it from Tower's auction table, and posts the weight extrinsic through a weights-api sidecar. Nothing about the compute's quality is judged, only its presence and attestation.
chain · hyperparams-2026-09-19.json
commit_reveal_weights_enabled True, weights_rate_limit 100, max_validators 64; chain-2026-09-09.json active_validators 7
How often it turns
Emission settles every 360 blocks, about 72 minutes, and the validator works in the same interval: it fetches auctions and the emission pool at the start, polls miner nodes every 30 blocks, and sets weights once per interval. A new registration is immune from deregistration for 7,520 blocks, about 25 hours, half again the usual 5,000.
chain · hyperparams-2026-09-19.json
tempo 360, immunity_period 7520
Who buys
Renters of confidential GPU and CPU machines by the hour, through the site and an unauthenticated inventory API. Prices run from $0.45 an hour for one RTX 4090 to $60 for an eight-card B300 class. On 2026-09-19 the inventory listed 33 SKUs, of which 7 showed any availability. No customer is named; the front page shows a logo strip and counters that read zero.
own docs · api.targon.com
"name":"rtx4090-small", "type":"rental", "cost_per_hour":0.45, "available":0
What has been gamed
Free compute drew abuse. The May 2025 post says subsidised inference on other companies' models attracts a queue that leaves when prices rise, so emission will not fund it. The guard in code is attestation: only nodes that pass TargonOS attestation and report at least the cluster minimum enter an auction, and weight no auction covers burns rather than paying idle registrations.
on record · manifold.inc · sn4-010
If we opened a gas station and cut our prices to $0, we would expect a line of cars
5 of 7 fields rest on the chain or the code. A manifest is the site's reading of the subnet's program, not a review; each field links its source. All programs · What a field means · Correct a field
What it says
The subnet in its own words: the line the index took from its front door, and what the team has said on the record. All 15 quotes in the map.
Their subnet
Targon is a compute marketplace in hyperscaler territory that refuses free gas and aims to fund every miner dollar with revenue.
Our aim is to move away from a plateau curve for the subnet, opting instead for an exponential curve where competition is not based on registration.
As of writing, the miners on Targon can afford approximately 1,400 H200s a day. This places us in hyperscaler territory.
If we opened a gas station and cut our prices to $0, we would expect a line of cars that stretched further than the eye could see.
We are working to create a market where every $1 that a miner receives is funded by $1 in organic revenue.
Introducing Targon Order Book: A new market mechanism enabling miners to set ask prices for their H200s, creating a more efficient and transparent compute marketplace with predictable hourly rates.
Bittensor, in their words
Manifold uses the foundation's marketplace-of-digital-commodities frame for Bittensor, but pairs it with open criticism of the network's tooling and miner economics, and a vision of a decentralised frontier lab.
Bittensor is a powerful platform that seems to be limiting its growth by forcing applications to be locked into one stack.
There is currently no standard/language-agnostic format for senders or receivers. This keeps the ecosystem trapped into using Python for almost anyone who wants to build on Bittensor.
Introducing Epistula v2: The first truly open source networking protocol in Bittensor history, designed for secure, efficient, and language-agnostic communication.
Our mission is to make Bittensor that kind of protocol. And to do that, we must think bigger.
Frames market network marketplace platform gate curve gas station hyperscaler lab order book
What it shows
Explained from zero by the index, from the subnet's own materials, for someone who has never heard of Bittensor.
SN4 Targon rents out computing machines by the hour, like a hotel whose rooms belong to strangers, with each room sealed so that not even the owner can see what the guest is doing inside.
The commodity, explained from zero
The room is a GPU, the graphics processor that does most AI computing. Targon lists three products: rentals (a GPU server a customer logs into), virtual machines, and Targon Virtual Machines, its own product. The last is the point of the subnet. A TVM runs inside a trusted execution environment, a locked box in the processor that encrypts memory and can prove to an outsider what ran inside, so the machine's owner cannot read the guest's data. Intel and Targon co-wrote a whitepaper on this, dated March 23, 2026.
Prices exist. The inventory API, which needs no login, quotes an H200 at $2.49 an hour in its example, and the live API returned $0.45 an hour for an RTX 4090 on September 9. The customer named on the site is a team doing "AI Training and Deployments". Whether anyone rents today is harder to see: the front page's counters for GPUs, uptime and utilization read zero to a fetcher, and every listed machine showed "Out of Stock" on the day of checking. A logo strip names OSS Capital, DCG, Google, Shopify and Bittensor without saying why.
Why it is on Bittensor at all
The mission page's claim is price: "a non-interruptible cloud experience at interruptible prices". Because supply can come from anyone with the right hardware, the whitepaper argues prices become "transparent, market-driven". The sealed room is what makes renting from strangers safe. The comparison breaks here: a hotel guest trusts the hotel, while a Targon guest trusts Intel's chip and a cryptographic proof, and the hotel itself never sees the guest's luggage. No number on the site compares Targon's price to a large cloud's.
How the work gets done
Miners (the room owners) install TargonOS, a signed operating system that wipes the machine, attests to the network on every boot, and re-attests roughly every 72 minutes. The subnet runs auctions: each names a target number of cards, a target price and a ceiling in US dollars per card per hour, and "while an auction is below its card target, every live card earns up to the payment cap". Validators (the judges) hold stake and set the weights that turn those rates into emission, the subnet's share of newly minted TAO; the docs describe the miner side in far more detail than the validator side. Emission not claimed by any auction is burned.
How you would know it works
The public inventory endpoint at api.targon.com/tha/v3/inventory returns live machines, specifications and hourly prices with no login. It is a price list, not a usage figure; the materials do not publish one.
What is missing
The chain contact devs@manifold.inc appears nowhere on the site or docs, which offer only a contact form and Discord. The Discord field on chain is empty. There is no status page (status.targon.com does not resolve), no SLA, and no stated response time. The miner docs give no scoring formula and no deregistration policy, and the registration cost is not stated (0.0005 TAO on chain today). The chain shows 5 active miners and 7 validators.
Go deeper
Sources for this explainer
Metaphor: a hotel of sealed rooms owned by strangers. Every claim is drawn from the evidence set or the subnet's own materials; "(inferred)" marks a conclusion rather than a quote. Corrections.
How legible it is
Four audiences, four questions each, rated on what a first-time reader can find in five minutes. The scale measures findability, not quality.
On-chain identity 5/7- subnet_name
- Targon
- github_repo
- github.com
- subnet_contact
- devs@manifold.inc
- subnet_url
- targon.com
- discord
- not set
- description
- Incentivized Compute Marketplace powered by the Targon Virtual Machine (TVM).
- additional
- not set
Stakers and validators●●●●● 3.0
Should I allocate here? · rank 18 of 32 for this audience
Q1What it is●●●●● 4
Scale Workloads with High Speed GPU & CPUs for AI Training and Deployments
with rentals, VMs and pay only for compute you use, hourly prices sit on the inventory page and public API, matches chain
Q2Who it is for●●●●● 2
Powered By OSS Capital DCG Google Shopify Bittensor
logo strip, GPU, uptime and utilization counters read 0 to a fetcher, stats page is JS rendered, every SKU Out of Stock, no dated usage figure
Q3How it resists gaming or fails●●●●● 3
emissions paid per card up to a cap against auction targets, unallocated emissions burned, TargonOS attests on boot, written for miners, no anti gaming section, no failure modes
Q4Identity and documentation●●●●● 3
identity 5 of 7 (discord and additional unset), github and url resolve, a validator setup page, chain contact devs@manifold.inc nowhere on the site, no staker or tokenomics page
Miners●●●●● 3.3
Can I compete, and what wins? · rank 18 of 32 for this audience
Q1What it is●●●●● 3
Mining on Targon runs on TargonOS, a signed, purpose-built operating system that provisions the machine end to end.
TDX GPU nodes for confidential VMs, no request or response spec
Q2Who it is for●●●●● 3
All prices are in USD per hour per card, reported in cents
with H200 and B200 auction targets and min_cluster_size 8, targets page JS rendered, registration cost not stated (burn 0.0005 TAO)
Q3How it resists gaming or fails●●●●● 3
per card rate settles toward the auction target and never exceeds max_price, attestation gates participation, no scoring formula, no immunity or deregistration policy
Q4Identity and documentation●●●●● 4
TargonOS install guide plus config.json and docker compose steps, repo pushed 2026-08-23, no tagged releases or changelog, docs pages undated
Buyers and enterprises●●●●● 3.5
Can I use this today? · rank 10 of 32 for this audience
Q1What it is●●●●● 4
Query available compute resources, their specifications, pricing, and current availability. This endpoint does not require authentication.
Rentals and VMs via an API, every SKU Out of Stock today
Q2Who it is for●●●●● 4
cost_per_hour
: 2.49 for an H200 in the example, the live API returns 0.45 per hour for an RTX 4090, customer type is AI training and deployment teams, no pricing page
Q3How it resists gaming or fails●●●●● 2
a contact form and Discord, no status page (status.targon.com does not resolve), no SLA, no response time, an api/health endpoint in the docs
Q4Identity and documentation●●●●● 4
REST reference for workloads, volumes, inventory and tokens, chain contact devs@manifold.inc not on the site, the site offers a contact form only
Newcomers●●●●● 4.0
What is this and why does it matter? · rank 5 of 32 for this audience
Q1What it is●●●●● 3
Targon is a decentralized compute network of trusted execution environments.
One sentence on the front page, TEE, TDX and confidential compute left unexplained, consistent with the chain description
Q2Who it is for●●●●● 4
non-interruptible cloud experience at interruptible prices, without compromising on the security our clients depend on
price and security set against centralized giants, no numbers
Q3How it resists gaming or fails●●●●● 4
Intel co authored whitepaper dated 03.23.2026 with a PDF, live inventory API with prices, front page counters read 0 to a fetcher, no benchmark of the network output
Q4Identity and documentation●●●●● 5
subnet_name Targon matches the site, description is a full sentence, url resolves, the mission page serves as about
Provenance
How this rating came to be. Verified means the scorer fetched the page and the words are on it; secondhand means concluded from code, absence, or a third party. The link badge is a separate automated check made before publication.
scored by agent c 2026-09-09, rubric v1.0merged 2026-09-09links verified 2026-09-09: 41 live, 0 unreachable, 1 manual
Something wrong? Corrections of fact are applied as they arrive during the window; rating disputes are batched at its close. How to file one · GitHub issue · email.