# Chutes: Bittensor's Revenue Machine, Subsidised

> Chutes is one of Bittensor's largest-revenue subnets (SN64). $1.3-2.4M ARR by Pine Analytics' estimate, against a far larger TAO emission subsidy. The OpenRouter contradiction explained.

- Published: 2026-04-10 · Updated: 2026-10-06
- Canonical: https://ownyourmind.ai/tokenomics/chutes-bittensor-revenue-machine/
- Cite as: Own Your Mind (ownyourmind.ai), licence CC BY 4.0. Link the canonical URL.

Charts and interactive widgets are omitted here; project data is at https://ownyourmind.ai/api/scores.json.

---

## The thesis

Chutes is the strongest revenue story Bittensor has. It's also losing money on every customer who uses the product, and the team has said so themselves.

The first part is what you read in the headlines. Chutes runs serverless inference for 50+ open-source models, ships an API that competes head-to-head with Together.ai and Replicate, and gets routed to by default on OpenRouter. Shipped product, paying users, external revenue.

The second part is what doesn't make the headlines. In February 2026, the team published a [community announcement](https://chutes.ai/news/community-announcement-february) that quietly killed the free tier and explicitly named the problem: heavy users were extracting "56x to 324x" their subscription value at equivalent pay-as-you-go pricing. Translation: paid users were getting more compute than their fees covered, by a factor of 50 to 300, and the gap was being filled by TAO emissions.

This article looks at both sides. Chutes is the most important Bittensor subnet to understand because it's the test case for whether the entire subnet model can produce a sustainable business or whether it produces subsidised cloud services that collapse when the subsidy stops. The next halving (~late 2029) is when the subsidy steps down materially; until then the runway holds, but the underlying unit economics still need to close.

## What it actually is

Chutes is a serverless inference platform. You hit an API endpoint, you get back a model response, you get billed per token. The service runs on top of Bittensor subnet 64, which means the underlying compute comes from miners running Chutes' validator-evaluated workloads in exchange for TAO emissions.

From a developer's perspective it looks like any other inference API. From the network's perspective it's a market: miners compete on quality and uptime, validators score them, the highest performers earn the TAO emissions allocated to subnet 64. At the time of Pine Analytics' March 2026 analysis, roughly 14.39% of all TAO daily emissions flowed to this subnet, which works out to approximately 518 TAO per day. That emission stream dwarfs the subnet's external customer revenue, and it is what funds the gap between what users pay and what the compute costs.

The model catalogue is broad. Chutes' [public model list](https://llm.chutes.ai/v1/models) includes Llama 3.2 and 3.3, DeepSeek V3.1, Qwen3 (4B through 235B), Mistral Small and Nemo, Kimi K2.5, GLM-5, Gemma 3, plus diffusion models, embeddings, speech, and 3D generation. The pricing undercuts centralised alternatives:

These numbers are pulled live from Chutes' API and Together.ai's pricing page. The discount holds up against published rates. The question is what funds it.

## The OpenRouter contradiction

Chutes claims 160 billion tokens per day across all surfaces. Rayon Labs has been citing this number consistently since March. It's also the foundation for the "approaching $10M ARR" claim that started circulating in early April.

The verifiable number tells a different story. Chutes is a [provider on OpenRouter](https://openrouter.ai/provider/chutes), a widely used open inference router. OpenRouter publishes daily token throughput data per provider, which is one of the few ways to externally verify a Bittensor subnet's actual usage. Chutes' OpenRouter throughput peaked on 7 February 2026 at roughly 42 billion tokens per day. By late March it had declined to 8-12 billion per day. In April 2026 it showed approximately 6.8 billion tokens per day across the 18 models Chutes served on the platform, and over the latest 89-day window it averaged 7B a day. (figures as of 2026-10-12)

Chutes also published its own daily revenue, keylessly, until mid-2026, then withdrew the series. We rebuilt it from the Wayback Machine. Over its last 90 published days it averaged $15.6K a day, peaking at $22.5K on 25 March 2026, and part of it through February was sponsored hosting paid by model makers. Those are Chutes' own numbers, but they put its revenue well under the "approaching $10M ARR" in circulation. (figures as of 2026-05-17)

Three possibilities, in order of charity:

1. **OpenRouter is a small fraction of Chutes' total volume.** Direct API customers and the chutes.ai web interface dwarf the OpenRouter traffic. If that's true, the team should publish a dashboard.
2. **The peak happened, the decline happened, the 160B figure is stale.** Chutes had a moment in February when one model (TNG/DeepSeek R1T2 Chimera) was getting heavy default routing on OpenRouter. That model has rolled off. The April 6.8B figure was the new baseline.
3. **The 160B number was never accurate.** It was an aspirational high-water mark that became a marketing line.

I'm not in a position to say which of these is right. Neither is anyone outside Rayon Labs. What I can say is that the "explosive growth" narrative from March doesn't match the publicly verifiable February-to-April trajectory, and any article that cites the 160B number without flagging this gap is misleading the reader.

## The subsidy economics

This is the part that matters more than any single throughput number.

**Update, 17 August 2026.** Every emission figure below is Pine's March 2026 measurement and the network has moved since. Our own chain index read Chutes at 6.55% of emissions on 15 August 2026 and 5.82% on 22 September 2026, well under half the 14.39% this piece was built on,  and TAO trades below the price the dollar conversions used. The shape of the argument holds, a revenue-light subnet carried by emissions, but the subsidy is smaller than these numbers imply. We have not restated the ratio: deriving TAO per day from our index would need a network emission denominator it does not carry.

[Pine Analytics published a bear case](https://pineanalytics.substack.com/p/the-bear-case-for-bittensor-tao) on Bittensor on 24 March 2026. Their estimate of Chutes' actual external customer revenue is $1.3-2.4 million annualised.  Against that, the subnet received roughly 518 TAO per day in emissions. Pine put the subsidy at 22 to 40 times customer revenue. Customer payments fund a small fraction of operations. TAO inflation funds the rest.

The pricing math follows directly. Pine calculates that if Chutes had to cover its costs from customer revenue alone (without emission subsidy), the break-even price would be approximately $1.41 per million tokens. Together.ai charges $0.88 per million tokens for Llama 70B. Unsubsidised Chutes would be 1.6 times more expensive than the centralised alternative, not cheaper. The cost advantage that makes Chutes attractive depends entirely on the emission subsidy continuing at its current rate.

The team confirmed this directly in February. The [community announcement](https://chutes.ai/news/community-announcement-february) eliminated the free tier (200 requests per day for roughly a year) and pulled four frontier models (GLM-5, Kimi K2.5, Qwen 3.5, MiniMax M2.5) out of the Base subscription tier. The stated reason: "heavy users were extracting 56x to 324x their subscription value at equivalent pay-as-you-go costs." That's not Pine Analytics talking. That's Chutes' own blog post.

Most projects don't admit this. Chutes did. Credit where it's due. The honesty doesn't fix the economics, but it does mean the team is aware of the problem and trying to address it.

## The next halving

Bittensor's emission schedule halves on a Bitcoin-style cadence. Halvings are supply-threshold events: each halving fires when total TAO issuance crosses the next 50% of remaining cap. The first halving fired at 10.5M TAO on 14 December 2025, dropping daily emissions from 7,200 to 3,600 TAO. The second halving fires at 15.75M TAO and is projected for approximately 11 December 2029, dropping daily emissions from 3,600 to 1,800 TAO. Recycled TAO (transaction fees and a few protocol charges) returned to the emission pool can push that date later but never earlier.

What that means for Chutes specifically: the subsidy halves on a fixed schedule, and the question is what revenue trajectory gets the subsidy ratio toward break-even by the time it fires. Holding the TAO price and Chutes' emission share (14.39%) constant, here's where different revenue growth rates land:

Two readings of this table. The optimistic one: doubling revenue every year for 3.5 years (or 35% YoY from the self-reported base) gets Chutes to structural break-even at the next halving. Together.ai cleared faster growth than that during early traction, so the trajectory is achievable for an inference business with product-market fit.

The pessimistic one: Pine's estimated base is $2M, not $10M. If the self-reported number is closer to marketing than measurement, the realistic starting point is the lower bound, and 100% YoY for 3.5 years is an aggressive ask for any SaaS business in any market. Even hitting it leaves a 1.3:1 ratio at the halving rather than a clean break-even, and the third halving (~late 2033) compresses the subsidy again.

Either way the runway holds. The harder question is structural: whether decentralised serverless inference has a reason to be cheaper than centralised serverless inference at scale, or whether the subsidy was the entire price advantage. If the latter, the runway buys time while the unit economics stay broken. The answer to that question is what the next 12-18 months of growth trajectory will tell us.

## The team and the entity question

Chutes was founded by Jon Durbin. Before Bittensor he was known as the author of [Airoboros](https://github.com/jondurbin/airoboros), a self-instruct fine-tuning framework that's been widely used in the open-source LLM training scene. Verifiable technical credentials, established public reputation, no anonymous founder problem at the individual level.

Rayon Labs is a different question. Rayon Labs is the umbrella name for the team that runs Chutes (SN64), Gradients (SN56), and Nineteen (SN19). Combined, those three subnets captured roughly 23.7% of all TAO daily emissions in early 2026, nearly a quarter of Bittensor's output flowing to a single team. Our 22 September 2026 chain read puts the three at 7.97%.

I went looking for the corporate entity. There isn't one I could find.  No Delaware C-corp registered to anyone named Jon Durbin or under "Rayon Labs." No Cayman entity. No UK Companies House filing. No Crunchbase or PitchBook profile. No disclosed funding round. The only published company description is "independent teams collaborating," which isn't really an entity. The only public GitHub org has one repo with two stars, while the actual Chutes platform code lives at [github.com/chutesai/chutes](https://github.com/chutesai/chutes), a separate org. The published Chutes repository contains the client SDK only. The miner, validator, and platform code are private.

This isn't necessarily a problem in isolation. Plenty of legitimate crypto projects operate as informal teams without traditional corporate structure. But the combination matters: emission concentration that reached 23.7% in early 2026, no disclosed legal entity, pseudonymous co-founders (Namoray is publicly known only by handle), and a closed-source core inference platform. For sovereignty-first readers, those are four risks compounding.

The bull case is that Jon Durbin's track record speaks for itself, the Airoboros work is well-regarded in the open-source ML community, and the team has consistently shipped product. The bear case is that a team with no published entity, no disclosed governance, and no formal accountability mechanism has drawn as much as 23.7% of Bittensor's economic output. Both are true at the same time.

## What to actually watch

Five leading indicators over the next 12-18 months tell you whether Chutes is on the trajectory to grow into the subsidy or whether the subsidy is the trajectory. None of these are demands; they're the signals that move the read on this project meaningfully.

1. **OpenRouter throughput trend.** The verifiable number. If the Feb-Apr decline reverses and throughput trends up consistently across multiple models, the volume narrative starts becoming credible. If it stays flat or keeps declining, the 160B self-report is increasingly hard to defend.
2. **Customer revenue disclosure.** Anything that closes the gap between Pine Analytics' independent $1.3-2.4M estimate and the "approaching $10M ARR" figure, which traces to a media post rather than Rayon Labs. A public dashboard. An audited number. A revenue range disclosed in a quarterly update. Any of those move Chutes from "trust the team" to "verify the team" on the metric that matters most.
3. **Pricing power.** Can Chutes raise prices toward the unsubsidised break-even of ~$1.41 per million tokens without losing customers? That's the cleanest test of whether the value is in the product or the price. A successful price increase says the customers are there for the inference quality. A failed one says they're there for the discount.
4. **Customer concentration.** If 80% of Chutes revenue comes from one OpenRouter integration or one large enterprise customer, that's a different risk profile than 80% coming from thousands of individual developers. Disclosure of the concentration distribution is itself a signal.
5. **Rayon Labs entity formalisation.** Whether the team publishes a corporate entity, jurisdiction, and accountability process over the next 12 months signals whether they're building for institutional revenue (which requires it) or staying informal indefinitely. The emission concentration, 23.7% in early 2026 and still the largest of any operator in our index, makes this more important than for any individual subnet.

Together.ai publishes most of these. So does OpenRouter. So does almost every centralised inference provider Chutes competes with. That none are available today is itself information; whether they become available over the next 12-18 months is information of a different kind.

## The honest assessment

Chutes is the most interesting thing happening on Bittensor right now. It's the clearest case of a subnet with a product that working developers use, revenue from external customers, and a team that's been honest about the problems.

It's also losing money on every customer at current prices, the volume narrative doesn't match the publicly verifiable data, and it sits on top of a large TAO emission subsidy (~518 TAO/day on Pine's March 2026 measurement)  that halves at the second Bittensor halving (~late 2029). The team will need to grow customer revenue roughly 15x from Pine's estimated base to reach sustainability by the time that subsidy steps down.

I think Chutes will get there mechanically: ~3.5 years of runway and a competent team is enough time to grow into the subsidy if the underlying inference business stands on its own. The harder question is what happens at the third halving (~late 2033) and beyond. Each halving makes the math harder, not easier, and there's no version of decentralised serverless inference where the structural cost is meaningfully below centralised serverless inference. The advantages of decentralisation in this category are sovereignty and censorship-resistance, not unit cost. The runway holds; the structural question is unchanged.

For [Bittensor](https://ownyourmind.ai/projects/bittensor/) holders, Chutes is the bet that one subnet can build a sustainable business inside the network. If it works, the entire subnet model gets validated. If it doesn't, the network's strongest revenue story turns out to have been a subsidy story all along. The next 18 months will tell us whether revenue growth tracks toward the 15x target or stalls; the 2029 halving is where the verdict gets enforced.

For the broader Bittensor revenue picture, see [Bittensor subnets: where the revenue actually is](https://ownyourmind.ai/tokenomics/bittensor-subnets-where-the-revenue-is/). For how to actually buy and stake into subnet 64, see our [Bittensor subnet staking guide](https://ownyourmind.ai/use/bittensor-subnet-staking/). For the underlying token economics, see [dTAO subnet economics](https://ownyourmind.ai/tokenomics/bittensor-dtao-subnet-economics/).

## Score change log

| Date | Score | Change | Reason |
| - | - | - | - |
| 2026-10-06 | Data | N/A | "Approaching $10M ARR" re-attributed: it traces to an Altcoin Buzz post, not Rayon Labs. Added Chutes' own daily revenue series, withdrawn mid-2026 and rebuilt from the Wayback Machine. Scores unchanged. |
| 2026-09-24 | Data | N/A | Rayon Labs' 23.7% emission share dated to early 2026 beside our 7.97% chain read, Chutes' current 5.82% share and the measured OpenRouter throughput added, and three unverified superlatives softened. Scores unchanged. |
| 2026-05-06 | Editorial | N/A | Deeper editorial reframe completed same day as the data correction below. Subsidy economics section rebuilt with multi-year scenario projections (flat / 50% YoY / 100% YoY / 35% YoY-from-self-report) showing where each lands on the subsidy ratio at the second halving. "What I'd want to see" reframed as "What to actually watch", with five leading indicators over 12-18 months rather than a list of demands. Position shifted from "acute deadline risk" to "structural concern with multi-year runway." Returns score impact deferred to June monthly review. |
| 2026-05-06 | Data | N/A | Halving date corrected throughout: original article repeatedly stated the next Bittensor halving was approximately 14 December 2026 with an "eight months" deadline framing. Bittensor halvings are supply-threshold events on a Bitcoin-style ~4-year cadence; the second halving is projected for approximately 11 December 2029 ([bittensor-halving.com](https://www.bittensor-halving.com/), corroborated by [docs.learnbittensor.org/concepts/halving](https://docs.learnbittensor.org/concepts/halving)). Frontmatter `seoTitle`/`description`, table labels, and conclusion timing all updated. Subsidy-to-revenue ratio analysis, OpenRouter contradiction, unit economics and entity questions unchanged. |
| 2026-04-10 | Both | N/A | Initial publish. |