Render’s Forth Vote: Can a Gateway Build AI Demand?

A Live Vote, an Old Question

This week, Render Network token holders are being asked to settle an argument the project has been having with its own price for most of 2026. The proposal in front of them is RNP-024. On the GitHub index it carries a dry name, a growth and adoption incentive program, but the product underneath it is ambitious. It is called Forth, and it is Render’s attempt to build the front door through which artificial intelligence demand enters the network, pool that demand in one place, and settle it in the RENDER token.

The timing is pointed. Render spent the late summer and early autumn as one of the strongest performers in the decentralized physical infrastructure sector, known as DePIN, and by one widely circulated accounting it closed the third quarter as the single highest revenue AI token in crypto. Yet RENDER changes hands near $1.52, roughly 89% below the $13.53 all-time high it set in March 2024, according to CoinGecko. The grid is busy, the headline numbers are large, and the token has gone sideways for the better part of two years.

The vote itself is the second of two stages. An initial 72-hour community vote cleared in late October; the binding final vote, run on Render’s Nation governance platform, opened on November 3 and runs for six days, needing at least half the votes in favor and a quorum of 15% of eligible supply to pass. A yes would authorize the Render Foundation to fund Forth’s rollout and its usage incentives from the treasury. It would not, by itself, ship a product. But it would commit the network to its clearest answer yet to the question that has dogged it: can Render manufacture demand for its own token rather than wait for the market to supply it, and if it can, will usage and price finally move together?

What Render Actually Is

Render began as a side effect of a rendering company. OTOY, founded in 2008 by Jules Urbach, builds OctaneRender, a GPU-accelerated path-tracing renderer that launched in 2012 and became a staple inside tools like Cinema 4D, Blender and Unreal. Rendering a complex 3D scene is embarrassingly parallel and brutally GPU-hungry, and most of the world’s GPUs sit idle most of the time. Render Network, launched in 2017, was the marketplace built to connect the two: artists and studios submit jobs, independent node operators run them on spare hardware, and the work is paid for in the network’s token.

The token started life as RNDR on Ethereum and migrated to Solana as RENDER, swapped one for one. CoinGecko notes that the legacy RNDR contract on Polygon was deprecated on July 18, 2025 following unauthorized access to that contract, a reminder that even a mature token carries operational history. Governance and the token now sit with the Render Foundation, a Cayman Islands entity spun out of OTOY in 2023.

What has changed most is what the network renders. Classic 3D work still runs, but AI compute, and inference in particular, has grown from under 10% of activity a couple of years ago to somewhere around 35% to 40% now, according to Crypto Briefing. The network reports more than 82 million cumulative frames rendered across roughly 5,600 active GPU nodes. Inference is a natural fit for a scattered grid: unlike training a frontier model, it does not need thousands of GPUs wired together in one room, so it can run on consumer cards spread across 180 countries.

That software lineage is the quiet advantage. Akash, io.net and Nosana rent raw GPUs; Render arrives with OctaneRender, a renderer that artists already use, and a pipeline that knows how to slice a scene into jobs, distribute them and stitch the results back together. Recent Octane releases have pushed into relightable Gaussian splats and neural rendering, the techniques behind a lot of 2026’s most striking motion work. Owning both the tool and the marketplace means Render is not just selling compute by the hour; it is selling a workflow, which is far harder for a commodity GPU bazaar to copy.

The Paradox: Record Revenue, a Sinking Token

Here is the paradox in two numbers. In late October, TokenPost published a third-quarter revenue ranking of AI-crypto projects, and Render sat at the top with about $23.3 million, ahead of Chainlink at $14.6 million and Aethir at $9.3 million, with NEAR, Bittensor, Virtuals and Internet Computer trailing and Akash at just $33,000. On that measure Render is not a minor player in AI crypto; it is the revenue leader.

The asterisk is large. About $21.4 million of that $23.3 million, roughly 92%, landed in September alone. A single strong month carrying a whole quarter is not obviously the same thing as durable demand, and TokenPost said as much, flagging the concentration as a reason to ask whether the figure reflects a trend or a spike.

A narrower lens tells a gloomier story. DefiLlama tracks on-chain render-job fees through the burn-and-mint contract, and by that measure fees fell by roughly half year over year through the third quarter. The two figures are not the same metric: one is a broad protocol-revenue tally, the other a narrow count of fees from rendering jobs settled on-chain, and the gap between them runs to roughly a hundredfold. Render has not publicly reconciled them. The honest reading is that the network is moving real volume, but how much of its revenue is recurring rendering demand, versus concentrated or one-off AI activity, is genuinely unsettled.

Against that backdrop, the token trades the way it trades.

RNP-024 and the Forth Gateway

Forth is Render’s name for the thing that has been missing. RNP-024, authored by Render Labs and created on October 16, describes it as a privacy-first AI gateway, routing and payment layer that gives users, developers, applications and autonomous agents a single API endpoint for reaching AI models and compute across many kinds of infrastructure. The proposal is explicit about what sits behind that endpoint: privately hosted and user-controlled infrastructure, anonymized third-party and frontier models, Render Network and other distributed GPU compute, and, in time, confidential-compute infrastructure, along with image, video, audio, inference, tool and agentic services.

Strip away the jargon and the strategic idea is simple. Render has spent nearly a decade building supply, the GPUs. Forth is an attempt to own the demand side by becoming the router that AI buyers pass through. Render Labs frames it as the front door through which AI demand enters the ecosystem. If buyers come in through Forth, and Forth meters and settles their spending in RENDER, then the token stops being something people trade in anticipation of demand and starts being the unit that demand is actually denominated in.

That is a meaningful step beyond the vertical-integration strategy Render has pursued with OTOY’s software stack. Owning the rendering tools and the marketplace gave Render a supply-side moat. Forth is the demand-side bookend: a single pane of glass over a messy field of model providers and compute vendors, with billing abstracted so a buyer never has to juggle separate credentials and invoices. Aggregators have been quietly powerful in other markets, from travel to ad buying; Forth is a bet that AI compute will reward one too.

The Incentive Program: Paying for Demand

Owning demand is not free, and the proposal is candid about the cost. RNP-024 authorizes up to 650,000 RENDER in total: 500,000 for a usage incentive program and 150,000 for marketing. Crucially, the money comes from the Foundation’s existing treasury, with, in the proposal’s words, no additional emissions beyond the existing BME emissions cap. That distinction matters, because the one thing a chronically inflationary token does not need is a new printing press. This is treasury spending, not fresh issuance.

The incentives roll out in three phases: a 100,000 RENDER pilot, then 150,000 if it works, then 250,000 after a sustained review, capped at twelve months or until the allocation runs out. Rewards are points-based, and the formula is engineered to push behavior in two directions. A payment multiplier pays 1.10 times for settling in RENDER versus 1.00 for USDC or fiat. A routing multiplier pays 1.25 times for sending the job to Render Network compute versus 1.00 for other infrastructure. A retention multiplier climbs toward 1.30 times for users who keep spending for twelve consecutive weeks. The multipliers stack to a cap of 1.50 times, only the first $1,000 of monthly spend per user is scored, and no user can earn back more than about half of their qualifying spend.

Read together, the levers say the quiet part out loud: Render will pay, within limits, for demand that is denominated in its token and routed to its own GPUs. That is a customer-acquisition subsidy, bounded and time-boxed on purpose. It also carries the risk every incentive program carries. Demand bought with rewards can leave when the rewards stop, the mercenary-liquidity problem that decentralized finance learned the hard way. The sunset clauses suggest Render knows this; whether Forth’s users stay once the multipliers fade is the test that matters.

Why Private Is the Whole Pitch

Why privacy-first, and why does it matter enough to put in the name? Because the buyers Render most wants are the ones a transparent public marketplace scares off. A studio rendering an unreleased film, an agency running a client’s brand assets through a model, a company doing inference on proprietary data: none of them want prompts, inputs or model weights sitting in the open on a public GPU bazaar. A gateway that can route to privately hosted infrastructure, anonymize access to third-party models and, eventually, lean on confidential compute is a pitch aimed squarely at that reluctance.

It also connects to something Render has talked about for years: provenance. The network’s Proof-of-Render mechanism began as a way to validate that a job’s output was genuinely produced by the node that claimed it, and OTOY has pushed to extend that idea into a provenance and intellectual-property layer for AI-generated media, tracking what model made what, from which inputs. Proving that an AI output came from a given model, run correctly, is its own fast-growing field; our look at opML and the push to make verifiable AI a plug-in covers the broader effort. For Render, a private gateway with built-in provenance is where its verifiability story meets an actual commercial need rather than a slide.

The ambition is not new. In 2024, announcing a provenance and IP-rights initiative with Stability AI and the entertainment group Endeavor, Urbach called it “a milestone that will shape the future of transparent, artist driven AI workflows and tools,” according to OTOY. Forth is the closest Render has come to turning that language into a product people pay for.

x402 and the Agent Economy

The payment plumbing is where Forth gets futuristic. It settles through x402, an HTTP-native payment protocol on Solana, in RENDER or USDC, with what the proposal calls accountless payment: a wallet signs and pays directly, no sign-up and no stored card, with conventional card payments available through MoonPay Commerce for everyone else. x402 revives an old corner of the web, the long-dormant HTTP 402 Payment Required status code, as a way for software to pay software per request.

That design only makes full sense if you believe the next big class of compute buyers is not people but programs. The bet behind wallet-native, pay-per-call rails is that autonomous AI agents will purchase inference and tools on their own, thousands of tiny transactions a day, without a human clicking checkout. Render explicitly lists autonomous agents among Forth’s intended users. If that economy arrives, a token an agent can spend without an account is useful infrastructure; if it does not, it is an elegant answer to a question nobody has asked yet.

The agent thesis cuts both ways. On-chain AI agents have been one of the most volatile corners of crypto, long on narrative and short on durable revenue, and the gap between agent hype and agent cash flow has burned plenty of tokens; the ai16z ghost-token saga is the cautionary extreme. There is also a human-factors question. Accountless, wallet-native payments are frictionless by design, and frictionless is not always safe; we have argued before that some friction in wallets is a feature, not a bug. Forth is making the opposite bet, that removing friction is what unlocks the agent economy. Both can be true at once.

Burn and Mint Equilibrium: the Engine and Its Ceiling

To see why demand matters so much to the price, you have to understand the engine. Render runs on a Burn and Mint Equilibrium, or BME. Jobs are priced in dollars; the buyer’s dollars are converted into RENDER at the market price at payment time; when the job completes, that RENDER is burned, minus a network fee of around 5%. Separately, node operators are paid in newly minted RENDER on a declining emissions schedule that Render’s documentation puts at roughly 9.1 million tokens in the first year and about 5.9 million in the second, which works out near 500,000 a month.

There is a subtlety here that trips up anyone reading burn charts as a straightforward bullish tell. Because jobs are priced in dollars, a higher token price burns fewer tokens for the same dollar of demand, and a lower token price burns more. A weak RENDER price mechanically inflates the token-denominated burn figure even if the real, dollar-measured demand is flat. Burn counts, in other words, flatter a falling price. It is the dollar value flowing through the network, not the token count going up in smoke, that tells you whether demand is actually growing.

And here is the ceiling. Render is still net inflationary. Burns have grown fast, up about 279% year over year by Crypto Briefing’s count, and on a good day the burn nearly cancels the mint: CoinMarketCap noted one recent day on which 15,693 RENDER, worth about $26,000, were destroyed across 164 transactions, offsetting roughly 98.85% of that day’s emissions. Cumulative burns have reached around 1.5 million RENDER. But monthly emissions near 500,000 tokens still outrun the burn, so the supply keeps growing. For BME to flip net-deflationary, dollar demand has to rise enough that burns consistently beat the mint. Closing that gap is precisely what Forth is for.

Whether a burn mechanism actually accrues value to a token, or just performs scarcity, is a live debate well beyond Render; our analysis of Hyperliquid and whether its burn is real yield or an expensive buyback frames the same question from the other end of the market.

How an RNP Becomes Real

Forth has to clear Render’s governance before it is anything more than a document. The process runs in four stages. A proposal is drafted as a pull request on GitHub and discussed on Discord, and the Foundation decides whether it is a full RNP or a smaller grant. It then faces an initial 72-hour community vote that needs only a simple majority and no quorum. If it passes, it goes to a binding final vote on Nation, Render’s Solana-native governance platform, which runs for six days and requires both a majority in favor and a quorum of 15% of the combined RNDR and RENDER supply. Only then do core contributors build it, with progress tracked back on GitHub.

Render migrated this machinery from Ethereum-based Snapshot voting to Nation, where holders vote with Phantom and other Solana wallets and voting power is snapshotted when a proposal is created. RNP-024 cleared its initial vote in late October and is now in the final stage. A pass is not a guarantee of success; it is permission to try. Render’s own history makes the point, with proposals that reached a vote and failed, such as the defeated RNP-020 on forward guidance, sitting in the record next to the ones that shipped.

Who Actually Renders on Render

For all the token drama, the network does real, visible work. The relightable Gaussian-splat visuals in A$AP Rocky’s Helicopter music video were rendered on Render through Octane. A full-CGI Santander and Formula 1 project by the studio Perhaps Creation ran on the network. White Rabbit, an animated short produced on Ethereum and rendered in an 18K version on Render, became the first Web3 project to win an Emmy in 2025. The immersive-art house ARTECHOUSE has used the network for its SUBMERGE shows, work that once took months to render and now takes weeks.

The people closest to the hardware describe the shift in blunt terms. NVIDIA chief executive Jensen Huang has said AI compute demand is increasing by an order of magnitude every single year, a curve no single data center can satisfy on its own. The digital artist David Ariew, describing a project rendered on the network, said work that “would have taken years” was “completed in days using Render.” Those are the twin claims the whole model rests on: demand growing faster than centralized supply can keep up, and a distributed grid that genuinely compresses the work.

Render has also built a dedicated AI-compute arm, Dispersed, launched at Solana’s Breakpoint conference in December 2025, where the large majority of payments are burned and early clients include the teams behind Jember, Scrypted and Intelligent Internet, per Messari. The calendar underneath all of this is busy: Render is showing at Motion Plus Design in Paris on November 14, running an immersive activation with Solana at Frameless in London during Breakpoint on November 16, and hosting a creators’ meetup at ARTECHOUSE in New York on November 20.

The DePIN Field and the Centralized Giant

Render is the largest token in its niche, but the niche is small and it does not trade alone. Its nearest decentralized peers are Akash, which is in the middle of migrating from its own Cosmos chain to Solana and trades near $0.61 for a market value around $182 million; io.net at about $0.125 and $53 million; and Nosana at roughly $0.40 and $40 million. Over the week into November 4 all of them were down together, Render about 11%, Akash about 10%, io.net 16% and Nosana 12%, which is the single clearest sign that this sector trades as a basket rather than on company-specific news.

Zoom out and the scale problem is stark. CoreWeave, the publicly traded AI cloud, reported $2.58 billion in revenue for the second quarter of 2026, up 112% year over year, on a backlog of about $104 billion, and guided to as much as $3.6 billion for the third quarter. One centralized provider books more revenue in a quarter than the entire decentralized GPU sector does in a year. The DePIN pitch was never to out-scale that. It is to be cheaper, more private and harder to switch off, running inference on idle consumer hardware that CoreWeave does not touch. Forth is a direct expression of that positioning: not a rival hyperscaler, but an aggregator that makes the long tail of private and distributed compute usable.

Beta or Fundamentals? Reading the Rally

So which is it, beta or fundamentals? The autumn evidence leans heavily toward beta. RENDER rallied hard in September, reclaiming a billion-dollar market value and testing $2 for the first time since a crash a year earlier, then gave much of it back in mid-October. The move looked less like the market repricing Render and more like money rotating through AI tokens as a group. A market-insights note from MEXC in early November described the latest bounce as a leverage-driven sector rotation rather than a confirmed breakout, pointing to thin liquidity, a long-to-short ratio skewed around 8.35, recent net outflows and a Fear and Greed reading near 31. That is the profile of positioning, not conviction.

The institutional bid is more durable but points the same way. RENDER is one of just four tokens in Grayscale’s accredited-only Decentralized AI Fund, sitting around 15% of the basket as of late October alongside NEAR, Bittensor and Filecoin. That is a real, standing source of demand, but it is also a reason RENDER moves as a sector proxy: a fund that rebalances a basket treats it as a slice of a theme, not as a single company with its own catalysts.

The deeper discomfort is the one running under all of AI crypto in 2026: token prices and token usage keep telling different stories. Render is a benign version of that gap, a network with genuine output whose token simply has not kept up. The same disconnect has an access angle too, as the broader altcoin-ETF wave gives funds new ways to buy themes wholesale; our one-year check-in on altcoin ETF approvals traces how that plumbing was built. Forth is Render’s wager that it can break out of the basket by giving RENDER a job that does not depend on sentiment: metering real demand.

Where RENDER Sits With the SEC

For a US holder, where does RENDER sit legally? Most straightforwardly, it reads as a utility and infrastructure token: it pays for a compute service and lets holders govern a network, which is not the profile of an obvious security. But the Securities and Exchange Commission applies a fact-specific framework for applying the Howey test to digital assets, and marketing a token around an appreciating, treasury-funded incentive program is exactly the kind of fact that can invite scrutiny.

The bigger uncertainty is structural. Legislation that would divide jurisdiction over spot crypto between the SEC and the Commodity Futures Trading Commission stalled in Congress in 2025, leaving the classification of tokens like RENDER unsettled, and RENDER was not among the assets named in the 2026 SEC and CFTC discussions of which digital commodities fall where, which centered on Bitcoin, Ethereum, Solana and a short list of large caps. For most US holders the practical exposure is not an enforcement action against the token; it is exchange and custody compliance, and tax. Paying for a compute service directly in RENDER, as Forth imagines, arguably bundles two taxable events, disposing of a token and buying a service, a wrinkle the rules have not cleanly addressed.

What to Watch Next

The near-term tells are specific. First, the vote: does RNP-024 clear the 15% quorum and the majority, and if it does, how fast does a pilot actually ship, given that governance approval and a working product are not the same thing. Then the substance: does Forth route real external demand, and does that demand pay in RENDER and land on Render’s own GPUs, which the incentive multipliers are built to encourage. Over a longer horizon the question is the one BME has posed all year: can dollar demand grow enough that burns finally outrun the roughly 500,000 RENDER minted each month and tip the token net-deflationary, and will September’s revenue concentration repeat or prove a one-off.

The honest summary is that Render has built an unusually complete stack: world-class rendering software, a global GPU grid, a dedicated AI-compute arm, and now a bid to own the demand funnel that feeds all three. What it has not yet shown is that any of this reliably pulls the token along with it. Forth is the most direct attempt so far to change that, to give RENDER a reason to rise that is not simply the AI sector having a good week. Whether a gateway can turn a busy network into a token that holds its value is the question Render carries into 2027.

Frequently Asked Questions

What is Render Network’s RNP-024 and Forth?

RNP-024 is a governance proposal authored by Render Labs to launch Forth, a privacy-first AI gateway, routing and payment layer. It gives users, developers and AI agents a single API endpoint to reach AI models and GPU compute across different kinds of infrastructure, with payments settled in RENDER or USDC. The proposal also authorizes up to 650,000 RENDER from the Foundation treasury to fund usage and marketing incentives.

Why is the RENDER token so far below its all-time high?

RENDER trades around $1.52, roughly 89% below its March 2024 peak of $13.53. The token set that high during the 2024 AI and crypto rally, and much of 2026’s price action has tracked the broader AI-token sector rather than Render-specific demand. The network is also still net inflationary, with monthly emissions near 500,000 RENDER exceeding the amount burned, so supply keeps growing.

How does Render’s Burn and Mint Equilibrium work?

Jobs on the network are priced in dollars and paid in RENDER, which is burned when the job completes, minus a small network fee. Node operators earn newly minted RENDER on a declining emissions schedule. Because jobs are dollar-denominated, a higher token price burns fewer tokens per dollar of demand and a lower price burns more, so token-denominated burn figures can look strong even when the dollar value of demand is flat.

Did Render really earn the most revenue of any AI crypto project in Q3 2026?

By one widely cited ranking from TokenPost, Render topped AI-crypto projects with about $23.3 million in third-quarter revenue, ahead of Chainlink and Aethir. About 92% of that came in September alone, which raises questions about how durable the demand is. A narrower measure of on-chain render-job fees from DefiLlama tells a weaker story, falling by roughly half year over year, and the two figures have not been publicly reconciled.

Is RENDER a security in the United States?

RENDER generally reads as a utility and DePIN token, used to pay for compute and to govern the network, rather than an obvious security, though the SEC applies a fact-specific Howey analysis to every token. US market-structure legislation that would clarify the split between the SEC and the CFTC stalled in 2025, and RENDER was not among the tokens named in 2026 digital-commodity discussions. For most US holders the practical issues are exchange compliance and tax treatment.

By Marcus Okafor, senior markets correspondent at HOGE Wire.