A live, no-KYC, crypto-payable GPU cloud and per-token LLM inference service (ASI:Cloud), but the CUDOS brand/token is historical after the ASI merger and GPU hourly pricing is login-gated.
by the bestcloudmining editorial desk · review version 2 · prose updated 2026-09-24 · facts as of 2026-09-24 · how we verify
Overview
CUDOS, now operating as ASI:Cloud under the ASI Alliance, is a GPU cloud and serverless LLM inference platform offering permissionless access via wallet login. It targets developers needing crypto-native, no-KYC compute for AI workloads. While live and technically functional, the CUDOS brand and token are defunct post-merger with FET, and GPU pricing is not publicly visible - requiring login to access.
What you can rent
You can rent GPU compute and serverless LLM inference through ASI:Cloud, a platform developed by the ASI Alliance (CUDOS x SingularityNET). Inference is priced per 1M tokens for models like Llama 3.3 70B, Qwen3-32B, and Mistral Nemo. Full GPU VMs are available but their hourly rates are not public. You get API/SSH access and pay in crypto, but you do not own any hardware or tokens - only usage rights on a centralized cloud backend.
Contract & terms
Contracts are flexible with no fixed term, typical of on-demand cloud services. There is no stated maintenance fee, but VMs are billed for the full first hour and continue to incur charges if stopped but not destroyed. Payouts aren’t applicable - this is a pay-to-use service. Refunds are not offered for crypto payments. Minimum top-up is typically $5, and full pricing access requires login.
Proof & verifiability
There is no documented proof-of-compute, hardware attestation, or TEE verification for ASI:Cloud’s GPU instances or inference API. The service operates as a conventional centralized cloud with crypto payment rails. While the inference endpoint is live (HTTP 401 confirmed), this does not verify underlying resource allocation - only that the API is active and access-controlled.
Reliability & history
CUDOS merged into the ASI Alliance in October 2024, with the blockchain and token fully retired. ASI:Cloud is a current product, backed by operator documentation and live endpoints. However, community engagement is minimal, with Reddit activity stalling in late 2025. The operator, Cudo Ventures, has a history in distributed compute, but the rebranding and merger mean legacy CUDOS is no longer active.
Pricing
Serverless inference pricing is public: for example, Qwen3-32B costs $0.17 per 1M tokens input and $0.43 output. The minimum top-up is $5. GPU VM hourly rates are not listed publicly and require login to view. Payments are accepted only in crypto (FET, USDC, USDT, ETH), with no fiat option. No KYC is required for cloud access, though a past early-access campaign reserved the right to enforce it.
Verdict
ASI:Cloud is a functional, no-KYC option for developers wanting crypto-paid GPU and LLM inference, but transparency is limited - especially on GPU pricing and infrastructure proof. Given the CUDOS brand’s retirement and opaque billing, test with small jobs first. Consider it only if wallet-based, non-custodial onboarding is a priority over verifiable resource allocation.
++ What works
Permissionless wallet login, explicitly no KYC, crypto-native
Public per-token inference price table, incl. a free asi1-mini model; $5 minimum top-up
Backed by CUDo Ventures' CUDO Compute GPU infrastructure plus SingularityNET inference
Cross-chain top-ups via Squid Router (70+ chains)
-- What to watch
No public GPU/VM hourly pricing (login-gated)
CUDOS token and chain are gone (merged into FET/ASI); only a legacy token still quotes
Docs were last updated September 2025; the main site returns HTTP 429 and the app 403 to automated access
Platform still labelled 'Beta'; no fiat and no refunds
Pricing reality check
Documented price
see provider site
Minimum
$5
Notes
Serverless inference (observed 2026-09-24) per 1M tokens in/out: asi1-mini free, Mistral Nemo $0.01/$0.04, gpt-oss-20b $0.03/$0.13, Gemma-3-27B $0.09/$0.20, Qwen3-32B $0.17/$0.43, Llama 3.3 70B $0.30/$0.43, GLM 4.5 Air $0.17/$0.93. GPU/VM hourly pricing is NOT public (requires login; docs point to CUDO Compute on-demand, whose pricing page returned HTTP 403 to this server).
Based on our checks as of 2026-09-24: CUDOS scores 6.1/10 in our weighted review. We logged no red flags. User reports and operator background are documented with dated sources in the meta-review below.
What do I actually get when I pay?
You get access to GPU VMs or serverless LLM inference via API, but no ownership of hardware or tokens. Usage is billed hourly (VMs) or per token (inference), with no proof of underlying resource allocation.
What are the total costs and fees?
Serverless inference is priced per 1M tokens (e.g., $0.17-$0.43 for Qwen3-32B). VMs are hourly but pricing is not public. No maintenance fee is listed, but stopped VMs still incur charges.
Is there a minimum spend or trial available?
Yes, the minimum top-up is typically $5. There’s no free trial, but the low entry cost allows small-scale testing. Serverless inference has no minimum beyond per-request cost.
How do payouts work?
Payouts don’t apply - you are paying for compute, not earning from it. This is a cloud rental service, not a mining or staking platform.
What happens if the service shuts down?
If ASI:Cloud shuts down, your access ends immediately and crypto payments are non-refundable. You have no recourse or asset recovery, as you own no infrastructure or tokens.
How does this compare to buying hardware or coins directly?
Renting avoids upfront hardware costs and maintenance. But unlike owning hardware or coins, you gain no long-term asset - only temporary compute. Price efficiency depends on workload and duration.
// The meta-review: what the internet says about CUDOS
We do not expect you to trust us. This review is a meta-review: below you see what other review sites claim, what users report on Reddit, forums and complaint boards, and who actually operates CUDOS. Every claim links to its source so you can verify it yourself. Our verdict forms from these sources plus our own fact checks - not the other way around.
// What other reviewers say (0)
Heads-up: most cloud mining "reviews" are affiliate content. We mark our reliability judgement on every source and include critical voices deliberately.
No third-party reviews with usable evidence found yet. In this niche that is itself a signal - most "review" sites are paid placements.
// What users say (1)
Reddit, forums, complaint boards. Anecdotes, not proof - but payout complaints and host-reliability patterns are the best early-warning system this market has.
Launch post describing ASI:Cloud as 'a permissionless GPU Cloud and AI inference platform developed by CUDOS and SingularityNET', offering open wallet-based access to GPU infrastructure, transparent per-token pricing, an open-source LLM catalogue via API, and ASI wallet/staking integration. Retrieved via pullpush.io.
// Who is behind CUDOS?
Cloud mining ownership is often deliberately opaque. This is what we could verify - with sources. Confidence: medium (checked 2026-09-24).
CUDOS merged into the ASI Alliance in Oct 2024: the vote closed 2024-09-24 with >96% approval, swap 118.344 CUDOS:1 FET, blockchain merger completed 2024-10-28, after which CUDOS had no separate chain or token. The surviving product is ASI:Cloud, 'developed by the ASI Alliance (CUDOS x SingularityNET)'; inference is provided by SingularityNET and GPU infrastructure is operated by CUDOS, the web3 arm of CUDO, which runs the CUDO Compute GPU cloud. Historical token terms name Cudos Limited, a Gibraltar private company (no. 119334, incorporated 2019-12-05); the site footer reads 'Cudos Ltd. 2019-2025'. ASI:Cloud docs were last updated September 2025.
Every external source used on this page, with the date we fetched it. Methodology: how we verify. Sources disappear or change - if a link is dead, tell us via contact.
No comments yet - be the first. Posting runs a short proof-of-work in your browser (anti-spam), no account needed.