{"id":"2869e3f5-335b-46c5-be47-a328effaa639","name":"Parasail","slug":"parasail","description":"Parasail is a US-based AI inference cloud that sells access to large language and open-weight models rather than raw GPU-hours. Instead of owning campuses, it aggregates on-demand capacity from dozens of neocloud and colocation providers and routes workloads across that fleet through a single OpenAI-compatible API. Co-founder and CEO Mike Henry (previously Chief Product Officer at Groq and founder-CEO of Mythic) and co-founder Tim Harris (founder-CEO of Swift Navigation) launched the company in 2023, began engineering in early 2024, and came out of stealth in early 2025.\n\nWhat differentiates Parasail from CoreWeave, Nebius, Crusoe and other capex-heavy neoclouds is that it never took the hyperscale infrastructure path. The company positions itself around what it calls the \"token economy\": customers commit to output (tokens served) on a fluid pool of Nvidia H100, H200, A100 and consumer 4090 GPUs sourced from partners, with dedicated deployments also available on newer parts including B300. That aggregator model let Parasail claim in April 2025 that its bookable on-demand fleet was larger than Oracle Cloud's, without Parasail itself owning any of it.\n\nParasail sits firmly on the inference side of the neocloud pack, aimed at AI-native startups rather than frontier training labs or Fortune 500 enterprise IT. Disclosed customers include Elicit, Weights and Biases, Rasa, MiniMax, Venice, mem0 and Skyfall AI. The company raised a $10 million seed in 2024 (Basis Set Ventures, Threshold Ventures, Buckley Ventures, Black Opal Ventures) and a reported $32 million follow-on round in April 2026, with Kindred Ventures, Touring Capital, Samsung Next and Banyan Ventures listed among backers.","status":"active","founded":2023,"hq":"United States (HQ city not disclosed publicly)","hqLat":null,"hqLng":null,"hqGeocodedAt":null,"hqAddress":null,"hqPrecision":null,"website":"https://parasail.io","fundingTotal":"42000000","type":null,"roles":[],"reviewStatus":"reviewed","lifecycleState":"active","supersededByCompanyId":null,"sources":[{"url":"https://parasail.io","title":"Parasail — The Inference Cloud for AI-native startups","publisher":"Parasail"},{"url":"https://parasail.io/about","title":"About Parasail","publisher":"Parasail"},{"url":"https://parasail.io/pricing","title":"Parasail pricing","publisher":"Parasail"},{"url":"https://techcrunch.com/2025/04/02/parasail-says-its-fleet-of-on-demand-gpus-is-larger-than-oracles-entire-cloud/","date":"2025-04-02","title":"Parasail says its fleet of on-demand GPUs is larger than Oracle's entire cloud","publisher":"TechCrunch"},{"url":"https://techcrunch.com/tag/parasail/","date":"2026-04-15","title":"Parasail coverage tag page (includes April 2026 $32M round article)","publisher":"TechCrunch"}],"keyFacts":[{"label":"Positioning","value":"Inference-focused aggregator: sells token-priced access to open + frontier LLMs, not raw GPU capacity it owns","sourceUrl":"https://parasail.io"},{"label":"GPU inventory (basis)","value":"Aggregates on-demand Nvidia H100, H200, A100 and 4090 capacity from dozens of third-party providers; Parasail does not disclose an absolute chip count and does not own the fleet","sourceUrl":"https://techcrunch.com/2025/04/02/parasail-says-its-fleet-of-on-demand-gpus-is-larger-than-oracles-entire-cloud/"},{"label":"Fleet-size claim","value":"Parasail claimed in April 2025 that its bookable on-demand GPU fleet was larger than Oracle Cloud's entire GPU footprint (company claim, not independently audited)","sourceUrl":"https://techcrunch.com/2025/04/02/parasail-says-its-fleet-of-on-demand-gpus-is-larger-than-oracles-entire-cloud/"},{"label":"Dedicated hardware offered","value":"Reserved deployments listed for Nvidia B300 (288 GiB) and H100 SXM (80 GiB), billed per minute with volume discounts","sourceUrl":"https://parasail.io/pricing"},{"label":"Serverless pricing examples","value":"Mistral Small at $0.09 input / $0.30 output per million tokens; Llama 3.3 70B at $0.22 / $0.50 per million tokens (per Parasail pricing page)","sourceUrl":"https://parasail.io/pricing"},{"label":"Throughput claim","value":"~750 billion tokens served per day across the platform (company figure)","sourceUrl":"https://parasail.io"},{"label":"Silicon partners","value":"Primarily Nvidia GPUs via third-party providers; disclosed technology partnerships with d-Matrix, Neuralwatt and Wafer AI for inference-efficiency work","sourceUrl":"https://parasail.io/blog"},{"label":"Named customers","value":"Elicit, Weights & Biases, Rasa, MiniMax, mem0, Venice, Reflow, Z, Skyfall AI","sourceUrl":"https://techcrunch.com/2025/04/02/parasail-says-its-fleet-of-on-demand-gpus-is-larger-than-oracles-entire-cloud/"},{"label":"Seed round","value":"$10M seed closed in 2024, led by Basis Set Ventures with Threshold Ventures, Buckley Ventures and Black Opal Ventures participating","sourceUrl":"https://techcrunch.com/2025/04/02/parasail-says-its-fleet-of-on-demand-gpus-is-larger-than-oracles-entire-cloud/"},{"label":"Follow-on round","value":"Reported $32M round in April 2026 per TechCrunch; Kindred Ventures, Touring Capital, Samsung Next and Banyan Ventures listed as backers on Parasail's site","sourceUrl":"https://techcrunch.com/tag/parasail/"},{"label":"Public/private","value":"Private company; no public ticker","sourceUrl":"https://parasail.io"},{"label":"CEO","value":"Mike Henry, co-founder and CEO; previously Chief Product Officer at Groq and founder-CEO of Mythic","sourceUrl":"https://parasail.io/about"},{"label":"Co-founder","value":"Tim Harris, co-founder; previously founder and CEO of Swift Navigation","sourceUrl":"https://parasail.io/about"},{"label":"HQ","value":"United States; Parasail has not disclosed a specific HQ city on its site or in coverage reviewed","sourceUrl":"https://parasail.io/about"},{"label":"Founding","value":"Company launched in 2023; engineering work began in early 2024; emerged from stealth in early 2025","sourceUrl":"https://techcrunch.com/2025/04/02/parasail-says-its-fleet-of-on-demand-gpus-is-larger-than-oracles-entire-cloud/"}],"aliases":[],"collisionRisk":"low","reviewNote":null,"atsProvider":null,"atsSlug":null,"factoryLocations":null,"manufacturingCapacity":null,"orgType":null,"burnSignal":null,"litigationEvents":null,"insurancePostureDisclosed":null,"unitEconomicsDisclosed":null,"tickerSymbol":null,"stockExchange":null,"secCik":null,"revenueUsd":null,"revenueBasis":null,"revenueAsOf":null,"employeeCount":null,"employeeCountBasis":null,"employeeCountAsOf":null,"marketCapUsd":null,"marketCapAsOf":null,"marketCapSource":null,"createdAt":"2026-08-28T01:48:32.832Z","updatedAt":"2026-08-28T03:45:17.421Z","jsonLd":{"@context":"https://schema.org","@type":"Organization","@id":"https://registry.deploy.report/companies/parasail","url":"https://registry.deploy.report/companies/parasail","name":"Parasail","description":"Parasail is a US-based AI inference cloud that sells access to large language and open-weight models rather than raw GPU-hours. Instead of owning campuses, it aggregates on-demand capacity from dozens of neocloud and colocation providers and routes workloads across that fleet through a single OpenAI-compatible API. Co-founder and CEO Mike Henry (previously Chief Product Officer at Groq and founder-CEO of Mythic) and co-founder Tim Harris (founder-CEO of Swift Navigation) launched the company in 2023, began engineering in early 2024, and came out of stealth in early 2025.\n\nWhat differentiates Parasail from CoreWeave, Nebius, Crusoe and other capex-heavy neoclouds is that it never took the hyperscale infrastructure path. The company positions itself around what it calls the \"token economy\": customers commit to output (tokens served) on a fluid pool of Nvidia H100, H200, A100 and consumer 4090 GPUs sourced from partners, with dedicated deployments also available on newer parts including B300. That aggregator model let Parasail claim in April 2025 that its bookable on-demand fleet was larger than Oracle Cloud's, without Parasail itself owning any of it.\n\nParasail sits firmly on the inference side of the neocloud pack, aimed at AI-native startups rather than frontier training labs or Fortune 500 enterprise IT. Disclosed customers include Elicit, Weights and Biases, Rasa, MiniMax, Venice, mem0 and Skyfall AI. The company raised a $10 million seed in 2024 (Basis Set Ventures, Threshold Ventures, Buckley Ventures, Black Opal Ventures) and a reported $32 million follow-on round in April 2026, with Kindred Ventures, Touring Capital, Samsung Next and Banyan Ventures listed among backers.","identifier":"2869e3f5-335b-46c5-be47-a328effaa639","foundingDate":"2023","address":"United States (HQ city not disclosed publicly)","publisher":{"@id":"https://deploy.report/#organization"}},"framework_metadata":{"framework_schema_version":"0.1.0","verification_status":"verified","maturity_stage":null,"lifecycle_state":null,"architectural_position":{"cohort":null,"sub_cohorts":[]},"within_cohort_verified_vs_claimed_pair":null,"cap_flags":[],"verification_depth":{"sources_count":5,"primary_source_types":["secondary-industry-publication"]}}}