Connect your agent to akta.pro and build company lists like this across sectors and geos.
Private Neocloud and GPU-as-a-Service companies, Series A or later, ranked by Q3 2026 primary-subject news volume — funding, business model, revenue model, news signals and headcount growth for each.
| Company | Funding | Business Model | Revenue Model | Signals (Q3 2026) | Headcount growth (Q3) |
|---|---|---|---|---|---|
| Nscale ↗ 🇬🇧 London, GBR Series C | Funding $9.81B | Product Led Growth Enterprise Field Sales Channel Partners | Usage Based Subscription Recurring Managed Services | 412 in Q3 View signals (94) → | +19% in Q3 884 → 1,052 employees |
| Together AI ↗ 🇺🇸 San Francisco, CA Series C | Funding $1.33B | Api First Product Led Growth | Usage Based Subscription Recurring Managed Services | 173 in Q3 View signals (50) → | +6% in Q3 421 → 445 employees |
| Crusoe ↗ 🇺🇸 Denver, CO Series F | Funding $7.20B | Enterprise Field Sales Sales Led | Subscription Recurring Managed Services | 161 in Q3 View signals (61) → | +8% in Q3 1,693 → 1,820 employees |
| Bittensor ↗ 🇨🇦 Toronto, CAN Series C | Funding $8.5M | Community Led Api First Event Driven | Transaction Fee Usage Based Licensing Royalties Marketplace Commission | 151 in Q3 View signals (120) → | -3% in Q3 35 → 34 employees |
| Lambda ↗ 🇺🇸 San Jose, CA Series E | Funding $5.11B | Product Led Growth Enterprise Field Sales Sales Led Channel Partners Community Led Api First | Usage Based Subscription Recurring Hardware Sales | 108 in Q3 View signals (21) → | +6% in Q3 787 → 831 employees |
| Fluidstack ↗ 🇺🇸 New York, NY Private Equity | Funding $3.78B | Enterprise Field Sales Api First Channel Partners Direct To Consumer | Subscription Recurring Usage Based Managed Services | 72 in Q3 View signals (33) → | +70% in Q3 358 → 608 employees |
| Fireworks AI ↗ 🇺🇸 Redwood City, CA Series D | Funding $1.83B | Api First Enterprise Field Sales Community Led | Usage Based Subscription Recurring | 59 in Q3 View signals (25) → | +44% in Q3 311 → 448 employees |
| Modal Labs ↗ 🇺🇸 New York, NY Series C | Funding $465.5M | Product Led Growth | Usage Based Freemium | 58 in Q3 View signals (23) → | +19% in Q3 216 → 256 employees |
| Prime Intellect ↗ 🇺🇸 San Francisco, CA Series A | Funding $200.4M | Product Led Growth Sales Led Community Led Channel Partners | Usage Based Subscription Recurring Marketplace Commission | 38 in Q3 View signals (16) → | +16% in Q3 70 → 81 employees |
| Aolani Cloud ↗ 🇸🇬 Singapore, SGP Series B | Funding n/d | Enterprise Field Sales Inside Sales | Usage Based Subscription Recurring | 18 in Q3 View signals (7) → | +6% in Q3 18 → 19 employees |
| PaleBlueDot AI ↗ 🇺🇸 Palo Alto, CA Series B | Funding $415.0M | Sales Led | Usage Based Subscription Recurring | 14 in Q3 View signals (5) → | +10% in Q3 30 → 33 employees |
| Thunder Compute ↗ 🇺🇸 San Francisco, CA Series A | Funding $37.9M | Product Led Growth | Usage Based | 13 in Q3 View signals (4) → | — |
| Deep Infra ↗ 🇺🇸 Palo Alto, CA Series B | Funding $153.6M | Api First Product Led Growth Enterprise Field Sales | Usage Based | 6 in Q3 View signals (5) → | +2% in Q3 42 → 43 employees |
| Runpod ↗ 🇺🇸 Mount Laurel, NJ Series A | Funding $122.0M | Product Led Growth Community Led Enterprise Field Sales Channel Partners Api First | Usage Based Subscription Recurring Transaction Fee | 5 in Q3 View signals (5) → | +21% in Q3 118 → 143 employees |
| io.net ↗ 🇺🇸 New York, NY Series A | Funding $32.5M | Marketplace Listing Api First | Usage Based | 5 in Q3 View signals (4) → | -1% in Q3 80 → 79 employees |
A neocloud is an infrastructure company built to sell GPU-dense compute directly to AI teams, financed and provisioned specifically for training and inference workloads rather than bundled into a general-purpose cloud platform. It is also written as GPU cloud, AI compute cloud, or compute-as-a-service; all describe the same shift away from the hyperscale public cloud model and toward capacity that is purpose-built, and often purpose-financed, for one workload: running large models.
The 15 companies here fall into four segments. Capacity builders such as Nscale, Crusoe, Lambda and Fluidstack acquire and operate GPU data centers directly, frequently financing the hardware itself, and sell that capacity as the core product. Inference and serving platforms such as Together AI, Fireworks AI, Deep Infra and Modal Labs sit a layer above raw capacity, offering API-first model hosting and serving on top of GPUs they lease or own. Decentralized compute networks such as Bittensor, io.net and Prime Intellect match idle or distributed GPU supply to demand through a marketplace rather than owning a fleet outright. Developer-first clouds such as Thunder Compute, Runpod, PaleBlueDot AI and Aolani Cloud serve smaller teams and specific workloads with simpler self-serve access.
The distinction that matters most is against the hyperscale public clouds — AWS, Azure and Google Cloud — which this list excludes entirely. A hyperscaler bundles GPU instances inside a much broader platform of storage, databases, networking and managed software, and a customer typically arrives already committed to that ecosystem. A neocloud sells nothing but the GPU-dense capacity itself, financed to track GPU refresh cycles, with the customer bringing its own training or inference stack. That changes the balance sheet: capacity builders carry debt-backed hardware as their core asset, while marketplace and serving-layer companies carry almost none.
The table substantiates the pricing consequence. 14 of 15 companies price on usage-based or metered terms, and 10 of those also layer subscription-recurring or managed-services revenue on top — evidence that even “raw compute” vendors are moving toward committed-capacity contracts to lock in enterprise demand. Only three companies — Bittensor, Runpod and Prime Intellect — run outright marketplace-commission or transaction-fee revenue, the signature of the decentralized-compute segment, where the company never takes the underlying hardware onto its own balance sheet.
Companies are tagged against standard industry codes and against akta.pro's internal taxonomy, so the same set can be rebuilt by code, by segment, or by a combination of both.
NAICSIndustry codes are provided for cross-referencing against public filings and procurement records. akta.pro classification is independent of them, so a company may carry an internal segment without a matching NAICS or SIC registration.
Connect your agent to akta.pro and build company lists like this across sectors and geos.
Help me integrate (15 mins, free) ↗