> ## Content Index
> Fetch the complete content index at: https://www.sotkn.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Neocloud vs Hyperscaler
- URL: https://www.sotkn.com/explained/neocloud-vs-hyperscaler/
- Published: 2026-09-19T17:18:36.000Z
- Updated: 2026-09-23T08:41:46.000Z
- Description: A neocloud bills accelerator time; a hyperscaler sells a broad cloud with GPUs as one line. Compare funding, pricing, lock-in, and what a demand dip hits.
- Author: Sean
- Tags: Versus, #desk-explained

EXPLAINEDLast reviewed: 20 September 2026Next review: after next CoreWeave or hyperscaler AI-capex update

Neocloud vs hyperscaler is the contrast between a specialist cloud that bills GPU accelerator time and a broad cloud that sells many services with GPUs as one line.

**In short:** A neocloud bills GPU hours and lives on utilisation and realised price. A hyperscaler funds AI from a broad book and sets CapEx demand for the stack with lag. Same accelerators can appear in both stories. Who gets paid, where money sticks, and what a demand dip hits are different.

For investors the money link is simple: who gets paid when GPU hours bind on a specialist fleet, versus who gets paid when CapEx orders hit scarce finish lines with lag. This page is the contrast — not a scorecard of names.

## What is the difference between a neocloud and a hyperscaler?

A [neocloud](https://www.sotkn.com/explained/neocloud-explained/) concentrates on accelerators and bills time by the hour or under reserved contracts. A [hyperscaler](https://www.sotkn.com/explained/hyperscaler-explained/) sells storage, databases, general compute, and AI as one menu. That difference drives revenue concentration, balance-sheet shape, and dip sensitivity.

![Neocloud vs hyperscaler — GPU revenue share and concentration](https://storage.ghost.io/c/cd/10/cd10e222-131f-4ba4-b4a9-ca4214e8c12d/content/images/2026/09/neocloud-vs-hyperscaler-concentration.png)

GPU revenue share: on a neocloud accelerators are effectively all of it; on a hyperscaler they are a slice.

| Axis                   | Hyperscaler                                    | Neocloud                                                |
| ---------------------- | ---------------------------------------------- | ------------------------------------------------------- |
| Service range          | Storage, databases, general compute, GPUs      | Accelerators, plus the minimum around them              |
| Revenue concentration  | GPUs are a slice                               | GPUs are effectively all of it                          |
| What funds expansion   | Cash from unrelated or broader lines           | Debt, leases, equity, customer prepayments              |
| Pricing behaviour      | Anchors the market; discounts strategically    | Prices against the anchor; competes on access and speed |
| Customer lock-in       | Data gravity and service breadth               | Supply certainty and price                              |
| Demand-dip sensitivity | Absorbed by other lines                        | Hits the whole business                                 |
| Deploy speed           | Slower, process-bound                          | Faster when GPUs are scarce                             |
| Customer type          | Enterprises and platforms needing a full stack | AI teams that need capacity now                         |

![Neocloud vs hyperscaler — demand-dip hit on the whole business vs absorbed by other lines](https://storage.ghost.io/c/cd/10/cd10e222-131f-4ba4-b4a9-ca4214e8c12d/content/images/2026/09/neocloud-vs-hyperscaler-dip-risk.png)

Demand-dip risk: a soft AI hour hits the whole neocloud book; a hyperscaler absorbs it inside other lines.

See the [neoclouds stack](https://www.sotkn.com/stack/neoclouds/) and the [hyperscalers stack](https://www.sotkn.com/stack/hyperscalers/) for how each layer sits next to chips, power, and halls.

## When does each win?

A neocloud often wins when scarce GPUs must land fast and the buyer needs committed blocks without a full platform move.

A hyperscaler often wins when data gravity and the multi-service stack outweigh raw GPU access speed. CapEx from Amazon, Microsoft, Alphabet, and Meta also sets supplier calendars. Amazon guided about US$200 billion of company-wide CapEx for 2026 (Amazon IR, 5 February 2026). Alphabet raised its 2026 CapEx guide to US$195–205 billion on the Q2 2026 call (CNBC, 22 July 2026). Those are guided envelopes, not supplier profit forecasts.

Customers often use both. Overflow hits neoclouds when hyperscaler queues are long; pricing pressure hits when hyperscalers expand AI instances.

## Who gets paid — and where money sticks?

On the neocloud path, cash sticks when billed hours cover hardware carry, power, cooling, and financing. CoreWeave reported Q2 2026 revenue of US$2.575 billion and company-defined backlog of about US$104 billion as of 30 June 2026 (CoreWeave earnings, 11 August 2026) — subject to delivery, not guaranteed profit. Commitments were 98% of revenue; Customer A was 36%, B 26%, C 10% of the quarter (Form 10-Q). Concentration cuts both ways: visibility, and first hit if a buyer pauses.

On the hyperscaler path, CapEx orders pay accelerators, packaging, memory, optics, and halls with lag. AWS net sales rose 37% year over year in Q2 2026 (Amazon IR, 30 July 2026) — reported growth inside a broad book. Meta narrowed FY2026 CapEx outlook to US$130–145 billion (Meta IR, 29 July 2026) — hyperscaler-scale demand even when ads, not cloud, are the main earn.

![Who gets paid — utilisation stuck step on neocloud vs CapEx lag on hyperscaler](https://storage.ghost.io/c/cd/10/cd10e222-131f-4ba4-b4a9-ca4214e8c12d/content/images/2026/09/neocloud-vs-hyperscaler-infographic-markets.png)

Stuck step: idle GPUs still carry capital and power on a neocloud; hyperscaler CapEx sticks at scarce finish lines with lag.

| Step                | Hyperscaler path                                                  | Neocloud path                                             |
| ------------------- | ----------------------------------------------------------------- | --------------------------------------------------------- |
| Who gets paid first | Chip, packaging, power, and hall suppliers when CapEx orders bind | Owners of live GPUs when reserved or on-demand hours bill |
| Where money sticks  | Scarce finish lines (packaging, HBM, watts) with lag              | Utilisation × realised price after discounts              |
| Stuck step          | Orders vs installs — guidance can rise while halls wait on power  | Idle cards still carry capital and power cost             |

Markets door for contracted neocloud demand: [CoreWeave neocloud contracts](https://www.sotkn.com/markets/us/coreweave-neocloud-contracts/). Power clock when halls wait on watts: [the AI trade’s 2028 grid-queue bottleneck](https://www.sotkn.com/analysis/speed-to-power-the-ai-trade-s-real-bottleneck-is-a-2028-grid-queue/).

## How does pricing compare?

Hyperscalers often anchor list rates and discount inside large deals. Neoclouds price against that anchor and compete on access speed and committed blocks. As at 20 September 2026, Lambda’s public on-demand list showed H100 PCIe at US$3.29 per GPU-hour and H100 SXM at US$3.99 on larger nodes (Lambda pricing page) — list price, not realised take-or-pay rates.

## Common misconceptions

- **People assume the same GPU means the same business.** Actually revenue concentration decides risk. A GPU SKU inside a hyperscaler is a slice; on a neocloud it is the earn.
- **People assume list price is what either side earns.** Actually large buyers negotiate off the page. Realised price after discounts is the number that matters.
- **People assume CapEx up means neoclouds lose immediately.** Actually orders, installs, and billed hours run on different clocks. Overflow can still fill specialist fleets while guidance rises.
- **People assume a big backlog equals booked profit.** Actually company-defined backlog and remaining performance obligations are subject to delivery and availability.

## FAQ

**What is the difference between a neocloud and a hyperscaler?**  
A neocloud concentrates revenue on accelerator time. A hyperscaler earns across many services and treats GPUs as one line. That changes funding, pricing power, and demand-dip risk.

**When does a neocloud win versus a hyperscaler?**  
A neocloud often wins when scarce GPUs must land fast and the buyer needs committed blocks without a full platform migration. A hyperscaler wins when data gravity, multi-service stack, and long-term platform lock-in matter more than raw GPU access speed.

**Who gets paid in each model?**  
On a neocloud, cash sticks when billed GPU hours cover hardware carry, power, cooling, and debt or lease cost. On a hyperscaler, CapEx orders pay chip, packaging, power, and hall suppliers with lag, while cloud fees land across many product lines.

**Is Oracle a neocloud or a hyperscaler?**  
Structurally a hyperscaler, because it sells a full cloud platform. Its AI infrastructure line can still behave like a neocloud when growth concentrates on committed GPU contracts.

**Why does customer concentration matter more for neoclouds?**  
Scaled neoclouds often disclose large share from a few customers. A hyperscaler can absorb a lost GPU deal inside a broad book. A neocloud cannot.

**Do list GPU prices tell you who earns more?**  
No. List price is a starting point. Large buyers negotiate off the page. Realised price after discounts, plus utilisation, decides the earn.

## Related terms

- [Neocloud](https://www.sotkn.com/explained/neocloud-explained/) — specialist GPU cloud that bills accelerator time.
- [Hyperscaler](https://www.sotkn.com/explained/hyperscaler-explained/) — broad cloud whose CapEx sets stack demand with lag.

## Markets and stack doors

Layers: [neoclouds](https://www.sotkn.com/stack/neoclouds/) · [hyperscalers](https://www.sotkn.com/stack/hyperscalers/). Markets: [CoreWeave neocloud contracts](https://www.sotkn.com/markets/us/coreweave-neocloud-contracts/).

## Further reading

Analysis teaser: [speed-to-power and the 2028 grid queue](https://www.sotkn.com/analysis/speed-to-power-the-ai-trade-s-real-bottleneck-is-a-2028-grid-queue/).

[Get the Wire](https://www.sotkn.com/#/portal/signup) — short notes when CapEx or neocloud clocks move.

**How we check this**

Figures come from company filings and government releases, each dated in the list below. Capacity and timing from press reports are labelled as reported, not guided. Where we work something out ourselves, the arithmetic is shown in full.

**Last reviewed:** 20 Sep 2026 · **Next review:** after next CoreWeave or hyperscaler AI-capex update

Spot an error? [Tell us](mailto:sean@thekopinotes.com) and we will correct it and note the change here.

[Our editorial standards →](https://www.sotkn.com/editorial-standards/)

## Sources

- [CoreWeave Q2 2026 results (11 Aug 2026) — revenue US$2.575B; backlog \~US$104B](https://investors.coreweave.com/news/news-details/2026/CoreWeave-Reports-Strong-Second-Quarter-2026-Results/default.aspx?ref=sotkn.com)
- [CoreWeave Form 10-Q, period ended 30 Jun 2026 — 98% committed; customer concentration](https://www.sec.gov/Archives/edgar/data/1769628/000176962826000366/crwv-20260630.htm?ref=sotkn.com)
- [Amazon Q4 2025 results (5 Feb 2026) — \~US$200B company-wide CapEx guided for 2026](https://ir.aboutamazon.com/news-release/news-release-details/2026/Amazon-com-Announces-Fourth-Quarter-Results/?ref=sotkn.com)
- [Amazon Q2 2026 results (30 Jul 2026) — AWS net sales +37% YoY](https://ir.aboutamazon.com/news-release/news-release-details/2026/Amazon-com-Announces-Second-Quarter-Results/default.aspx?ref=sotkn.com)
- [CNBC — Alphabet raises 2026 CapEx guide to US$195–205B (22 Jul 2026)](https://www.cnbc.com/2026/07/22/google-earnings-q2-goog-live-updates.html?ref=sotkn.com)
- [Meta Q2 2026 results (29 Jul 2026) — CapEx US$31.08B; 2026 outlook US$130–145B](https://investor.atmeta.com/investor-news/press-release-details/2026/Meta-Reports-Second-Quarter-2026-Results/?ref=sotkn.com)
- [Data Center Dynamics — Microsoft FY2026 CapEx \~US$145.3B; FY2027 \~US$175B (30 Jul 2026)](https://www.datacenterdynamics.com/en/news/microsoft-brought-88-data-centers-online-in-fy2026/?ref=sotkn.com)
- [Lambda GPU cloud pricing page — H100 list rates accessed 20 Sep 2026](https://lambda.ai/pricing?ref=sotkn.com)

S

**[Sean](https://www.sotkn.com/author/sean/)**

Writes Second Order by TKN — a plain-English map of AI infrastructure and semiconductors for non-specialist investors. Focus: who gets paid, where money sticks, and which step is stuck.

Covers AI infrastructure, semiconductors, and second-order stack reads.

[Author page →](https://www.sotkn.com/author/sean/) · [Editorial standards →](https://www.sotkn.com/editorial-standards/)