Are Neoclouds a Lasting Part of the AI Compute Value Chain?

One logical question to be asked about the neocloud segment of the artificial intelligence compute value chain is how sustainable the role might become, over time, as some amount of consolidation occurs. 

History suggests a new and sustainable role in the AI computing value chain could emerge. 

Hardware and platform layers tend to consolidate first, which might suggest to some that neocloud service providers (as infrastructure) could consolidate and eventually be absorbed by hyperscalers.

Middleware and applications repeatedly re-fragment, on the other hand (databases, runtimes, ML frameworks). .

But there also is an argument to be made that intermediation layers survive. As time-sharing bureaus to value-added resellers and managed service providers have emerged as sustainable niches, neoclouds could emerge as a permanent part of the value chain, providing customers (hyperscalers, for example):

• Price discovery

• Flexibility (financial and operational)

• Vendor neutrality.

Hyperscalers dominate integrated platforms, but merchant compute and specialized capacity might be sustainable positions in the value chain. 

In every computing era, the dominant platform provider tries to absorb adjacent layers. But a neutral or merchant layer re-emerges when:

• Utilization is volatile

• Customers resist lock-in

• Economics differ by workload

That pattern strongly suggests neocloud is not an anomaly, even if there are business reasons the hyperscalers providing “AI compute as a service” might prefer the role be limited.

For starters, to the extent there are supply constraints for graphics processing units, neoclouds compete for that supply, and reduce hyperscaler leverage over chip vendors. 

Neoclouds also can expose:

• High gross margins on certain workloads

• Cross-subsidies inside hyperscaler pricing

• Arbitrage opportunities hyperscalers don’t want visible

For the hyperscalers, non-existence of neoclouds strengthens the “buy from us; there is no alternative” positioning. Without neocloud alternatives, there are fewer opportunities for customers to ask “why is this cheaper elsewhere?”

So there will be some logic for hyperscalers to absorb, starve, or outflank neoclouds. 

On the other hand, there are structural reasons an independent neocloud role persists. Hyperscalers are bad at merchant compute, one might argue. 

Hyperscalers prefer:

• Platform lock-in

• Long-lived customer relationships

• Bundled services

• Predictable utilization. 

They are not optimized for, or do not prefer:

• Bursty, price-sensitive workloads

• Short-term GPU leasing

• Single-workload economics

• Pricing experimentation. 

Even if hyperscalers can do neocloud-style offerings, they often won’t, because doing so:

• cannibalizes higher-margin SKUs

• disrupts enterprise sales narratives

• complicates investor messaging

• introduces volatile revenue sources.

On the demand side, customers (including the hyperscalers themselves) want a neutral compute layer that supports multi-cloud capabilities, without a “platform” agenda. Cost and balance sheet advantages (moving capex to opex) also exist. 

Neoclouds might also offer faster access to new silicon and more flexible or negotiable terms. 

In terms of value chain positioning, the hyperscalers will control:

• Developer ecosystems

• Data gravity

• AI platforms & models. 

The value for their customers will include convenience, integration and trust. 

The neoclouds, on the other hand, providing a merchant compute layer, will provide capacity arbitrage, 

specialized hardware and price-performance leadership. The value is raw compute, predictable economics and speed to deployment.

Era	Hardware, Infrastructure	Systems, Platform Layer	Operating System	Middleware,  Runtime	Applications	Services,  Intermediation
Mainframe (1960s–1970s)	Vertically integrated mainframes (IBM-dominated)	Proprietary system architectures	Proprietary (IBM OS/360, etc.)	Embedded in OS	Enterprise custom apps	Systems integrators, time-sharing bureaus
Minicomputer (1970s–1980s)	DEC, HP, Data General	Vendor-specific platforms	UNIX variants, VMS	Early databases, transaction monitors	Departmental apps	VARs, integrators
Client–Server (1980s–1990s)	Commodity servers (x86)	Wintel standard	Windows, UNIX	Databases (Oracle), app servers	Enterprise packaged software	Hosting, VARs, IT outsourcers
Early Cloud (2000s–2010s)	Hyperscale data centers	Virtualized compute platforms	Linux	Cloud middleware, containers	SaaS	MSPs, CDNs, colocation
Mature Cloud (2015–2022)	Hyperscalers dominate scale	IaaS / PaaS platforms	Linux	Kubernetes, managed databases	Cloud-native SaaS	MSPs, FinOps, cloud brokers
Emerging AI Era (2023– )	Accelerators (GPUs, TPUs, ASICs); power & data centers	Hyperscale AI platforms + neocloud capacity	Linux	ML frameworks, inference runtimes	AI-native apps, copilots	Neoclouds, AI infra brokers, model hosts

So there is reason to believe that neoclouds will emerge as a permanent part of the AI compute value chain, supplying:

• Merchant GPU capacity

• Independent AI compute

• Pricing-led infrastructure specialists/
  
  

The value chain seemingly always creates a layer where price discovery, specialization, and customer leverage are the values. Neocloud is that layer, some will argue. 

And while enterprise compute will be part of the market, much of the current market is driven by compute needs of the hyperscalers themselves. 

Company	Percentage from Hyperscalers	Key Details/Notes
CoreWeave	~80-100%	Primary revenue from hyperscalers and AI labs. Microsoft alone: 62% (2024 full year), rising to ~70-72% in early 2025 periods. Top 2 customers (likely Microsoft + Meta/OpenAI): 77% in 2024. Additional contracts with Meta ($14B+), OpenAI, and others. Acts as overflow capacity for hyperscalers.
TeraWulf	~14-20% (growing rapidly)	Primarily Bitcoin mining revenue; HPC/AI (hosting for hyperscalers via partners like Fluidstack/Core42, backed by Google) contributed ~14% in recent quarters, with major multi-year contracts ramping in 2025-2026.
CleanSpark	~0-5% (early stage)	Still primarily Bitcoin mining (>95% revenue). Pivoting with AI data center hires and site wins (e.g., beat Microsoft for Wyoming site), but minimal hyperscaler revenue recognized yet; focus on future diversification.
Hut 8	~10-30% (growing rapidly)	Shifting from mining; major 15-year $7B+ lease (potentially $17B+) with Fluidstack (Google-backed) for AI hosting starting ramp in 2025. Earlier GPU-as-a-Service for AI clients; hyperscaler deals driving pivot.
Others (e.g., Core Scientific, IREN)	20-50%+ (varies)	Similar miners pivoting: Core Scientific ~21-30% from HPC (deals with CoreWeave/hyperscalers); many in 10-30% range amid transition. Full pivot companies approach 100%.

Some might question the “permanence” of neocloud providers in the “AI compute as a service” space, but current thinking tends to be that a new role within the value chain is being created. 

Analysts view NeoClouds as emerging with enduring roles through specialization, partnerships, and niche dominance, rather than widespread buyouts.

Hyperscalers (Microsoft, Google, Amazon, Meta) prefer massive long-term offtake contracts and partnerships to secure capacity quickly, while building their own infrastructure. This hybrid approach allows them to use NeoCloud balance sheets for off-balance-sheet scaling without full integration risks.

Others might argue that the window for neoclouds is somewhat less certain, to the extent it is driven by hyperscale inability to rapidly supply the current demand for AI compute. Eventually, the argument goes, the hyperscalers will be able to build and operate their own internal capacity, reducing reliance on neoclouds. 

Source/Estimate	Timeframe	Unmet/Shortfall Capacity	Key Notes/Reasons
McKinsey	By 2030 (incremental 2025-2030)	~125-205 GW (AI-related global)	Total AI demand 156-260 GW by 2030; hyperscalers capture ~70%, but build lags due to power/grid.
CBRE / Utility Requests	US hyperscale 2025-2026	~14-40 GW incremental (2025 surge)	Vacancy at record low 1.6%; requests far exceed grid additions; multi-year delays in key markets.
Seaport Global / Industry	Near-term (2025-2027)	Significant GPU/power shortage	NeoClouds fill "shortage of graphics chips and electricity"; temporary 3-5 year window.
NVIDIA / Analyst Backlogs	Blackwell supply 2025-2026	3.6M units backlog (hyperscalers)	Sold out through mid-2026; drives outsourcing to NeoClouds for immediate access.
Overall Analyst Consensus	2025-2028	Tens of GW + millions of GPUs unmet soon	Power as #1 bottleneck; hyperscalers' $350-600B annual CapEx still constrained by grid/energy.

When scale providers win on unit economics, merchant or brokerage layers appear wherever customers value flexibility, neutrality, or pricing innovation. In the case of AI compute, hyperscale AI compute suppliers, no less than enterprise customers, will have such needs. 

Content delivery networks provide a good example of how new specialist roles can emerge. CDNs are specialized data centers whose value is edge location and latency reduction for media and content delivery. 

Ironically, Home Owner Solar Power Damages Universal Service

A shift of consumer electrical service charges from “usage” to “a connection fee,” as happened in the telecom industry, will upset traditional thinking about how to support an effective and yet affordable electrical grid under changing usage patterns. 

Some will argue that such a switch is harmful to lower-income customers. Others will argue higher electricity costs are equally harmful to lower-income customers. As always, there are trade-offs. 

Assume a traditional utility rate that might include charges for fixed and variable costs (generation and customer usage) of:

• Fixed costs: $100/month (embedded)

• Energy rate: $0.20/kWh

• Average bill (500 kWh): $100

But self generation even by consumer customers changes the business model, to say nothing of large business customer self generation. 

If, for example, a home owner installs a solar system, the grid-delivered electrical usage bill could fall to $20 to $30 a month. But the common costs of maintaining the grid do not change. Eventually, that causes a capital recovery problem. 

So assume pricing changes to include a heftier “access to the grid” fee, with lower usage fees, possibly including a grid access fee of perhaps $60 per month; usage of $20 per month at $0.08/kWh. 

The customer still pays $80 per month to remain connected to the grid and the utility does not go out of business. 

The telecom industry had to make this explicit shift over the last couple of decades as usage of fixed networks dropped dramatically, replaced by use of mobile networks. At the same time, demand for the core “voice service” changed as internet access became the anchor product on the fixed network. 

So today the fixed network is supported more by fixed “access” charges than “usage.” And even at that, policymakers argue that fixed cost recovery mechanisms are insufficient. 

Still, the advantages are that explicit revenue mechanisms to cover shared and fixed costs are available, even as more customers remove themselves from the fixed network. 

Subsidies are targeted and transparent and cross-subsidies are policy choices, not accidents. Even on the replacement mobile networks, the widespread use of flat fee access, with “unlimited” data usage, national voice calls and text messages, show the reliance on “access to the network” pricing, rather than “minutes of use” or “bytes consumed” usage models. 

The implications for the electricity network, as more customers move to self generation, is rather obvious. The sunk costs of the grid must be paid for, irrespective of individual customer usage. 

“Access to the network” becomes the “product,” rather than usage.

What’s really happening is a decoupling of value from volume, something that also happens in other infrastructure contexts. 

The grid’s value is optionality and insurance, but it’s priced like a commodity pipeline. Distributed generation exposes that mismatch.

As for the argument that access fees hurt low-income customers, consider today’s situation, where solar power benefits homeowners with the means to self generate; living in sunnier climates. 

Renters, those without capital or physical means to generate their own electricity and people living in less-sunny climes are disadvantaged. 

Access fees help ameliorate such problems, while still protecting an electrical utility’s ability to build and maintain universal access networks under conditions where “best customers” are creating their own substitutes. 

Access fees, rather than usage, now dominate telecom service fees. People actually pay for ability to use the networks, not the amount of usage of those networks, which once was the case. 

If electrical energy networks must have a “universal service” character, then we also have models for ensuring such access for lower-income customers. We use subsidies.

Electricity Business Can Learn from Telecom Evolution

Oddly enough, local electricity generation by businesses and homeowners exposes a key problem for electricity supplier economics. Traditional pricing assumes energy consumption is equal to grid usage. 

But distributed generation breaks that assumption. Essentially, customers remove themselves, at least partially, from the system, but retain the optionality of using the grid for reliability, backup, and peak load balancing. 

But fixed costs stay embedded in the price of per-kiloWatt hour charges, so rates will rise as sales fall. At the same time, new demand driven by high-performance computing and associated data centers increases the need for new investments in transmission infrastructure as well as generation, increasing the fixed costs. 

The basic problem is a combination of high fixed costs; low marginal costs per additional kWh and the impact on ability to cover fixed costs when demand is reduced by local generation. 

Since fixed costs do not decline proportionally with local generation, all remaining sales must cover more fixed cost per kWh consumed. 

This pushes per-kWh rates upward for customers who still rely heavily on the grid. 

But the network still must be designed for peak load, sized to serve customers when solar output drops (night, winter, clouds). So self-generation reduces energy delivered, not the need for the grid.

Share of Customers with On-Site Generation	Utility Retail Sales (as % of original)	Fixed Cost Recovery per kWh	Average Retail Rate Impact for Non-Solar Customers
0% (baseline)	100%	$0.10/kWh	Baseline
10%	93%	$0.108/kWh	+8%
25%	82%	$0.122/kWh	+22%
40%	68%	$0.147/kWh	+47%
60%	52%	$0.192/kWh	+92%

What’s really happening is a decoupling of value from volume, something that also happens in other infrastructure contexts. 

The grid’s value is optionality and insurance, but it’s priced like a commodity pipeline. Distributed generation exposes that mismatch.

So what might be done to fix this problem? Fixed monthly connection charges are one way of “socializing” grid costs. Time-of-use pricing and demand charges also can help. But as with mobile and fixed telecom networks, “access” to the network might be more important than usage charges. 

So reframing the product might be conceptually necessary. The “product” electrical utilities sell is reliability, capacity, and load balancing, not just energy. 

Energy is a commodity that is part of the service, but grid access becomes the actual “product.” 

Beyond all that, perhaps more explicitly cross subsidies are needed, as once was the case for communications services, where business user profits subsidized consumer usage. Perhaps business customers and self-generators must subsidize customers unable (for financial or physical reasons) to participate in self generation. 

Until pricing reflects capacity and availability, not just kWh, rising self-generation will continue to raise rates for those most dependent on the grid.

This reminds me very much of how economics of the “telecom” business changed with competition. 

Both electrical grids and telecom networks have the same core traits:

• Extremely high fixed costs

• Very low marginal cost per additional unit of usage

• Peak demand, not average usage, drives capital investment

• Universal-service expectations layered on top of commercial economics

Historically, both industries solved this with implicit cross-subsidies. But widespread technology changes and deregulation changed the telecom business model. 

Traditionally, high prices for business customers (especially long distance calling) provided the profits that allowed affordable service for consumers. 

This worked as long as high-margin users couldn’t easily bypass the network and suppliers had pricing power. 

Self generation in the electricity business has the same dynamics. When high-value customers (commercial, industrial, affluent residential) can self-generate, electricity providers lose the profits that allow them to serve mass-market customers reliant on the grid with affordable rates.

The cross-subsidy that once flowed invisibly is exposed. The analogy with telecom after deregulation, mobile substitution for fixed voice, embrace of internet protocol and reliance on internet access as a core service for the fixed network illustrate the issues. 

Dimension	Traditional Telecom Access	Electric Grid (Emerging)
Core asset	Nationwide access network	Transmission & distribution grid
Cost structure	High fixed / low marginal	High fixed / low marginal
What drives capex	Peak simultaneous usage	Peak demand & reliability
Primary pricing unit	Minutes / lines	kWh
Implicit subsidy source	Business & long-distance margins	High-usage / high-income customers
Subsidy recipient	Residential & rural users	Low-income & non-solar customers
Bypass mechanism	VoIP, wireless, OTT apps	Rooftop solar, storage, microgrids
Resulting problem	Access prices no longer cover costs	Volumetric rates no longer recover fixed costs
Regulatory response	Access charges, USF fees	Grid access charges, demand charges (emerging)
Political constraint	Universal service obligation	Universal service + decarbonization goals

The problems are similar. Neither industry can simultaneously have volume-based pricing; high fixed costs; widespread abandonment of the core network and stable rates for mass-market customers. 

The telecom industry adapted by shifting its revenue model. Today,  customers do not primarily pay for minutes or megabytes anymore. They pay for “access to the network.” Think of it like Wi-Fi access. One pays to be connected, not for usage (bytes consumed or time connected or bandwidth provided). 

The analogy is a mobile phone service plan offered at a flat fee per month that includes “unlimited” data usage; “unlimited” national calling and text message. 

The customer pays for the ability to use the network, not consumption in a strict sense. 

Today’s electrical energy service problem is that self generation reduced kWh sales while fixed costs remain. As rates rise to cover fixed costs borne by fewer customers, there is more incentive to defect. 

So an access-fee model more effectively recovers shared fixed costs. So self generation no longer erodes fixed cost recovery. And the grid stays healthy.