Key facts at a glance

• AI is increasing demand for cloud computing, but the biggest visible surge is in physical infrastructure such as chips, networking gear, power systems, and data centers.
• Large US technology companies are expected to spend about $650 billion on AI-related infrastructure in 2026, up from roughly $410 billion in 2025, according to industry analysis.
• Nvidia plans to invest $2 billion each in photonics companies Lumentum and Coherent, a signal that data movement and networking are becoming as important as raw compute.
• Public cloud remains the fastest place to experiment with AI because it offers immediate access to GPUs, foundation model APIs, vector databases, orchestration tools, security controls, and integration services.
• Once AI workloads reach production scale, cost, utilization, data gravity, governance, and network traffic can push enterprises to reconsider where those workloads run.
• On-premises infrastructure and specialized neocloud providers are emerging as alternatives for steady-state, cost-sensitive, or compliance-heavy AI workloads.
• The likely future is a segmented market in which public clouds, on-premises environments, and neoclouds each play a role in the AI life cycle.

The infrastructure race behind AI cloud demand

AI is clearly accelerating demand for cloud computing, but not in the tidy, software-led pattern that many executives expected. The most important story is not a new application layer or a fresh productivity suite. It is the extraordinary amount of capital flowing into the physical foundation required to support AI at scale. Chips, networking equipment, power systems, cooling, and massive data centers are becoming the strategic center of gravity for the cloud market as providers race to support model training and inference workloads.

The spending figures explain why this shift matters. Large US technology companies, including Alphabet, Amazon, Meta, and Microsoft, are expected to spend about $650 billion on AI-related infrastructure in 2026, up from roughly $410 billion in 2025, according to analysis cited by industry observers. That kind of growth is not a normal upgrade cycle. It suggests that AI is forcing a redesign of the cloud stack itself, from the processor to the rack to the building that houses the cluster.

That redesign reaches deep into networking and data movement. Nvidia recently announced plans to invest $2 billion each in photonics companies Lumentum and Coherent, underscoring where pressure points are emerging. The limiting factor is no longer only raw compute. It is also how quickly data can move between processors, racks, and clusters without creating unacceptable bottlenecks or power inefficiencies. As AI systems scale, latency, throughput, and energy usage become first-order economic concerns rather than secondary engineering details.

Why most AI starts in the public cloud

When companies are experimenting, speed matters more than optimization. Public clouds give teams immediate access to GPUs, foundation model APIs, vector databases, orchestration tools, security controls, and integration services. They also allow businesses to start pilots without waiting for procurement cycles, data center expansions, or specialized infrastructure teams. For many organizations, that speed is the difference between testing an idea this quarter and debating it for the next year.

Given the high level of uncertainty, the public cloud is often the right choice for first-generation AI. Enterprises do not yet know which use cases will deliver value, how much inference traffic they will see, or which architecture will ultimately survive. At this stage, the ability to try many things quickly is more important than squeezing every dollar from the underlying infrastructure. Managed services reduce friction, and friction is the enemy of early adoption.

This is why strong initial demand for AI is landing in public cloud environments. Enterprises are building chatbots, copilots, knowledge assistants, document automation systems, and code generation tools there because the cloud lowers the barrier to entry. It provides compute as well as a full operating environment for AI experimentation. The public cloud also offers a global footprint and a broad set of supporting services that few individual enterprises can replicate on their own.

When production economics change the calculus

The second generation of enterprise AI systems looks different. Once a use case proves its value and usage becomes persistent, the financial model changes. A workload that looked inexpensive during a proof of concept can become shockingly expensive when it runs at production scale, especially if it depends on premium GPU instances, high-performance storage, constant network traffic, and managed services layered on top of one another.

That is where repatriation enters the conversation. A pattern is beginning to emerge in which enterprises build first-generation AI systems on public clouds, learn what works, and then move some workloads back on-premises or onto specialized neocloud providers that offer AI-optimized infrastructure at a lower cost. The move is not a rejection of cloud computing. It is a recognition that different stages of the AI life cycle have different economic and operational requirements.

On-premises deployment is attractive when utilization is steady, data gravity is high, governance requirements are strict, and the organization has enough scale to justify owning or directly controlling the infrastructure. Neocloud options become attractive when enterprises still want an external provider but do not want to pay the full premium often associated with large hyperscalers. These specialized providers are increasingly positioning themselves around dense GPU capacity, simpler pricing, and architecture built specifically for AI rather than for general-purpose enterprise IT.

This adoption pattern dispels the old assumption that cloud migration is always one-way. In the AI era, workload placement is becoming more fluid. Enterprises are learning that the best place for experimentation may not be the best place for steady-state production and that AI economics can punish architectural laziness much faster than traditional enterprise applications ever did.

Neoclouds and the return of workload placement

The rise of neocloud providers is one of the clearest signs that the market is segmenting. These providers are not trying to be all things to all enterprises. They are focusing on AI-specific requirements such as dense GPU capacity, high-speed interconnect, predictable pricing, and simplified access to accelerated computing. For enterprises that want external capacity without paying every premium associated with a broad hyperscaler platform, neoclouds can be a practical middle option.

At the same time, on-premises infrastructure has regained relevance for workloads that are cost-sensitive, steady-state, and compliance-heavy. Many enterprises already own data centers or have hybrid arrangements that can be extended to support AI inference. When data cannot leave a facility, when latency requirements are tight, or when utilization is high enough to justify capital investment, keeping workloads closer to the data can make both economic and operational sense.

How much AI demand will stay with hyperscalers?

AI will drive a great deal of demand for public cloud computing, especially in the near term. Every major enterprise AI initiative will likely engage the public cloud in a meaningful way, whether for model development, training bursts, integration services, security tools, or global deployment. But it would be a mistake to assume that all demand will remain locked in traditional hyperscalers over time.

Some AI workloads will stay in the public cloud permanently because they are bursty, globally distributed, hard to predict, or tightly coupled to cloud-native services. Other workloads, especially those with stable usage patterns and heavy inference volume, will be candidates for relocation. Economics will drive those decisions more than ideology. The likely outcome is a more segmented market. Public clouds will dominate the front end of AI adoption and continue to play a major role in hybrid operations. On-premises environments will regain relevance for cost-sensitive, steady-state, and compliance-heavy workloads. Neocloud providers will grow as a middle option for enterprises seeking external AI capacity without paying full hyperscaler prices.

Three factors enterprises should weigh

First, speed and cost are distinct metrics. The public cloud is usually the fastest way to get an AI initiative off the ground, and that speed has real business value. But the architecture that wins a pilot may end up destroying the production budget. Enterprises need a placement strategy from day one, even if they start in the cloud.

Second, AI workload economics differ from those of traditional applications. Training, inference, data movement, storage, and model serving can interact in ways that quickly create cost surprises. Organizations should model not only compute usage but also utilization patterns, network flows, and the costs of managed services surrounding the core AI stack. Without that discipline, they risk designing systems that are technically elegant but financially unsustainable.

Third, future flexibility matters more than short-term convenience. Enterprises should avoid building AI systems so tightly around a single provider's proprietary stack that moving becomes painful or impossible. The winners in this market will be the companies that preserve optionality, enabling them to shift workloads across public clouds, on-premises environments, and emerging neocloud platforms as economics, regulations, and business requirements evolve.

What the next phase of AI adoption will look like

The next phase of AI adoption will be less about a single destination and more about matching workloads to the right environment. Public cloud will remain the default starting point for experimentation and for workloads that benefit from elasticity, global reach, and managed services. On-premises infrastructure will handle workloads that require tight control, predictable performance, or close proximity to data. Neoclouds will fill the gap for enterprises that want specialized AI capacity with simpler economics.

The practical question for enterprises is not whether AI belongs in the cloud, but which workloads should stay there, for how long, and under what conditions they should move. AI will unquestionably generate significant new demand for public cloud computing. For most enterprises, AI workloads will stay in the cloud long enough to enable rapid innovation, but they will not necessarily remain there forever. The organizations that plan for portability, cost visibility, and workload placement from the beginning will be better positioned than those that assume the first deployment model is the permanent one.

Source:InfoWorld News