Three years into the AI era, the market has advanced quickly, but the value pyramid remains bottom-heavy

Although the generative AI (GenAI) cycle began three years ago, the market is still translating that progress into broad-based economic value. Model performance has improved, infrastructure investment has accelerated, enterprise experimentation has broadened, and buyers are recognizing value from early deployments. This progress reinforces TBR’s early view that AI would reshape enterprise work over time through workforce augmentation, task automation and role redesign. Although the long-term promise of AI has not dimmed, realizing the technology’s potential will take longer than was initially expected. The issue is whether enough value can move into the software, services and workflow layers where enterprise productivity is realized. That is the central question about the next phase of the AI market: When — and how meaningfully — will the AI value pyramid begin to invert?
 

 

AI infrastructure dominates the broader opportunity, with AI services representing only a small portion of the market today

The AI story remains trapped in a phase dominated by infrastructure investment, but early success in coding platforms offers a glimpse of how AI services revenue will scale over time. GPU and memory chip makers are capturing the largest revenue pools at the base of the pyramid, hyperscalers and neoclouds are monetizing compute delivery in the middle, and AI services providers are capturing a comparatively smaller share at the top. This imbalance is a sign that the cycle remains in its early stage. The first wave of monetization has accrued to the vendors supplying the capacity required to train, host and scale AI. The next wave will need to come from production usage, recurring inference consumption and AI services that deliver clear enterprise outcomes.
 
Over the long term, the AI value pyramid will need to invert. The market cannot remain structurally dependent on infrastructure build-out if AI is to become a durable enterprise productivity platform. Value will need to move closer to where work is executed, governed and measured. That is the future hyperscalers are investing in as large cloud commitments continue to flow from model developers such as Anthropic and OpenAI. The current distribution is rational for an early technology supercycle, but it cannot be the steady-state structure of the market indefinitely.

The AI value pyramid will need to invert over the long term

To put rough numbers to the AI value pyramid, across the three buckets of semiconductors, hyperscale delivery and AI services, semiconductor vendors are still capturing the lion’s share of the market. As noted in TBR’s 4Q25 AI Infrastructure Market Landscape, AI server and systems revenue reached an estimated $235 billion in 2025, inclusive of NVIDIA, AMD and hyperscalers. By comparison, TBR estimates 2025 AI cloud delivery and platforms revenue was under $50 billion, inclusive of Amazon Web Services (AWS), Microsoft, Google Cloud and ServiceNow, while AI services revenue was around $25 billion, inclusive of OpenAI, Anthropic, Microsoft, AWS, Google Cloud and ServiceNow.
 
Put another way, the foundational layer of the AI market saw nearly 10x the revenue opportunity of AI services providers in 2025. This gap is the clearest evidence that the AI market remains early in its value shift. Infrastructure monetized first because the market needed capacity before broad production usage could scale. Over the long term, however, this structure will need to change. If AI is to realize its full enterprise potential, the revenue generated via inference consumption on AI services, workflow management and agentic execution needs to increase.
 

 

AI end-user perception remains positive, supporting future adoption

Upfront investment ahead of long-term opportunity is normal in a technology supercycle, but the AI services market needs to expand for the broader AI opportunity to be sustainable. Hyperscalers are investing for this future, and TBR’s customer research continues to point to an AI world where customer appetite is growing as early adopters recognize more value than expected from AI-related projects.
 
In TBR’s upcoming 2H25 AI Applications Customer Research, 55% of respondents are increasing their IT budgets to support AI adoption, up from 48% in 2H24, while 58% of respondents stated AI tools exceeded their expectations for value creation. With buyers seeing value and adjusting budgets accordingly, TBR continues to view end-user demand as robust. The demand side of the pyramid is not the main constraint, but it remains to be seen how quickly vendors can convert that appetite into production usage, recurring consumption and AI services revenue.

Agentic systems and usage-based revenue will accelerate revenue growth, helping close the value gap

The AI services market will not scale simply because the models improve. Instead, the market will scale when AI systems become embedded deeply enough in enterprise workflows to generate sustained inference demand. Agentic systems are the clearest path to that outcome. Unlike first-generation copilots, which primarily assist with discrete tasks, agents are designed to execute multistep work across applications, data sources and governance layers. That matters economically because every step in an agentic workflow creates additional model calls, tool calls and consumption events.
 
This shift gives AI services vendors a more credible path to monetization. Fixed-seat copilots can demonstrate productivity, but they disconnect usage from revenue and can pressure margins when consumption rises. Agentic systems practically require usage-based pricing because the work performed is more measurable and the consumption profile is more directly tied to business activity. In this model, AI revenue growth comes not from selling more experimental seats but from running more enterprise work through AI-enabled execution layers.
 
Infrastructure capacity is being built ahead of demand, but that investment will only be justified if inference workloads expand across real production environments. The vendors best positioned to capture that opportunity will be those that control where AI work is governed and managed. Model quality remains important, but workflow control is becoming the more durable source of value.

Code-generation tooling is becoming the first scaled, enterprise agentic AI market

This thesis is already playing out in software development, which is the first scaled example because it combines high labor cost, measurable productivity gains, structured workflows and clear willingness to pay. The adoption of AI tools has significantly increased software developers’ productivity, accelerating code generation and reshaping expectations around the cost of producing software. Executives from several notable software incumbents, including SAP and Salesforce, have boasted about their ability to limit hiring for new developers, with Salesforce leaders stating the company did not hire any in 2025. Software developer job postings on Indeed remain nearly 70% below their highs in 2022, reinforcing the view that the market for developer talent has shrunk significantly over the past three years, while revenues from leading AI code-generation tools, including Claude Code and OpenAI Codex, grow in the triple digits year-to-year.
 
The market for AI-powered development platforms is also moving toward agentic systems very quickly, trusting tools to leverage codebases, generate changes, test outputs, open pull requests and operate across parts of the development life cycle. As such, code generation is the first scaled enterprise proof point for agentic AI. It is a market where the productivity impact is measurable, user adoption is already material and monetization is scaling very rapidly. It also shows why execution control matters. GitHub has an advantage because it owns the repository, pull request, CI/CD (continuous integration and continuous delivery) and governance workflow. Cursor is gaining traction as an agent workspace. Anthropic and OpenAI are competing through model-native coding agents in Claude Code and Codex. The market is a contest not only between models but also across the workflow control layer, the agent workspace and the model-native execution layer.

Conclusion

The AI market is not failing to meet its promise, but it is still early in translating that promise into broad-based economic value. Three years into the GenAI cycle, the largest pools of revenue remain concentrated in the infrastructure layers required to train, deploy and scale frontier models. That concentration is rational given the scale of compute demand, but it is not a sustainable end state if AI is to become a true enterprise productivity platform.
 
The next phase of market development will depend on whether the AI value pyramid can begin to invert. Infrastructure investment has created the capacity, but inference demand will need to scale through production usage, workflow integration and consumption-based pricing. Agentic systems are the most credible path to that outcome because they increase AI usage, extend AI into multistep workflows and create a stronger foundation for usage-based monetization. Improvements in code generation show that the market is beginning to change, but broader enterprise adoption will require the same combination of measurable productivity, workflow integration, governance and pricing alignment.
 
The AI value pyramid should begin to shift over time, but not simply toward model developers. Durable value will accrue to the vendors that control execution. That includes model providers building agent platforms, hyperscalers hosting and metering workloads, SaaS vendors embedding agents into business processes, developer platforms controlling software workflows, and services firms operationalizing AI across complex enterprise environments. For model providers specifically, this means competitive positioning can no longer be measured by benchmark performance alone. Revenue mix, infrastructure access, enterprise adoption, partner alignment, agentic capabilities and workflow integration are becoming more important indicators of vendors’ positioning long-term.

U.S. Analyst and Advisor Meet 2026, New York City, March 5, 2026 — Infosys hosted industry analysts and advisers for an afternoon in the newly branded Infosys Theater within Madison Square Garden. Using client stories amplified through technology partner support to reinforce Infosys’ role in the IT services, cloud and enterprise AI market, company executives consistently noted that enterprise AI success depends on combining strong data foundations, responsible governance, talent transformation, domain-specific use cases and partner-led execution, which can help turn AI from isolated experimentation into measurable business value.  

 

Americas’ scale and market maturity present an opportunity for Infosys to treat AI as a horizontal capability rather than an ad hoc solution

Similar to previous meetings, the event began with an update on the company’s strategy and performance in the Americas region. Anant Adya, EVP and head of Americas Delivery, led the presentation, highlighting key elements of the company’s success in the region, such as its hub-first strategy, including the addition of a hub in Costa Rica. Adya described the hubs as serving four roles: innovation, client cocreation, centers of excellence, and training and enablement. He specifically mentioned the hub in Hartford, Conn., which focuses on insurance and healthcare domain solutions and client prototyping, as well as a ServiceNow Center of Excellence, which focuses on onboarding and training local university students and other talent.
 
A key nuance in Adya’s presentation at last year’s event was Infosys’ positioning of AI and moving from the earlier pentagon-shaped strategy to a new hexagon-shaped strategy centered on AI. The message was that AI is no longer a horizontal add-on but is now meant to reshape the full services value chain. He emphasized that enterprise data readiness is the critical prerequisite, with cloud already assumed and data becoming the true launchpad for AI at scale. He used several client examples to illustrate the strategy in action: AI-infused managed services for a packaging company, AI-first manufacturing operations for a semiconductor client, AI-led SAP transformation for a utility, and AI-enabled wealth operations for a financial services firm. Across those examples, the focus was on measurable business outcomes like productivity, manufacturing uptime, revenue impact, and value creation, not just IT efficiency.
 
TBR appreciated these examples as we see them as a bridge into opportunities for Infosys to drive awareness of its broader capabilities and strengthen relationships among enterprises. We believe that pivoting from driving AI conversations centered on efficiency improvement to discussing business growth at scale will pressure-test Infosys’ commercial model readiness, especially as the former leads buyers to expect perpetual cost savings, thus pressuring Infosys’ margins, while the latter provides a pathway for increasing volume, delivered by fewer employees.
 

 
Infosys has maintained an operating margin within its guided range of 20% to 22% for a few years, yet headcount growth began to rebound in 4Q25, at 4.2% year-to-year, following several quarters of decline. Additionally, grouping partners into strategic foundation partners, conventional core partners, and innovation-edge AI partners highlighted Infosys’ recognition of the value of the ecosystem and the importance of where each party plays. Aligning to core strengths from both a capability and a messaging perspective can help Infosys and partners demonstrate depth and strengthen trust with buyers who seek mutual accountability and understanding of go-to-market priorities.

Crystalizing AI talent strategy will provide Infosys with the necessary support to drive business value at scale

Following Adya’s strategy update, Ranjana Joshi, AVP and HR leader, Americas, discussed how Infosys is redesigning its workforce model for an AI-first services business. Although TBR has previously heard some of the details Joshi outlined, it was apparent that Infosys has crystallized what comes next in developing an AI-ready workforce. Infosys is building the talent strategy around three pillars: a new talent operating model, which she described as an “ambidextrous” model; a new career model; and a new talent development model.
 
According to Joshi, these pillars are set to help Infosys do two things at once: “augment the whole organization with AI while also building deeper engineering and domain expertise to deliver AI-first services.” Under its new “ambidextrous” approach, Infosys is changing how it hires. The company is shifting toward recruiting more specialized talent, including specialist programmers from U.S. campuses, alongside continued hiring of full-stack engineers and domain experts.
 
Additionally, Infosys continues to invest in internal bridge programs and hands-on sandbox-based development paths to ensure it develops the right-skilled bench. Meanwhile, career paths are being redesigned away from the traditional linear ladder. Instead of the old one-dimensional path from engineer to executive, Infosys has set up a Y-shaped career structure. One track is for the broad workforce that is AI-augmented, and the other is a specialist stream for people with deep engineering, domain and functional expertise. Further, the company is adding distinguished specialists in roles such as AI strategist and responsible AI engineer. These professionals are meant to act as catalysts who accelerate innovation and create specialist ecosystems around them.
 
Although these are important investments, especially as Infosys looks to build messaging centered on AI-readiness at scale, peers often cite numbers that suggest the majority of their workforce is AI-ready, challenging it to differentiate. Infosys’ opportunity lies in its ability to develop an AI-ready sales bench, especially as the technology forces it to act more like a software company than a traditional services company. We view Infosys’ investments in forward-deployed engineers (FDEs) as a bridge into an AI-ready sales force. Striking the right balance between pushing AI-first sales and managing relationships with ecosystem partners will be key, as going too far into selling tech could sour relationships with partners that try to do the same.

Helping clients scale, orchestrate and operationalize AI for business value will test Infosys’ commercial and operating model readiness

Throughout the rest of the afternoon, Infosys’ executives, clients and partners continued their efforts to separate hype from reality regarding scaling enterprise AI adoption. Compared to previous events, Infosys’ emphasis on domain-specific use cases amplified through panel discussions across financial services, energy, insurance, consumer packaged goods and life sciences strengthened the company’s message and AI strategy.
 
Joydeep Mukherjee, EVP and global head, Data & Analytics, Digital & Creative Services; and Srinivas Gopal, VP and Global Business Head – Data, Analytics & AI, noted that the market has moved beyond AI experimentation and is now demanding measurable enterprise value. Mukherjee’s core point was that most companies still struggle to scale AI because they are held back by weak strategy, fragmented data, governance gaps, integration challenges, and immaturity in their operating models. Scaling AI requires a coordinated blueprint spanning value discovery, operating model design, technology readiness, performance management, responsible AI and talent transformation.
 
Gopal built upon that point by showing what the next stage looks like in practice: agentic AI applied to real business processes, where enterprises use data products, intelligence layers and orchestrated agents to improve decision making, accelerate workflows and create business outcomes faster. Across both sessions, the message was that enterprise value does not come from isolated pilots or stand-alone agents but rather from combining strong data foundations, contextual business knowledge, governance and production-scale execution. This message also highlighted the breadth of Infosys’ capabilities in helping customers make that shift with Infosys Topaz as the orchestration layer.
 
In TBR’s 4Q25 Infosys report, we wrote: “Infosys’ recent cluster of GenAI [generative AI] announcements signals a deliberate pivot from AI-enabled services toward an agent-first delivery platform strategy, with Topaz Fabric positioned as the control plane. By pairing Fabric with hyperscaler-native capabilities like Amazon Q Developer for IT modernization and productivity, while embracing emerging AI software engineer tooling through Cognition’s Devin and scaling vertical agents such as its energy-operations solution on the Microsoft stack, Infosys is effectively building a multipartner operating model that can meet customers where they are without ceding the orchestration layer.
 
For competitors, differentiation is shifting from headcount, pricing and even domain expertise toward agent operational maturity, where productizing services compresses delivery timelines and forces a margin reset. For alliance partners, the upside (and risk) is that fabric-owning SIs become high-leverage distribution for platforms and models, while increasingly establishing the integration patterns, governance standards and customer experience — shifting partnerships from joint marketing to control of the runtime and commercial attach adds value.”
 
According to TBR’s December 2025 Digital Transformation: Voice of the Customer Research report, the market is signaling that vendors are getting better at packaging and surfacing IP, but monetization is still anchored to services constructs, not product economics, especially as automation and AI start to do a greater share of the work.
 
Against that backdrop, Infosys’ recent AI investments look like an attempt to bridge this gap rather than leap over it. The inflection point to watch is whether Infosys can translate its agent-first strategy into commercial pioneers that buyers can procure — clear SKUs, governance guarantees, and outcome and/or consumption mechanisms that do not simply relabel labor. If Infosys can prove repeatable productivity and quality gains and then price around outcomes with credible controls, it can move from AI-enabled services to a platform-led model that aligns with where buyers say they want vendors to go but have not consistently shifted toward yet in terms of their procurement behavior.
 
Alternatively, Infosys’ AI announcements can be interpreted as commercial positioning ahead of change. Topaz Fabric, agents and FDEs absolutely help industrialize delivery, but they also create a convenient story in which Infosys can claim platform-led differentiation while still monetizing primarily through large, multiyear services contracts. The recent launch of AI-enabled global capability center (GCC) framing, in particular, can be read in one of two ways: Either it is a structured route to platformizing captive operations, or it is a mechanism to lock in demand and defend wallet share by embedding Infosys’ tooling, processes and governance so deeply that switching becomes hard without changing the pricing model as much as the marketing implies. The near-term expectation is that Infosys will capture value the way the industry typically does during transitions: sell transformation rhetoric, deliver productivity gains, and then negotiate hard to keep most of the savings — or recycle those savings into scope expansion rather than price reductions.

Domain-aligned services backed by relentless and quality service execution will help Infosys sustain trust with key buyers

Across the panel discussions, speakers noted that most organizations have already experimented with pilots and proofs of concept. The real challenge now is turning those isolated efforts into repeatable, enterprisewide capabilities. Panelists described this as an execution problem as much as a technology challenge, requiring alignment across strategy, operating model, governance, data and talent.
 
Within a financial services discussion, banking was presented as one of the sectors most ready to use AI because it can improve both growth and efficiency while also supporting resilience and compliance. The insurance panel showed how AI is being applied in a regulated industry through underwriting, pricing, unstructured data extraction, faster service and operational efficiency. The insurance leaders repeatedly stressed that in the human-plus-AI equation, especially in specialty insurance, AI can speed intake and improve accuracy, but the knowledge and judgment of experts still matter for decision making and evaluating risk.
 
The energy discussion centered on the growing complexity of energy and commodity markets and the role Infosys wants to play in that transformation. Speakers described energy trading and risk management as a stabilizing force in volatile markets, especially as geopolitical disruption, AI-driven power demand and shifting wholesale-retail dynamics make forecasting and execution more difficult. Infosys’ acquisition of MRE Consulting, which featured prominently during the panel, will play a key role in strengthening trust with existing buyers and expanding the company’s addressable market opportunity across EMEA and APAC. Infosys’ Energy, Utilities, Resources and Services (EURS) vertical share of revenue has hovered around 13% to 14% of Infosys’ total sales. We believe MRE Consulting, along with future similar investments in the vertical, will help boost that share to between 16% and 17%, similar to how manufacturing grew from around 10% of Infosys’ total sales in 2020 when it signed its megadeal with Daimler to now hovering close to 17%.

Balancing steady services-enabled revenue growth with new product sales channels will test Infosys’ otherwise strong culture

Over the next 12 to 18 months, we will keep a close eye on whether Infosys can convert AI breadth into depth, which will be evidenced by higher attach rates of Topaz Fabric in large programs, clearer packaging and commercial models for GCC platformization, and consistent outcome metrics (not just activity metrics) tied to modernization, data readiness and workflow redesign.
 
Infosys has an opportunity to stand out from its peer group if it can make Fabric the default operating layer for agent delivery and use FDEs to accelerate time to value, creating durable differentiation versus peers, as many of them are still largely focused on tooling plus services. Key risks include execution complexity, partner dependence blurring differentiation, and value-capture pressure if Infosys is forced to compete on price, relinquishing the value of the productivity gains it gives clients. Additionally, the use of FDEs, even on a smaller scale than Infosys’ traditional labor arbitrage pyramid, can raise questions about the company’s use of a time-and-materials model versus agent-based wrapped pricing.
 
Further, accelerating revenue from value-based selling would bolster profitability but compel the company to recalibrate its staffing pyramid across legacy and new areas, pressure-testing attrition levels. Lastly, quick-hit wins within the agentic AI space could entice Infosys’ leadership to pursue a more aggressive product sales strategy, disrupting its ongoing success with traditional services deals. In an intense and rapidly changing competitive market for AI-enabled IT services solutions, Infosys has a fighting chance to stand out. And we think the company’s leadership and recent success will help Infosys separate from peers over the next few years
 
TBR will continue to cover Infosys across the IT services, ecosystems, cloud, and digital transformation spaces, including publishing quarterly reports that assess Infosys’ financial model, go-to-market strategy, and alliances and acquisitions strategies. For a comparison with Infosys’ peers and other IT services vendors, TBR includes Infosys in our quarterly IT Services Vendor Benchmark, our semiannual Global Delivery Benchmark and Cloud Ecosystem Report, and our annual Adobe and Salesforce Ecosystem Report, SAP, Oracle and Workday Ecosystem Report, and ServiceNow Ecosystem Report. Click here to learn more about accessing the data and analysis within these reports.

Neocloud providers challenge a decade of hyperscaler control

Before the explosion of AI interest in 2023, the hyperscalers had enjoyed a nearly unimpeded growth trajectory since 2017, when the last of the telco cloud competitors exited the business. Verizon, IBM, Rackspace and a multitude of others all started their own cloud platforms in a bid to challenge Amazon Web Services (AWS) in the cloud infrastructure space but ultimately exited the space and focused on the periphery of the cloud market opportunity. Even the challenges posed by General Data Protection Regulation (GDPR) and other regulations were ultimately not enough to dislodge the major cloud providers, as the three leading U.S.-based firms still hold the largest market share, even in the heavily regulated European Union markets.
 

More than a decade after the market came into existence, the cloud infrastructure market remains a steady source of double-digit growth, with a total opportunity size of roughly $500 billion globally in 2025. Most of this market opportunity is claimed by AWS, Microsoft and Google Cloud, which together account for half of the total market. Although we expect the collective share of those three vendors to continue growing through 2029, neocloud providers will cause real disruption and limit the leading hyperscale cloud providers’ nearly unimpeded ability to expand. Although neocloud providers will not realistically capture leadership of the cloud infrastructure space, they will disrupt the hyperscaler market by capturing AI-related market growth, pressuring pricing for AI workloads, building a platform ecosystem around their services, forcing hyperscalers to partner with major neocloud providers, and directly targeting enterprise customers.

It is not just the revenue but also the growth that hyperscalers will miss

The most direct and visible impact of neocloud providers will be the revenue they generate. The neocloud market is estimated to have surpassed $25 billion in 2025, representing triple-digit growth year-to-year. That presents a significant opportunity and represents one of the fastest-growing segments of the overall cloud market. However, there is overlap, as many hyperscalers are also the largest neocloud customers, and the fact that this group of companies is capturing tens of billions of dollars and growing at a rapid pace is a complicating factor in the market. CoreWeave, for instance, earns nearly $2 billion in revenue each quarter, a figure that is roughly doubling year-to-year.

Neocloud providers will also pressure hyperscale pricing and margins

CoreWeave’s $2 billion in quarterly revenue is even more significant given the pricing advantages it offers its customers. Although the specifics vary based on a number of factors, in general, neocloud providers’ prices are 30% to 60% lower than what the major hyperscalers charge for the same services. That means CoreWeave is taking between $2.6 billion and $3.2 billion in market opportunity off the table from hyperscalers. The total revenue impact is even more severe after accounting for the pricing pressure those hyperscalers are forced to grapple with while trying to minimize the disparity between their AI service prices and neocloud offerings.
 

Hyperscalers’ expenses and margins will also reflect the impact of neocloud providers, as their operating models vary significantly, as shown in Figure 1. CoreWeave, for instance, is operating at basically a break-even profit level, investing internally, largely through R&D and infrastructure build-outs. Neoclouds’ increased pricing and investment pressures will also impact AWS’ double-digit operating margins.
 

Building out a platform will solidify neocloud competitive positioning

While selling access to the raw GPU capacity remains the primary way in which neoclouds are disrupting the hyperscale landscape, expanding into the platform layer of services will have a sustained impact. As GPU supply expands and competition intensifies, platforms have become critical for differentiation, customer retention and higher-margin services. This is a strategy straight out of the cloud hyperscaler playbook, as AWS, Microsoft and Google have entrenched themselves with customers through additional development, integration, marketplace and data services on top of their core infrastructure capabilities. Neoclouds, by layering orchestration, AI tooling and developer environments on top of specialized infrastructure, aim to move up the value chain and compete more directly with hyperscaler AI environments.
 

CoreWeave provides one of the clearest examples of this platform strategy. The company now positions its offering as a purpose-built AI cloud platform that combines high-performance infrastructure with intelligent software tools. The platform integrates services such as managed Kubernetes environments, GPU-native scheduling systems and AI storage layers designed to support large distributed training workloads. For example, the CoreWeave Kubernetes Service (CKS) and Slurm-on-Kubernetes orchestration stack allow customers to efficiently schedule massive GPU jobs and maximize cluster utilization.
 

Beyond orchestration, CoreWeave has expanded into tools that support the broader AI life cycle, including training infrastructure, inference deployment and integrations with AI development ecosystems. The company’s platform is designed to enable developers to build, train and serve AI models within a single environment, rather than relying on fragmented tooling across multiple providers. These additional services target some of the fastest-growing addressable markets for hyperscalers, large-scale AI training and inference. The impact becomes more distinct in the long term, as customer adoption of platforms is quite sticky, preserving neocloud advantages even as GPU scarcity and price-to-performance advantages potentially fade over time.

Hyperscalers are forced not only to compete with neoclouds but also to partner and purchase from them

Although there is a competitive element between hyperscalers and neoclouds, both groups also rely on each other to capitalize on the AI market opportunity. This reflects the explosive demand for AI compute and the limits of hyperscalers’ ability to build capacity quickly enough to meet that demand. Neocloud providers such as CoreWeave have built their businesses around specialized AI infrastructure and have an inherent advantage in the space due to their unique relationships with NVIDIA, which allow preferential access to supply-constrained GPUs. As demand for generative AI accelerates, these specialized environments have become valuable sources of additional compute capacity — even for the largest cloud providers, leading to a paradoxical relationship between hyperscalers and neoclouds.
 

On one hand, hyperscalers must compete with neoclouds for AI customers, particularly startups and AI labs seeking large GPU clusters at competitive prices. On the other hand, hyperscalers also benefit from accessing the infrastructure that neoclouds have rapidly built. In some cases, hyperscalers purchase compute capacity from neocloud providers to supplement their own data center supply while their internal AI infrastructure continues to scale.
 

This hybrid competitive and cooperative dynamic reflects a broader shift in the cloud market. AI infrastructure demand has grown so quickly that no single provider can fully control the supply of compute resources. As a result, the emerging AI cloud ecosystem is becoming more interconnected, with hyperscalers, neoclouds and infrastructure providers operating within a complex web of competition and collaboration. In the near term, this dynamic is likely to persist, limiting the aggressiveness with which neoclouds and hyperscalers compete.

 

Expanding to enterprise customers is the next stage of neocloud diversification

Although platform services and capabilities represent neoclouds’ functional diversification, enterprise customers represent diversification efforts within the neocloud customer base. The supply-constrained environment and tremendous amounts of investment made AI startups and hyperscalers lucrative early customers for neocloud providers. That concentrated customer base has created a significant market very quickly, but expansion and diversification represent the next phases in the market’s evolution. Given the areas of competition with hyperscalers, it is particularly important to expand the customer base to large enterprises that build and run their own AI models. To expand into enterprise customer accounts, neocloud providers are combining specialized AI infrastructure with enterprise-grade platforms and long-term capacity agreements that appeal to organizations deploying large-scale AI workloads.
 

For instance, in addition to providing GPU-dense infrastructure optimized for AI training and inference and large clusters of NVIDIA accelerators connected through high-bandwidth networking, CoreWeave is also targeting enterprise customers by building a platform around AI workload management and deployment. The company’s managed Kubernetes environments and GPU-aware scheduling capabilities allow enterprises to run large distributed training workloads more efficiently. Integrations with AI development tools and machine learning frameworks enable customers to build, train and deploy models within a unified environment rather than assembling multiple infrastructure and software layers independently. CoreWeave has also pursued enterprise adoption through large, multiyear compute agreements that give customers guaranteed GPU capacity while ensuring predictable revenue streams for the company. Although neoclouds’ efforts to diversify their customer bases are still new, the trend will eventually lead to greater stability for the providers as the AI and GPU markets mature.