AI Drives Cloud Player Capex Amid Cautious Overall Spend

  • Cloud service providers’ capex is expected to grow by around 8% YoY in 2023 due to investments in AI and networking equipment.
  • Microsoft and Amazon are among the highest spenders as they invest in data center development. Microsoft will spend over 13% of its capex on AI infrastructure.
  • AI infrastructure can be 10x-30x more expensive than traditional general-purpose data center IT infrastructure.
  • Chinese hyperscalers’ capex is decreasing due to their inability to access NVIDIA’s GPU chips, and decreasing cloud revenues.

New Delhi, Beijing, Seoul, Hong Kong, London, Buenos Aires, San Diego – July 25, 2023

Global cloud service providers will grow capex by an estimated 7.8% YoY in 2023, according to the latest research from Counterpoint’s Cloud Service. Higher debt costs, enterprise spending cuts and muted cloud revenue growth are impacting infrastructure spend in data centers compared to 2022.

Commenting on the large cloud service providers’ 2023 plans, Senior Research Analyst Akshara Bassi said, “Hyperscalers are increasingly focusing on ramping up their AI infrastructure in data centers to cater to the demand for training proprietary AI models, launching native B2C generative AI user applications, and expanding AIaaS (Artificial Intelligence-as-a-Service) product offerings”.

According to Counterpoint’s estimates, around 35% of the total cloud capex for 2023 is earmarked for IT infrastructure including servers and networking equipment compared to 32% in 2022.

Global Cloud Service provider's Capex
Source: Counterpoint Research
2023 Capex Share
Source: Counterpoint Research

In 2023, Microsoft and Amazon (AWS) will account for 45% of the total capex. US-based hyperscalers will contribute to 91.9% of the overall global capex in 2023.

Chinese hyperscalers are spending less due to slower growth in cloud revenues amid a weak economy and difficulties in acquiring the latest NVIDIA GPU chips for AI due to US bans. The scaled-down version – A800 of the flagship A100/H100 chips – that NVIDIA has been supplying to Chinese players may also come under the purview of the ban, further reducing access to AI silicon for Chinese hyperscalers.

Global Cloud Service Provider's AI spends as % of Total Capex, 2023
Source: Counterpoint Research

Based on Counterpoint estimates, Microsoft will spend proportionally the most on AI-related infrastructure with 13.3% of its capex directed towards AI, followed by Google at around 6.8% of its capex. Microsoft has already announced its intention to integrate AI within its existing suite of products.

AI infrastructure can be 10x-30x more expensive than traditional general-purpose data center IT infrastructure.

Though Chinese players are investing a larger portion of their spends towards AI, the amount is significantly less than that of the US counterparts due to a lower overall capex.

 The comprehensive and in-depth ‘Global Cloud Service Providers Capex’ report is available. Please contact Counterpoint Research to access the report.


Counterpoint Technology Market Research is a global research firm specializing in products in the technology, media and telecom (TMT) industry. It services major technology and financial firms with a mix of monthly reports, customized projects, and detailed analyses of the mobile and technology markets. Its key analysts are seasoned experts in the high-tech industry.

Analyst Contacts

Akshara Bassi


Peter Richardson


 Neil Shah


Follow Counterpoint Research

Related Posts

Global Server Revenues to Grow 17% YoY in 2022

  • Global server revenues will grow 17% YoY in 2022 to reach $111.7 billion.
  • Global server shipments will rise 6% YoY in 2022 to reach 13.8 million units.

New Delhi, Boston, Toronto, London, Hong Kong, Beijing, Taipei, Seoul – June 7, 2022

The global server market’s revenue will grow 17% YoY in 2022 to reach $111.7 billion, according to Counterpoint’s Global Server Sales Tracker. From an enterprise perspective, transformation to hybrid cloud and upgrades of existing infrastructure to handle increased workloads will provide growth impetus after a pause during the COVID-19 pandemic. 5G, automotive, cloud gaming and high-performance computing will remain the key drivers for cloud service providers in data center expansion.

The market is evolving with the introduction of “As-a-Service” and “pay per use” models by server companies, like Pointnext from HPE, APEX from Dell and TruScale from Lenovo. The business model pivot has been enjoying great success due to more flexibility given to the customer at minimum capital expenses due to the emergence of workload-based infrastructure.

Looking at the overall market, Research Analyst Akshara Bassi said, “In an inflationary but extremely dynamic environment, companies are looking to strengthen their infrastructure as they prepare for Web 3.0 demands from the infrastructure end. Companies are diversifying their IT infrastructure to meet the needs of data evolution and making customers cloud-ready.”

Counterpoint Research Global Server market Revenue

Global Server Market Observations, 2018-2021

  • Revenues were flat in 2020 and early 2021 owing to the pandemic and less spending by enterprises on IT infrastructure upgrades and expansion.
  • In H2 2021, there was a backlog due to supply chain disruptions across the whole value chain from accelerator chips to power ICs, which has trickled into H1 2022.
  • The demand in 2021 picked up due to investments by hyperscalers in the fields of supercomputing, edge computing and 5G deployment. From an enterprise perspective, the upgrades and expansions happened to enable digital transformation along with cloudification of processes.
  • The shipments rose at a single digit owing to better attach rates and higher configurations within the servers.
  • Higher compute and storage requirements by customers along with supply chain disruptions raised the average unit price considerably for the server units.

Counterpoint Research Global Server market Revenue by Company

Commenting on the server market, Bassi said, “Dell and HPE are the server market icons but are seeing companies like Lenovo, Inspur and Supermicro giving strong competition as demand for flexible customised configurations in bare metal option continues to rise.”

ODM Direct grew at a higher pace by 3 percentage points than the overall market for 2021 indicating shift towards ODM direct as choice of hardware for large scale Data Center deployments. Foxconn and Quanta have been gaining significant market share over the years as hyperscalers continue to expand and favor ODM Direct for their data center orders.

Key Market Drivers, 2022

  • Edge servers: Edge server configurations will be a key driver of growth in server shipments as companies start enabling chip-to-cloud features within the devices and 5G deployments make IoT use cases penetrate across all verticals – consumer, industrial, healthcare and banking.
  • Metaverse: The infrastructure spends to enable the Metaverse will give an impetus to servers as they are the building blocks of the Metaverse. Major internet corporations and hyperscalers have already detailed their plans for spends on Metaverse infrastructure.
  • Supercomputing: Many enterprises are building their own supercomputers to develop software that can unleash Web 3.0 on consumers. From a government perspective, supercomputers are becoming ubiquitous in every research department within it – from being used only in defense initially to climate and healthcare now.
  • Cloudification of services: Many services are increasingly becoming cloud-based offerings as opposed to device-based earlier, with device storage and compute needs being taken care of by cloud-based services. Data centers, and thereby servers, are an integral part of the infrastructure needed to enable these services.

Note: ODM Direct constitutes sale to hyperscalers and whitelabel device sales to Enterprises directly.


Counterpoint Technology Market Research is a global research firm specializing in products in the TMT (technology, media and telecom) industry. It services major technology and financial firms with a mix of monthly reports, customized projects and detailed analyses of the mobile and technology markets. Its key analysts are seasoned experts in the high-tech industry.

Analyst Contacts:

 Akshara Bassi


 Dale Gai


 Neil Shah


Follow Counterpoint Research


Related Posts

Nokia and Kyndryl Partner To Target Campus Wireless Market

Enterprise revenues today constitute a very small proportion of any incumbent infrastructure vendor’s total revenues. However, all vendors are looking to grow their network businesses beyond the traditional CSP market and are investing heavily in a range of new technologies and applications, with a key focus on private networks – and the campus wireless market in particular.

Campus Wireless Go-To-Market Strategy

The private networks market can essentially be split into two categories: Wide Area Networks (WAN) and Campus Wireless. While the WAN market is an established business that continues to grow, the campus wireless market is very much a nascent market. At present, it is small but growing very quickly. The WAN market typically consists of a small number of large regional or nationwide deployments such as public safety or utility-type networks. In contrast, the campus wireless market – as the name suggest – is characterised by much smaller, site-based deployments such as office buildings, hotels, shopping malls, schools, universities – and of course – factories.

Unlike public mobile networks and large private networks, the campus segment is a business where distribution channels and partnerships are critically important. This is very different to selling cellular networks to CSPs. In particular, a different approach is needed for the Industry 4.0 market, with specific consultancy, design and managed services capabilities to cater for individual verticals. As a result, vendors need to partner with a plethora of companies, including consulting companies, systems integrators, industrial technology companies, IT and cloud companies, device OEMs, etc. They will also will need to develop a host of different business models and a multitude of different distribution channels.

Nokia-Kyndryl Partnership and Ecosystem

Nokia recently announced a partnership with Kyndryl, IBM’s old global managed infrastructure services business, which was recently spun off as an independent company. Kyndryl is one of the world’s largest IT infrastructure providers involved in the design, building and management of complex, mission-critical information systems across a range of verticals. The partners plan to combine Nokia’s Digital Automation Cloud (DAC) application platform with Kyndryl’s consulting, design, implementation and managed services expertise, with the aim of supporting companies as they transition to Industry 4.0. Several Proof of Concepts applications have already been developed for US chemical company Dow Inc. which support applications such as worker safety and collaboration, asset tracking, etc. These applications are based on a blueprint which Dow hopes to expand and deploy across its sites worldwide. Exhibit 1 shows Nokia’s current ecosystem partners.

Exhibit 1:  Nokia: Private Wireless Growth and Industry 4.0 Ecosystem

Verdict: An Excellent Strategic Fit

Although private 5G can provide higher speeds and a better mobile experience with new capabilities, manufacturers may fail to see the benefits unless they talk to people with specific vertical expertise who understand their requirements – which may be very different to the requirements of a mobile operator. This is why partnerships with companies such as Kyndryl are vital for vendors such as Nokia. In fact, Counterpoint Research believes that the Nokia-Kyndryl partnership is an excellent strategic fit for both companies.

For instance, Kyndryl possess considerable scale, it has experienced teams with lots of expertise and knowledge in IoT, the edge and also AI. In addition, it mostly serves large global companies and has an enterprise customer base of around 4,000 companies. With 450 private wireless customers globally, Nokia is a leader in the private networking market and possesses a comprehensive portfolio of mission-critical network options, 5G SA core and multi-cloud solutions, vertical market applications, edge-based platforms as well as a range of industrial devices.  The partnership thus creates an entity that is global both in IT consulting and in communications networks – i.e. two complementary sectors critical for comprehensive 5G and edge implementation for global enterprises.

With the CSP market only likely to record modest growth at best over the next few years, it will interesting to see how quickly the Nokia-Kyndryl partnership starts to contribute meaningful revenues and to what extent – and how soon – Nokia’s new enterprise initiatives can redress the vendor’s 11:1 CSP/enterprise revenue imbalance.

Enterprise Wireless In Focus

The enterprise wireless market is a major growth opportunity for all incumbent vendors. As perhaps a recognition of its importance to its future growth, Ericsson this week announced the creation of a new “enterprise wireless” business unit formed from the merger of its revamped private networks unit and its Cradlepoint wireless WAN business. The unit will also include the Vonage API acquisition, once completed, as well as other businesses such as IoT platforms. However, it is inevitable that incumbents will face stiff competition across all of these markets, and from a multitude of new entrants, particularly cloud providers such as Microsoft Azure. As no player can offer end-to-end solutions across all verticals, developing the best ecosystems through partnerships will be critical for all of them, including the big cloud players.


Related Reports

Infrastructure Insights – Selected Highlights from MWC, Barcelona

Subscriber Adds and Network Roll-Out Key Challenges for Rakuten, Dish and 1&1 Drillish

Monetizing 5G Will Be The Challenge For Incumbent Vendors in 2022


Edge Computing Space Gets Crowded as Hyperscalers Move on 5G MEC Opportunity

In recent months, several hyperscale cloud providers have signed deals with major MNOs, signalling their intent to target the Multi-Access Edge Computing (MEC) space.

In December, Amazon launched its new AWS Wavelength 5G edge computing platform and announced that Verizon will be running Wavelength on its new edge computing platform, 5G Edge. This follows the partnership between AT&T and Microsoft announced earlier in 2019, which integrates Azure cloud services into AT&T’s 5G network edge. To date, Microsoft has partnered with eight managed service providers which will enable enterprises to connect to its Azure cloud, including Reliance Jio, which is building data centers for Microsoft in its home Indian market. Google and IBM are also signing similar deals with MNOs.

AWS Wavelength

AWS’s Wavelength embeds cloud compute and storage services at the edge of MNOs’ 5G networks. This enables developers to build next generation ultra-low latency applications using familiar AWS services, APIs, and tools which they can potentially use across other 5G networks around the world, thus providing a consistent developer experience. First customers using AWS Wavelength include Mapbox, Vario and games company Bethesda. AWS is also collaborating with Vodafone, SK Telecom and KDDI to launch AWS Wavelength across Europe, South Korea and Japan in 2020 with more global partners expected later in the year.

AWS’s Wavelength software will be installed on Verizon’s own computing resources located inside the MNO’s network and developers pay for access to those computing resources. Verizon claims that the partnership with Verizon is non-exclusive, which may suggest that it will offer similar cloud solutions with other hyperscalers such as Microsoft’s Azure, etc. However, Verizon has seemingly ruled out using third-party edge computing infrastructure offered by companies such as, EdgeConneX, etc.

In addition to partnering with MNOs, the hyperscalers are investing in developing their own edge infrastructure as well as leveraging the facilities of other big data center providers such as Equinix. The company recently launched a $1 billion-plus joint venture to operate data centers in Europe for Amazon and Microsoft. Other major cloud companies are also leveraging Equinix’s data centers to increase their edge presence.

To bolster this expansion, Equinix announced this month that it was acquiring US start-up Packet, which provides bare-metal “Computing-as-Service” hardware to more than 20 data center locations worldwide. The acquisition will allow it to offer dedicated “single-tenant” servers (often preferred by enterprises for regulatory and security reasons) rather than the multi-tenant servers more common in the cloud and will enable it to expand its existing data center footprint to central offices, tower base stations and C-RAN hubs.

Start-ups lining up

The AWS Wavelength deal with Verizon is not good news for the various telco-backed start-ups hoping to build an industry-based standard middleware platform, such as Deutsche Telekom-backed MobileEdgeX and Ericsson’s Edge Gravity.

Nor is it particularly good news for the host of micro data center providers eyeing opportunities with MNOs such as Vapor io, EdgeConneX, EdgeMicro, DataBank, Compass Datacenters, Baselayer, Switch, DartPoints, vXchnge and 365 Data Centers. For example, is proposing using its “Vapor Chamber” to provide edge facilities at MNOs’ C-RAN hubs (Exhibit 1). The company recently raised $90 million from backers Berkshire Partners and tower giant Crown Castle.

However, Verizon’s deal with AWS suggests that MNOs (or at least the major MNOs) will insist that MEC services should run on MNO-owned hardware located at their own edge premises. As a result, it is likely that these small players may need to re-focus on other opportunities, such as private networks, for example.

Exhibit 1: Vapor Edge Telecom Architecture

Counterpoint Vapor Edge Telecom Architecture

MNOs versus Hyperscalers

The edge is now shaping up to be a future battleground between telecom operators and the hyperscalers. Telcos believe that the fact that they own real estate at the edge (central offices, C-RAN hubs, etc.) gives them an advantage which means that the Internet companies will be forced to partner with them. However, there is mounting evidence that AWS, Microsoft et al. are moving fast and developing their own edge infrastructure. There is therefore a real risk that MNOs could lose edge business opportunities to the Internet companies if they prevaricate.

As recent events show, it seems that the big MNOs believe that they have more to benefit than to lose by partnering with the likes of Amazon, etc. Given the impact these giant Internet companies have already had on their businesses, Counterpoint believes that they should be wary of such partnerships, as the danger is that they will be relegated to the role of connectivity provider. However, there is also the chance that 5G MEC could be disruptive for the hyperscalers themselves, particularly as most enterprises do not want all their operational data going to AWS or Azure clouds. This could therefore be an opportunity for the telcos to offer an alternative, neutral, edge platform. However, the industry must move quickly and overcome challenges regarding standards, etc. if they are to seize the opportunity – a seemingly tall task at present!


IBM Watson: a milestone in artificial intelligence

On 2nd December 2015 IBM announced that it is going to deploy its cognitive computing technology platform “Watson” in India to support medical staff treating cancer. Watson is an example of a cognitive intelligence system, which, in turn, is also part of the wider development of artificial intelligence.

Cognitive Computing consists of computing systems with some level of self-learning ability and natural language processing that can be used for data analysis, pattern recognition and complex problem solving. Such systems can handle huge sets of heterogeneous data and can analyse data at greater execution speeds than conventional computing systems. The execution speed of is achieved by running multiple algorithms at the same time. The systems are capable of adaptive ‘learning’, meaning they’re well suited to responding to human needs where there are multiple possible answers; computing systems are usually poor at dealing with ambiguity. However cognitive intelligence is moving toward supporting complex decision making processes where different outcomes need to be evaluated.

IBM's Watson
IBM’s Watson

Example use cases might include an engineer designing a complex structure – for example part of a space craft. The system is able to analyse bundles of algorithms in fractions of a second. It will evaluate all possible answers and use predictive capabilities to select the best answer to the problem among the multiple possible answers. Such ability can be applied in diverse areas including healthcare, automotive, analytics and finance.

Currently IBM, Google, Apple and Microsoft are the key players in this nascent area of development. Facebook and Baidu are also active. Among these, IBM has launched and deployed Watson in healthcare and analytics. Google and Microsoft are still largely in the R&D phase with, respectively, Brain and Adam. Both these systems are working on image recognition problems that can have application in, for example, examining the nutritional content of a meal simply by taking an image of it, or performing diagnostics based on photographing skin lesions. By contrast, Apple is conspicuous by its silence. So far all that has been seen is the less than impressive Siri personal assistant – which is some distance behind Microsoft’s Cortana and Google’s Now.

The global cognitive computing and artificial intelligence space remains in a largely nascent stage. 2016 will see more real world applications of systems like IBM’s Watson. Google, Microsoft and other AI pioneers will apply machine learning and start to offer services – for example through a cloud-based AI system. IBM has developed an early lead for its particular flavour of AI. However IBM’s approach with Watson is only one way to deliver benefits of AI.

The real mystery remains Apple’s activity in this area. Ever secretive, Apple has published almost zero research into its AI development work, however it is building a substantial team and has huge resources to deploy in this area. Much of the development work currently being undertaken in AI is achieved through collaboration. Apple’s lack of collaboration may therefore mean its as behind the curve as Siri is.

By- Nitish Pande

Term of Use and Privacy Policy

Counterpoint Technology Market Research Limited


In order to access Counterpoint Technology Market Research Limited (Company or We hereafter) Web sites, you may be asked to complete a registration form. You are required to provide contact information which is used to enhance the user experience and determine whether you are a paid subscriber or not.
Personal Information When you register on we ask you for personal information. We use this information to provide you with the best advice and highest-quality service as well as with offers that we think are relevant to you. We may also contact you regarding a Web site problem or other customer service-related issues. We do not sell, share or rent personal information about you collected on Company Web sites.

How to unsubscribe and Termination

You may request to terminate your account or unsubscribe to any email subscriptions or mailing lists at any time. In accessing and using this Website, User agrees to comply with all applicable laws and agrees not to take any action that would compromise the security or viability of this Website. The Company may terminate User’s access to this Website at any time for any reason. The terms hereunder regarding Accuracy of Information and Third Party Rights shall survive termination.

Website Content and Copyright

This Website is the property of Counterpoint and is protected by international copyright law and conventions. We grant users the right to access and use the Website, so long as such use is for internal information purposes, and User does not alter, copy, disseminate, redistribute or republish any content or feature of this Website. User acknowledges that access to and use of this Website is subject to these TERMS OF USE and any expanded access or use must be approved in writing by the Company.
– Passwords are for user’s individual use
– Passwords may not be shared with others
– Users may not store documents in shared folders.
– Users may not redistribute documents to non-users unless otherwise stated in their contract terms.

Changes or Updates to the Website

The Company reserves the right to change, update or discontinue any aspect of this Website at any time without notice. Your continued use of the Website after any such change constitutes your agreement to these TERMS OF USE, as modified.
Accuracy of Information: While the information contained on this Website has been obtained from sources believed to be reliable, We disclaims all warranties as to the accuracy, completeness or adequacy of such information. User assumes sole responsibility for the use it makes of this Website to achieve his/her intended results.

Third Party Links: This Website may contain links to other third party websites, which are provided as additional resources for the convenience of Users. We do not endorse, sponsor or accept any responsibility for these third party websites, User agrees to direct any concerns relating to these third party websites to the relevant website administrator.

Cookies and Tracking

We may monitor how you use our Web sites. It is used solely for purposes of enabling us to provide you with a personalized Web site experience.
This data may also be used in the aggregate, to identify appropriate product offerings and subscription plans.
Cookies may be set in order to identify you and determine your access privileges. Cookies are simply identifiers. You have the ability to delete cookie files from your hard disk drive.