Telecom Capex – Past, Present and the Future
Together with the team from Strand Consult, I have written the report The Nature of Telecom Capex until the End of the Decade, comprehensively analyzes capital expenditure (Capex) trends in the Western European telecommunications industry, examining the forces driving investments, emerging challenges, and projections for the coming years. This report builds upon last yearโs report on โThe Nature of Telecom Capex,โ which can be found on my Technology-Economics blog, techneconomyblog.com.
The purpose is to describe what drives the investments in the industry and how we believe these will develop in the future. The report should be seen as inspiration that can be used when discussing the trends that are affecting the industry.
Recent developments suggest the Industry may be entering a new phase marked by what some analysts have labeled an “End to History of Telecom” โ an era where the apparent exponential growth, high-intensity Capex spending, and unbounded data traffic growth no longer define the sector’s trajectory. Some evidence in the 2023 spending compared to previous years may indicate that some European telcos have begun scaling back on Capex, primarily due to an overbuilt and underutilized infrastructure landscape, high capital costs, a deceleration in data growth, and the expected cyclic spending dynamics that usually is observed after new technology has been deployed (e.g., 5G, Fiber, …).
The recent (minor) decline in Telco Capex between 2022 and 2023 has led some industry followers to announce that this is a sign of a persistent future decline in Capex and an End to Big spending in our Industry.
These arguments are mainly driven by:
๐๐ป๐ฑ ๐๐ผ ๐๐ฟ๐ผ๐๐๐ต of data traffic. There is no doubt, and there should be no surprise either, that the data traffic growth rate in Telecom Networks is declining irrespective of whether it is fixed broadband or mobile broadband data traffic we observe. We are past the inflection point in Europe and will be below 10%, heading towards 0% within the next 5 to 10 years (if the current trend persists). This is a good thing as it enables our telco networks to more easily support the traffic for the foreseeable future without capacity Capex spinning out of control, which it would have done if traffic growth followed a sustainable exponential growth path. However, Capex will still be needed for capacity but at a manageable rate.
๐๐ป๐ฑ ๐๐ผ ๐ฆ๐ฝ๐ฒ๐ฒ๐ฑ of data services. It has been argued, rightly so, in my opinion, that the actual demand for throughput for the services we use has plateaued, and striving for ever higher throughputs is unnecessary. As there is a duality between speed (bps) & volumetric capacity (Bytes), increasing the effective speed of our Telco networks also amounts to providing more capacity or allowing more users to experience excellent quality. While, on average, it is true that the capacity of Telco Networks far exceeds the demand, it is not universal when de-averaging demand vs. capacity on a cell level (yet!). More innovative solutions are needed (e.g., Opanga Networks RAIN solution) rather than adding more capacity, spending unnecessary mobile Capex better spend elsewhere (e.g., Fiber deployment).
๐๐ป๐ฑ ๐๐ผ ๐ฆ๐๐ฝ๐ฝ๐น๐ถ๐ฒ๐ฟ๐ of telco infrastructure. When Analysts’ wishful thinking about dramatic reductions in Capex gets into gear, they often forget that there is a huge debt to be paid remaining in many Telco networks in Europe. Some of the biggest European countries still have to remove their Chinese infrastructure from their Mobile Networks (although they have fought to delay). So far, The European Commission has focused on 5G via their 5G Toolkit for Security and raised concerns about several member states’ reluctance to comply with the recommended removal of high-risk supplier infrastructure. Some leading member states (e.g., DK) have legislation in place that is technology agnostic, extending the high-risk supplier ban to fixed broadband as well (and beyond).
๐๐ป๐ฑ ๐๐ผ ๐๐ถ๐๐๐ผ๐ฟ๐ of Telecom. Whether we are seeing the beginning of the end of telecom history, as William Webb “hints” at, is an interesting question. I expect some big fundamental questions to remain before we can call it a day and sleep on our “laurels.” A call for resetting how we think of 6G is one that I support. See, for example, Dean Bubley’s โThe 6G Vision Needs a Resetโ for additional inspiration for the โ6G Resetโ thinking. Hopefully, we will see a shift in how we look (design, plan, build & operate) our networks instead of today’s siloed approach, moving towards an integrated approach.
Several overlapping trends have shaped telecom Capex in Western Europe. Between 2017 and 2023, Capex levels rose significantly to support the rollout of 5G networks and fiber infrastructure, reaching a peak of โฌ51 billion in 2022. However, Capex spending dropped in 2023 as operators began questioning the sustainability of high expenditure levels amid slow data growth, challenging financial conditions, and limited interest in the next generation of connectivity, namely 6G. This decline has led some analysts to forecast an “End to High Capex Intensity,” suggesting that telecom companies are unlikely to return to their previous levels of investment. This shift is partly influenced by economic factors like inflation and rising interest rates, which have increased the cost of capital and made large-scale projects more costly and riskier. Meanwhile, demand for high-speed connectivity appears to stabilize, raising questions about future network investments. This slowdown has sparked debates around the “End to Growth of Data Traffic” as industry observers question whether data consumption can continue growing in an era where most of the population and their โthingsโ already have access to high-speed internet. What is often forgotten in the discussion of slowing growth is that on the path to near-zero growth, which may easily take 5+ years after the inflection point has been passed, there is still a lot of data growth within the future growth path, and thus, investments into capacity will not just magically disappear.
Even though global network traffic growth show signs of deceleration, emerging data-intensive services and applications like artificial intelligence (AI), augmented reality (AR), and virtual reality (VR), may drive additional traffic demand significantly. According to Nokiaโs projections, applications powered by AI, along with the growing adoption of immersive technologies, are expected to add substantial data flow across both consumer and enterprise networks. For instance, consumer AI traffic, encompassing applications such as generative AI, immersive media, and cloud-based XR experiences, is forecasted to grow at an impressive compound annual growth rate (CAGR) of 51%, while enterprise AI traffic is projected to increase at a CAGR of 57%, illustrating the possible intensifying role of these technologies in driving network demand. The overall effect, if such contributions turns out to be true, may result in an overall data growth rate that appears constant for the foreseeable future (e.g., 5 โ 10 years).
Another major factor in the recent Capex reduction is the role of market consolidation. Across Europe, a wave of telecom mergers and acquisitions has led to significant structural changes in the Industry. Notable examples include Vodafone UK and Three UK, which announced their intent to merge in 2023, along with similar moves in Spain, Italy, Greece, and Denmark. In the short term, consolidation often reduces Capex spending as companies delay investments during merger approvals and initial integration phases. However, once the mergers are finalized, Capex typically rises again as the new entities consolidate infrastructure, update systems, and integrate operations to realize the benefits of their union. Such integration efforts require substantial investment to achieve operational synergies, enhance network performance, and secure a competitive edge, offsetting the initial reduction in Capex. As consolidation activity is expected to continue in the region, the associated Capex trends will likely reflect these temporary dips followed by gradual increases.
Security considerations have also introduced new challenges (and opportunities) for Capex planning. In recent years, European telecommunications companies have faced mounting pressure to replace high-risk suppliers from non-EU countries, notably Huawei and ZTE. This shift is primarily driven by concerns over data privacy, national security, and the need for a more resilient supply chain. The European Union has, with their โThe EU Toolbox for 5G Securityโ, actively pushed to remove Chinese-origin infrastructure from critical areas, especially 5G networks. In response, telecom operators in countries like Germany have claimed that the high-risk regulation may led to significant Capex increases as they allegedly work towards replacing Chinese equipment in their networks while continuously resisting to purge their networks of those high-risk suppliers. The European Commission has established guidelines to aid telecom operators in this transition, urging member states to define short transition periods to phase out high-risk suppliers, thus promoting a more secure telecom environment free of high-risk suppliers. Given the heavy reliance on Chinese technology in many of Europeโs largest telecom networks, these replacement efforts could lead to substantial, albeit temporary, increases in Capex for affected operators above and beyond what may be embedded in the cyclic renewal process of their infrastructures. The โbestโ investment case for telcos is that local regulators and politicians allow their telecom companies to replace their high-risk infrastructure when it is financially written off or should be replaced anyway due to obsolescence. The โworstโ scenario for some laggard telcos that have successfully resisted regulatory pressures to replace high-risk infrastructure is that they will experience regulation that forces by law replacement before a high-risk infrastructure’s economic or operational end-of-life has been reached. I have experienced analysts who believe the replacement Capex is similar to the regular replacement Capex of obsolete equipment, ignoring the substantial back-end support systems and processes that also need to change. Moreover, maybe they also forget that several big European telcos have a very significant share of high-risk suppliers installed that they are continuously investing into (even today!) and, by that, incurring substantial capital replacement debt that may have to be paid back over a much shorter period than a regular obsolescence transformation that more often than not would involve the same supplier.
The re-election of Donald Trump is expected to intensify pressure on China and its technology suppliers, notably Huawei. The US policy is likely to further restrict Chinese manufacturers’ access to American technology, complicating operations for firms like Huawei and the telecom businesses relying on its technology solutions. This aligns with a broader strategy of “derisking”โreducing dependency on Chinese technologyโanticipated to become a central policy for the US and its allies. Consequently, global supply chains may experience significant disruptions, necessitating urgent and time-pressured adjustments by companies worldwide to adapt to the evolving trade landscape. This may also pose substantial operational and financial risks for telecom businesses in Europe that have been highly resistive to adapting the European Commissionโs recommendations towards high-risk suppliers.
This all said, there is no clear evidence that operators that have โripped and replacedโ their Chinese telco infrastructure (e.g., Core for most and RAN for some) have incurred substantial increases in their Capex spending as they transitioned to 5G. Although given the financial boundaries and guardrails that telecom operates under, it may also not have been expected to see such increases directly. The reason most likely has been that โEU 5G Security Toolboxโ compliant non-Chinese suppliers, like Ericsson, Nokia, and Samsung, have provided substantial rebates softening the replacement and, at the same time, buying 5G (and 4G) market share, taking Huawei out of the equation. Our report also discusses the relatively high likelihood of the high-risk supplier ban being extended to fixed broadband network infrastructure.
A core theme of the report is the transformation of data growth dynamics, which previously fueled continuous Capex expansions. Until recently, data demand followed what appeared to be an exponential growth trajectory, necessitating frequent network upgrades and capacity expansions. This also drove the need for next-generation access technology to leapfrog spectral efficiency and throughput (e.g., 4G, 5G, XGPON, …). However, current data growth now appears to be approaching what economists call an “S-curve” โ a self-limiting growth pattern where demand eventually stabilizes.
๐๐ ๐ฝ๐ผ๐ป๐ฒ๐ป๐๐ถ๐ฎ๐น growth dynamics have the same (percentage) growth rate indefinitely. It is rare and unsustainable in natural phenomena. While exponential growth can occur under ideal conditions, such as abundant resources and minimal constraints, these conditions rarely persist in nature for long.
๐ฆ๐ฒ๐น๐ณ-๐น๐ถ๐บ๐ถ๐๐ถ๐ป๐ด growth dynamics, or S-curve behavior, will have a declining growth rate. Natural systems are generally self-limiting, although they might exhibit exponential growth over a short term, typically in the initial growth phase.
By 2023, Western Europe’s data demand reached 11 Exabytes (EB) per month, which, though substantial, remains significantly below the network’s total capacity. For instance, the existing mobile infrastructure could theoretically support 130 EB monthly, almost 12 times the demand. This surplus capacity, or “overproduction,” implies that operators have invested more than current demand necessitates. As a result, Capex is expected to shift away from expansion and towards optimization and efficiency in the future, emphasizing more intelligent network management over constant expansion. This change aligns with the “End to Growth of Data Traffic,” suggesting that data consumption will not continue on its previous upward trajectory indefinitely. Operators are increasingly focused on optimizing the existing network rather than expanding capacity, especially as the high-speed broadband needs of most consumers are already met.
Our report further explores how Capex is transitioning from traditional hardware investments to software-centric solutions. Industry shifts in accounting practices, like adopting IFRS16, have led to reclassifying leasing expenses as Capex, thus impacting reported financial metrics. Moreover, cell tower spin-offs, cloud migrations, and adopting software-defined networking (SDN) are reshaping the Capex landscape. Many telecom operators now lease cell towers and data centers instead of owning them, which reduces direct Capex while increasing operational expenses (Opex). For example, telecoms have moved from investing heavily in data centers to using cloud-based services, which allows them to operate more flexibly and reduce upfront capital requirements. Software-defined networks enable operators to replace expensive proprietary hardware with software solutions, reducing the need for continuous Capex on physical assets like routers and switches. This shift reflects a broader industry trend where telecoms increasingly rely on operational expenditure to meet infrastructure needs.
Our report also emphasizes Capex’s broader financial implications for telecom companies’ income statements, balance sheets, and cash flow statements. Capex investments enhance a company’s infrastructure and capacity but also increase depreciation expenses, which can reduce net income in subsequent years. In my experience, 60% to 80% of a Capex plan is almost mathematically determined. Even if mathematical, it will never be assumption-free or super objective as we look into the crystal ball of the future (typically 3 to 5 years), where matters of opinion may differ a lot (e.g., exponential vs. self-limiting growth, fixed vs. mobile broadband, terrestrial vs. LEO satellite vs. stratospheric drone balance, when 6G, etc.). The report examines some budgetary human and organizational behaviors that drive a portion of Capex that is less mathematical and more driven by personal and organizational needs. These behaviors are often ignored when analysts assess telcos’ future capital plans.
Many European operators are now adopting new strategies, such as dividend commitments and debt management, to balance the need for Capex with their overall financial health. By committing to stable dividend payouts, telecoms must carefully allocate resources to ensure that Capex does not overly strain their financial position or limit shareholder returns. This often results in prioritizing critical projects while delaying less urgent ones, especially when interest rates and capital costs are high. Some operators have opted not to issue dividends but instead channel profits into Capex to pursue long-term growth. This approach allows telecoms to reinvest earnings into network upgrades, innovation, and expansion, which may yield greater returns over time than immediate cash distributions to shareholders.
We thoroughly examine telecom Capex trends in Western Europe, highlighting key themes such as the “End to History of Telecom” and the gradual shift away from high-intensity Capex and exponential data growth. As the Industry matures, telecom companies are likely to adopt more strategic approaches to investment, focusing on network efficiency, security, and cost-effective operations. The future of telecom Capex will thus be characterized by targeted investments aimed at maintaining and optimizing existing networks rather than pursuing perpetual growth. While the Industry may never entirely abandon Capex-intensive projects, the emphasis should shift towards sustainable, carefully managed investments that align with market demands and financial realities.
It is doubtful that Capex will substantially decline towards the end of the decade. We see increased urgent regulatory pressure to remove high-risk Chinese infrastructure from European mobile and fixed broadband networks. Moreover, while the intensity of fiber optical network rollout reduces across the member states of the EU, laggard countries such as Germany and the UK will continue to commit substantial capital spend to close the fixed broadband gap to many of their peer countries towards the end of the decade. Furthermore, the regular obsolescence transformations of the radio access infrastructure will peak between 2025 and 2030. One would expect at least one to two replacement cycles of a radio access technology (e.g., 4G, 5G NSA, 5G SA, โฆ) before the next one is introduced. Some Telcos may decide to wait for 6G-compliant infrastructure, pushing investments toward the decade’s end. However, I would expect that most would have one 5G modernization cycle and attempt to postpone the 6G introduction to the 30s rather than be the first movers here. Expecting substantial reductions of absolute capital investment spend and Capex intensity (i.e., Capex to Revenue ratios) over the remainder of the decade seems at the moment like engaging in too much wishful thinking.
