Following four consecutive years of modest telecom equipment growth across the six telecom programs tracked at the Dell’Oro Group*, preliminary findings show that the aggregate telecom equipment market moderated somewhat from the 8% revenue increase in 2021 to 3% year-over-year (Y/Y) in 2022.
Looking back at the full year, the results were slightly lower than the 4% growth rate we projected a year ago going into 2022. In addition to more challenging comparisons in the advanced 5G markets and the supplier exits in Russia, the strengthening USD weighed on the broader telecom equipment market. Supply issues also impacted the market negatively during 1H22 but eased somewhat in the second half.
Regional developments were mixed, underpinned by strong growth in North America and CALA, which was enough to offset more challenging conditions in EMEA and the Asia Pacific. With China growing around 4%, we estimate global telecom equipment revenues excluding China increase around 3% in 2022.
From a technology perspective, there is a bit of capex shift now underway between wireless and wireline. Multiple indicators suggest Broadband Access revenues surged in 2022, however, this double-digit growth was offset by stable or low-single-digit growth across the other five segments (Microwave Transport, Mobile Core Network, Optical Transport, RAN, SP Router & Switch).
Vendor dynamics were relatively stable between 2021 and 2022, with the top 7 suppliers driving around 80% of the overall market. Despite on-going efforts by the US government to limit Huawei’s TAM and access to the latest silicon, our assessment is that Huawei still leads the global telecom equipment market, in part because Huawei remains the #1 supplier in five out of the six telecom segments we track. At the same time, Huawei has lost some ground outside of China. Still, Nokia, Ericsson, and Huawei were the top 3 suppliers outside of China in 2022, accounting for around 20%, 18%, and 18% of the market, respectively.
Following five consecutive years of growth, the Dell’Oro Group believes there is more room left in the tank. Collectively the analyst team is forecasting the overall telecom equipment market to increase 1% in 2023 and record a sixth consecutive year of growth. Risks are broadly balanced and the analysts will continue to monitor the 5G rollouts in India, capex cuts in the US, and 5G slowdown in China (preliminary data by MIIT suggest new 5G BTS volumes will drop by a third in 2023 relative to 2022), wireless and broadband investments in Europe, forex fluctuations, and inventory optimization.
*Telecommunications Infrastructure programs covered at Dell’Oro Group, include Broadband Access, Microwave & Optical Transport, Mobile Core Network (MCN), Radio Access Network (RAN), and SP Router & Switch.
2023 witnessed a remarkable resurgence of the OFC conference following the pandemic. The event drew a significant turnout, and the atmosphere was buzzing with enthusiasm and energy. The level of excitement was matched by the abundance of groundbreaking announcements and product launches. Given my particular interest in the data center switch market, I will center my observations in this blog on the most pertinent highlights regarding data center networking.
The Bandwidth and Scale of AI Clusters Will Skyrocket Over the Next Few Years
It’s always interesting to hear from different vendors about their expectations for AI networks, but it’s particularly fascinating when Cloud Service Providers (SPs) discuss their plans and predictions regarding the projected growth of their AI workloads. This is because such workloads are expected to exert significant pressure on the bandwidth and scale of Cloud SPs’ networks, making the topic all the more astounding. At OFC this year, Meta portrayed their expectations of how their AI clusters in 2025 and beyond may look like. Two key takeaways from Meta’s predictions:
The size and network bandwidth of AI clusters are expected to increase drastically in the future: Meta expects the size of its AI cluster will grow from 256 accelerators today to 4 K accelerators per cluster by 2025. Additionally, the amount of network bandwidth per accelerator is expected to grow from 200 Gbps to more than 1 Tbps, a phenomenal increase in just about three years. In summary, not only the size of the cluster is growing, but also the amount of compute network per accelerator is skyrocketing.
The expected growth in the size of AI clusters and compute network capacity will have significant implications on how accelerators are currently connected: Meta showcased the current and potential future state of the cluster fabric. The chart below presented by Meta proposes flattening the network by embedding optics directly in every accelerator in the rack, rather than through a network switch. This tremendous increase in the number of optics, combined with the increase in network speeds is exacerbating the power consumption issues that Cloud SPs have already been battling with. We also believe that AI networks may require a different class of network switches purpose-built and designed for AI workloads.
Pluggable Optics vs. Co-packaged Optics (CPOs) vs. Linear Drive Pluggable Optics (LPOs)
Pluggable optics will be responsible for an increasing portion of the power consumption at a system level (more than 50% of the switch system power @ 51 .2 Tbps and beyond) and as mentioned above, this issue will only get exacerbated as Cloud SPs build their next-generation AI networks. CPOs emerged as an alternative technology that have the promise to reduce power and cost compared to pluggable optics. Below are some updates about the state of the CPO market:
Cloud SPs are still on track to experiment with CPOs: Despite rumors that Cloud SPs are canceling their plans to deploy CPOs due to budget cuts, it appears that they are still on track to experiment with this technology. At OFC 2023, Meta reiterated their plans to consider CPOs in order to reduce power consumption from 20 pJ/bit to less than 5 pJ/bit using Direct Drive CPOs (Direct Drive CPOs eliminate the digital signal processors (DSPs)). It is still unclear, however, where exactly in the network they plan to implement CPOs or if it will be primarily used for compute interconnect.
The ecosystem is making progress in developing CPOs but a lot remains to be done: There were several exciting demonstrations and product announcements at OFC 2023. For example, Broadcom showcased a prototype of its Tomahawk 5-based 51.2 Tbps “Bailly” CPO system, along with a fully functional Tomahawk 4-based 25.6 Tbps “Humboldt” CPO system that was announced in September 2022. Additionally, Cisco presented the power savings achieved with its CPO switch populated with CPO silicon photonic-based optical tiles driving 64×400 G FR4, as compared to a conventional 32-port 2x400G 1 RU switch. During our discussions with the OIF, we were provided with an update on the various standardization efforts taking place, including the standardization of the socket that the CPO module will go into. Our conversations with major players and stakeholders made it clear that significant progress has been made in the right direction. However, there is still much work to be done to reach the final destination, particularly in addressing serviceability, manufacturability, and testability issues that remain unsolved. Our CPO forecast published in our 5-year Data Center Forecast report January 2023 edition takes into consideration all of these challenges.
LPOs present another alternative to explore: Andy Bechtolsheim of Arista has suggested LPOs as another alternative that may address some of the challenges of CPOs. The idea behind LPOs is to remove the DSP from pluggable optics, as the DSP drives about half of the power consumption and a large portion of the cost of 400 Gbps pluggable optics. By removing the DSP, LPOs would be able to reduce optic power by 50% and system power by up to 25% as Andy portrayed in the chart below.
Additionally, other materials for Electric Optic Modulation (EOM) are being explored, which may offer even greater savings compared to silicon photonics. Although silicon photonics is a proven high-volume technology, it has high voltage and insertion loss, so exploring new materials such as TFLN may help lower power consumption. However, we would like to note that while LPOs has the potential to achieve power savings similar to CPOs, they put more stress on the electrical part of the switch system and require a high-performance switch SERDES and careful motherboard signal integrity design. We expect 2023 to be busy with measurement and testing activities for LPO products.
800 Gbps Pluggable Optics are Ready for Production Volume and 1.6 Tbps Optics are already in the Making
While we are excited about the aforementioned futuristic technologies that may take a few more years to mature, we are equally thrilled about the products on display at the OFC that will contribute to the market growth in the near future, such as the 800 Gbps optical pluggable transceivers, which were widely represented at the event this year. The timing was perfect, as it is aligned with the availability of 51.2 Tbps chips from various vendors, including Broadcom and Marvell. While 800 Gbps optics started shipping in 2022, more suppliers are currently sampling, and the volume production is expected to ramp up by the end of this year, as indicated in the chart below from our 5-year Data Center Forecast report January 2023 edition. In addition, several 1.6 Tbps optical components and transceivers based on 200 G per lambda were also introduced at OFC 2023, but we do not expect to see substantial volumes in the market before 2025/2026.
OFC 2023 was a great conference this year with the exhibit hall packed with people exchanging thoughts and ideas. One of the main technology themes at the show was around terabit-capable coherent wavelengths. Hence, my main takeaway from the conference was that we are now in the Terabit Era.
During the conference, practically every component and systems manufacturer showed off their plans for a 1.2 Tbps-capable coherent wavelength. The only exception was Ciena with an advanced 1.6 Tbps single carrier solution.
New High-Performance Coherent DSPs
The next generation of coherent DSPs was announced and displayed by several companies. Based on conversations at the show, the first significant quantity of shipments could begin as early as 2H 2023 with a ramp in production during 2024. The following is a list of announcements related to high-performance sixth-generation coherent DSPs.
Cisco/Acacia Jannu: 1.2 Tbps-capable DSP on a 5 nm node process operating at up to 136 Gbaud. General availability target is mid-2023.
NEL ExaSpeed GAIA: 1.2 Tbps-capable DSP on a 5 nm node process operating at up to 140 Gbaud. General availability target is mid-2023.
Nokia PSE 6s: 1.2 Tbps-capable DSP on a 5 nm node process operating at up to 130+ Gbaud. General availability target is end-2023.
Infinera ICE7: Customized 1.2 Tbps-capable DSP from a partner supplier on a 5 nm node process operating at up to 148 Gbaud. General availability target is mid-2024.
Ciena WaveLogic 6e: 1.6 Tbps-capable DSP on a 3 nm node process operating at up to 200 Gbaud. General availability target is mid-2024.
On a side note, Huawei was not at this conference. However, in separate meetings with the company, we understand they have plans to develop a 1.2 Tbps-capable DSP operating at up to 130+ Gbaud. The semiconductor process technology is unknown at this time, but it may be 7 nm if SMIC can produce it.
Among the vendors, Ciena is the only one launching a coherent DSP that can do a single carrier 1.6 Tbps wavelength. To accomplish this, the company is not only going with a more advanced semiconductor node process (3 nm rather than 5 nm), but is also developing advanced coherent drivers and receivers that can push the transponder to 200 Gbaud (no other company has 200 Gbaud drivers at this time). To remove any concerns about timing and progress of the WaveLogic 6e development, especially since it is using such an advanced semiconductor process node, Ciena took the time to detail out its development progress behind closed curtains.
Terabit Era Ready
There were a number of announcements by service providers confirming that 1.2 Tbps-capable devices are ready, and that they themselves were looking forward to these new high-performance transponders. Three of those announcements were as follows:
Windstream did a field trial in a live network to confirm technology readiness in a real world setting. The field trial used the Acacia CIM8 based on the Jannu DSP and showed that a 1 Tbps wavelength could be carried over at least 541 km with excess performance margin. The wavelength operated over SMF28 fiber with approximately 6 WSS filters along the way.
GlobalConnect did a field trial of Nokia’s PSE 6s in a live network. In this trial, GlobalConnect tested the 1.2 Tbps-capable DSP for a long haul route, operating at a wavelength speed of 800 Gbps. The route length achieved in this trial was 2,019 km.
China Mobile did a field trial of the Acacia CIM8 in an ultra-long-haul application, demonstrating that a 1.2 Tbps-capable transponder can transmit a 400 Gbps wavelength 5,616 km in their live all-optical network.
These field trials are a great demonstration of how the coherent 1.2+ Tbps-capable technology delivers on both higher capacity and longer spans. In the January 2023 Optical Transport five-year forecast report, we projected that demand for 400+ Gbps wavelength shipments will grow at a 40% CAGR, and that one of the key enablers for this growth will be these new sixth-generation coherent DSPs.
We just returned from MWC after a couple of intense days of meetings. RAN might not be the fastest growing market but from a technology perspective, the pace of change simply remains impressive and the MWC is a great event to witness the progress. Although we are not the right source to capture all of the PR activity involving incremental product announcements, we want to point out a few RAN-related observations that could potentially impact the RAN forecast or vendor dynamics.
Virtualized RAN is gaining momentum
As we now know, vRAN started out slow but picked up some speed in 2022 in conjunction with the progress in the US. The challenge from a forecasting perspective is that the visibility beyond the greenfields and the early brownfield adopters is limited, primarily because purpose-built RAN still delivers the best performance and TCO. As a result, there is some skepticism across the industry about the broader vRAN growth prospects.
During MWC, we learned four things: 1) Near-term vRAN visibility is improving – operators in South Korea, Japan, US, and Europe are planning to deploy vRAN in the next year or two. 2) vRAN performance is firming up. According to Qualcomm, Vodafone (and Qualcomm) believes the energy efficiency and performance gap between the traditional and new Open vRAN players is shrinking (Vodafone publicly also praised Mavenir’s OpenBeam Massive MIMO AAU). Samsung also confirmed (again) that Verizon is not giving up any performance with Samsung’s vRAN relative to its purpose-built RAN. 3) vRAN ecosystem is expanding. In addition to existing vRAN suppliers such as Samsung, Ericsson, Mavenir, Rakuten Symphony, and Nokia announcing improvements to their existing vRAN/Cloud RAN portfolios, more RAN players are jumping on the vRAN train (both NEC and Fujitsu are expecting vRAN revs to ramp in 2023). And perhaps more interestingly, a large non-RAN telecom vendor informed us they plan to enter the vRAN market over the next year. 4) The RAN players are also moving beyond their home turf. During the show, Nokia announced it is entering the RAN accelerator card segment with its Nokia Cloud RAN SmartNIC (this is part of Nokia’s broader anyRAN strategy).
256T256R Massive MIMO
For a technology that was initially viewed as being mostly fit for high-traffic locations, Massive MIMO has come a long way in just a few years, ramping at a much faster pace than initially expected. As we recently outlined in the Massive MIMO 2022 blog, the days of exponential growth are in the past but there is still Massive MIMO upside ahead to support TDD MBB expansions in the less advanced markets, FDD hotspot deployments, FWA, and TDD product refresh to take advantage of continuous product improvements such as wider IBW, lower power consumption, smaller form factors, and more transceivers/antenna elements.
With limited new sub-7 GHz spectrum on the horizon, improving the spectral efficiency using more didoes and transceivers will continue to play a fundamental role with 5G, 5G-Advanced, and 6G. This is not a surprise. However, since we previously assumed the 128T128R would be the de-facto configuration for the 6 to 7 GHz band, it was somewhat surprising to see Huawei’s 256TRx prototype in the booth (ZTE had a 128T128R AAU in its booth).
Private 5G ecosystem is evolving
Private 5G is running behind schedule. We recently adjusted the private wireless forecast downward to reflect the current state of the market. Still, the slow uptake is not dampening the enthusiasm for private wireless. If anything, the interest is growing and the ecosystem is evolving as suppliers with different backgrounds (RAN, core, Wi-Fi, hyperscaler, in-building, SI) are trying to solve the enterprise puzzle. Below is a summary of the private RAN, core, and SI/services providers that we are currently monitoring.
More suppliers want a piece of Europe
Per our 4Q22 RAN report, Ericsson, Nokia, and Huawei collectively accounted for around 95% of the European RAN market in 2022. Other RAN suppliers have tried to expand their respective footprints over the past couple of years without much success. One of the takeaways from the MWC discussions is that activity with the smaller suppliers is on the rise:
Samsung highlighted its vRAN/Open RAN progress with Vodafone in the UK, Germany, and Spain and remains optimistic its European RAN revenues will soon become more material.
NEC started recognizing European RAN revenues in 2022 (primarily driven by 1&1 in Germany).
Fujitsu believes its European RAN growth prospects will improve in 2024.
Mavenir has 10 K brownfield Open RAN sites in the pipeline for 2023/2024 – Deutsche Telekom is part of the mix, however, the exact upside in Europe remains unclear.
Rakuten Symphony recently estimated its Open RAN pipeline is worth around $2B (half of the total pipeline). We don’t know the size of the European component. However, the company recently reported that nearly half of the # (not value) of prospecting clients (for RAN and non-RAN) are located in Europe per the chart below.
While it is unlikely that management over at Ericsson, Nokia, and Huawei are losing a ton of sleep at this juncture from these announcements, it might be worthwhile to check back in a year if the collective share of the top 3 is still around 95%.
Skepticism is on the rise
Not surprisingly, disconnects between vision and reality are common when new technologies are introduced. Even if this is expected, we are sensing more frustration across the board this time around, in part because RAN growth is slowing and 5G still has mostly only delivered on one out of the three usage scenarios outlined in the original 5G use case triangle. With 5G-Advanced/5.5G and 6G starting to absorb more oxygen, people are asking if mMTC+/mMTC++ and URLLC+/URLLC++ are really needed given the status of basic mMTC and URLLC. Taking into consideration the vastly different technology life cycles for humans and machines, there are more questions now about this logic of assuming they are the same and will move in tandem. If it is indeed preferred to under-promise and over-deliver, there might be some room to calibrate the expectations with 5G-Advanced/5.5G and 6G.
And most importantly, this was our first live MWC event since 2019. It was great to be back and meet people in person without any Covid restrictions. The energy level at the show was amazing – now if this energy could somehow start showing in the RAN numbers, that would be even more exciting.