During the first half of 2024, e& United Arab Emirates (U.A.E) set a new benchmark for 5G speeds, outperforming all other operators worldwide. This achievement directly results from e&’s continued investment in its mobile infrastructure and the deployment of 5G Standalone (SA), ensuring it stays ahead of international competitors. In this article, we examine the network performance of the world’s leading 5G operators, analyze the factors contributing to their speeds, and discuss the importance of 5G speed competition as the industry prepares for 6G.

Key Takeaways:

• The U.A.E. is the fastest 5G market. According to Ookla® Speedtest Intelligence® data, The U.A.E. led the world in 5G speeds during Q1-Q2 2024, with a median download speed of 660.08 Mbps.
• e& UAE is the top-performing 5G operator in the UAE and globally: e& UAE recorded the highest median 5G  download speed globally at 749.63 Mbps and an upload speed of 43.52 Mbps in Q1-Q2 2024. This performance has earned e& UAE the highest Speed Score®, showcasing continuous and steady improvements since 2022.
• Competing on high-performing, robust networks can help operators better prepare for 6G. Mobile operators that focus on delivering fast and reliable service now not only gain a competitive edge but also position themselves as leaders in industry innovations like 6G, helping them capitalize on the growth in next-generation digital services.

The U.A.E. leads in 5G performance globally

According to Ookla Speedtest Intelligence data, the U.A.E. was the fastest 5G market globally in Q1-Q2 2024, with a median download speed of 660.08 Mbps. Building on recent improvements in 5G performance and continued fiber deployments, the U.A.E. secured a leading position in July and August 2024 on the Speedtest Global Index™ for both mobile and fixed broadband services across all technologies — a first since the index’s launch in 2017.

Government support and operators’ investments were vital to achieving this result. For example, the U.A.E’s telecom regulator TDRA waived the fee for 5G frequencies of 3GHz and above for five years, and operators had access to large amounts of spectrum in the 2500-2600 MHz and 3300 MHz-3800 MHz ranges. Furthermore, as of March 2024, local operators e& UAE and du have invested more than $816 million (AED3 billion) in their mobile infrastructure and collectively rolled out more than 11,000 sites.

The neighboring country, Qatar, was in second position with a median 5G download speed of 613.08 Mbps. Operators have been deploying additional sites and activating new 5G carriers in the lead-up to and during the AFC Asian Cup at the beginning of 2024. This helped to achieve significant advancements in 5G network speed in the country from September 2023.

South Korea is considered a pioneer in 5G deployment and adoption of 5G technology, launching the world’s first commercial 5G network in April 2019. The nation’s telecommunication providers have undertaken substantial investments in expanding their 5G infrastructure, supported by governmental initiatives to achieve nationwide coverage exceeding 90% by 2026. According to Speedtest Intelligence data, the median download speed in the first half of 2024 was 524.99 Mbps.

Brazil showcased impressive download speeds at 443.23 Mbps considering its size compared to all other countries featured in the leaderboard. The activation of 5G in July 2022 and access to large amounts of dedicated spectrum, combined with the continuous expansion of 5G services since 2023, propelled the country into the top 4 fastest 5G markets.

e& UAE tops the global 5G performance leaderboard

We examined the 5G performance of the top operators globally during Q1-Q2 2024 using Ookla’s Speed Score metric. We considered operators that lead in Speed Score in their domestic markets. The score combines upload and download speed metrics, with download speed contributing 80% and upload speed 20% to the final score. Both upload and download speed components are measured using a modified trimean, which combines speeds from the 10th percentile, 50th percentile (median), and 90th percentile in a weighted average using a 1:8:1 ratio.

e& UAE consistently led in 5G Speed Score, showing a strong upward trend over time, indicating solid and sustained improvement in 5G performance. 5G median download speed steadily increased to 749.63 Mbps during H1 2024, with a peak speed of 1380.54 Mbps. e& UAE’s upload speed performed well, with a median of 43.52 Mbps and a peak performance of 134.78 Mbps.

Since Q1-Q2 2023, Ooredoo (Qatar) has gradually improved its 5G Speed Score, moving to second position in Q1-Q2 2024 with a download speed of 666.14 Mbps. Ooredoo’s upload speed also improved slightly, though less significantly than the download speed. 

LG U+ (South Korea) displayed a relatively stable Speed Score with modest improvements over time. Its download speed reached 546.83 Mbps in H1 2024, and its upload speed followed a similar pattern, with a slight jump during the first half of 2024 to 44.28 Mbps, securing the second position.

Claro (Brazil) experienced a significant rise in Speed Score since H1 2023, following the launch of 5G SA, reflecting the sharp rise in download and upload speeds during that period, before stabilizing at 501.29 Mbps and 38.45 Mbps, respectively, in H1 2024.

Claro‘s Speed Score in Guatemala declined in the second half of 2022 before recovering in the first half of 2024. However, upload speed exhibited a sharp decline throughout 2022 and 2023 before recovering by the end of the period.

The Speed Score of A1 (Bulgaria) declined consistently, along with reduced download and upload speeds, indicating network capacity constraints. However, it achieved the sixth-highest Speed Score in the world, with a median download speed of 439.37 Mbps and a median upload speed of 34.08 Mbps.

E& UAE’s leadership in 5G speeds translated into a higher customer satisfaction score

e& UAE emerged as the fastest 5G operator in the U.A.E. during the first half of 2024, achieving a median download speed of 861.52 Mbps, significantly ahead of du’s 651.30 Mbps in Q3 2024. e& UAE’s steady improvement in 5G performance since 2022 boosted the U.A.E’s 5G download speed.

Since 1H 2022, e& UAE has outperformed du in terms of 5G download speeds. From mid-2022 to mid-2023, the performance gap between e& UAE and du widened significantly as e& UAE demonstrated consistent growth in download speeds, reaching 861.52 Mbps in Q3 2024. This growth highlights e& UAE’s substantial investment in infrastructure and success in delivering faster 5G services. 

e& UAE achieved near-nationwide coverage of 5G and deployed 5G SA technology in 2023, first as a fixed wireless access (FWA) solution and subsequently as a mobile service. The deployment of 5G SA enables higher throughput, enhances basic services such as voice, and improves advanced consumer and enterprise applications such as AR/VR and IoT.

Net Promoter Score (NPS), a key performance indicator of customer experience, generally positively correlates with improved network performance. Speedtest Intelligence data shows that mobile subscribers who use 5G universally rate their network operators a higher NPS than those on 4G LTE.

The U.A.E. has a similar trend, with 5G users more likely to recommend their operators than those on 4G. e& UAE, recognized as the fastest 5G operator in the market, leads with an NPS of 53.8 in Q1-Q2 2024, outperforming du with an average NPS of 36.7.

Competing on 5G speed can help operators be better prepared for 6G

Speed competition remains critical for mobile telecom operators, even though current network capabilities allow users to accomplish virtually any task. The pursuit of higher speeds is not merely a technical ambition; it reflects a commitment to excellence that appeals to both consumers and investors and positions the country as economically competitive on the global stage. 

Speed is also a highly relatable concept for consumers, driving their choice of service providers and stimulating spending. This makes speed an essential marketing tool for operators, who should leverage their network’s performance to maintain customer loyalty and protect their market share.

Beyond its immediate consumer appeal, speed competition plays a crucial role in supporting the development of the broader ICT industry. Speed helps companies stay competitive by leveraging technologies such as AI, cloud, and edge computing to enable real-time decision-making, operational efficiency, and rapid innovation.

In essence, competing in 5G speed is not just about delivering exceptional service today; it’s about laying the groundwork for tomorrow’s innovations, as illustrated by the U.A.E’s recent announcement of its 6G strategy. By investing in scalable and adaptable infrastructure now, operators ensure that they remain at the forefront of technology innovation. This proactive approach to network development also helps operators cultivate the expertise and skills necessary to anticipate market demand better and inform service enhancements and new offerings. Ultimately, telecom operators are building a foundation for long-term success to support next-generation digital services.

Ookla will continue to monitor 5G performance and deployments globally and benchmark countries and operators. For more information about Speedtest Intelligence data and insights, please contact us.

In the last three years, generative AI (GenAI) has driven most of the hype around AI. Its ability to understand and respond to queries in a human-like way and produce high-quality programming code, music, and video content raises critical questions about network requirements to support these applications and how best to measure network performance. This article covers current and emerging GenAI use cases and discusses the implications of growing consumer adoption on network performance and measurement strategies.

Key Takeaways:

• The growing adoption of GenAI and its evolution has the potential to disrupt traffic usage patterns significantly. The surge in demand for GenAI and the diversity of the content types it can generate could lead to a significant increase in data traffic and a heightened need for improved bandwidth and latency in the medium to long term. Advancements in data compression and prompting techniques could mitigate demand pressures and lead to less severe outcomes than anticipated.
• GenAI drives the need for new approaches to measure the quality of experience. The evolution of GenAI from chatbots to real-time conversations with richer and more interactive applications will require a new content distribution infrastructure and unlock new use cases. This will prompt a rethink of how we assess network performance and user experience.
• Addressing data misuse, misinformation, and mistrust concerns will be crucial for widespread GenAI adoption. As GenAI content becomes more sophisticated, privacy, security, and trustworthiness issues will grow in importance. Establishing robust regulatory and technical safeguards will be vital to building consumer and business confidence in these technologies.

GenAI has captivated users, businesses, and investors alike

GenAI is a small part of AI that specializes in generating novel content using computer models trained on terabytes of existing digital datasets of text, computer code, images, and videos. Despite the recent decline in overall AI private investment, GenAI continues to attract the most funding. Private funding for GenAI surged to $25.3 billion in 2023, nearly nine times the amount raised in 2022. In October 2024, OpenAI, the company behind ChatGPT, the leading GenAI chatbot, raised $6.6 billion in new funding, valuing it at $157 billion despite incurring $5 billion in losses that year.

GenAI is not just capturing the attention of investors; it is also a hot topic among Fortune 500 companies. GenAI was the most frequently cited AI theme on earnings calls, appearing in nearly 20% of all earning calls during 2023. This shows that corporate leaders see significant potential in this technology.

GenAI services have also become very popular among consumers. For example, ChatGPT made headlines with its record-breaking growth. After the release of ChatGPT 3.5 in May 2023, the service reached 100 million users in just 64 days – the fastest-growing consumer application to date. Today, the ChatGPT website attracts over 600 million monthly visitors (down from its peak of 2 billion visits a month), and the service has more than 180 million users. While growth has slowed, ChatGPT is expected to capture 75.2% of GenAI users globally, even as its market share dips due to the rise of many proprietary and open-source alternatives. The U.S. is home to the largest ChatGPT user base, expected to increase to 87.9 million users by 2025.

Businesses are eagerly embracing GenAI for its perceived benefits. A CCS Insight survey found that 82% of U.S. and European companies have already deployed GenAI or plan to do so in the next 12 months. GenAI has been mostly used for administrative-type tasks such as generating draft contracts and summarizing notes, content creation, marketing, and customer support.

The telecom industry is increasingly bullish on GenAI. Operators are using the technology to introduce more sophisticated chatbots that offer a more human-like customer service experience. According to a recent GSMA Intelligence survey, 81% of global telecom operators are currently testing GenAI. While the initial focus is on cost reduction and operational efficiency, GenAI could eventually create new revenue streams by enabling new AI-powered consumer and B2B services. A 2023 McKinsey report suggests that GenAI could boost telecom revenue by 2.3% to 3.7% (as a percentage of total industry revenue), corresponding to an additional $60 to $100 billion. 

The rapid evolution of GenAI services could lead to increased network requirements 

GenAI is a broad category of applications that can generate different types of content. Most users are familiar with large language models (LLM), such as GPT-4, which powers ChatGPT. LLMs are a subset of GenAI trained on vast amounts of text to generate responses in natural language. As such, most current interactions with GenAI services are basic text-based conversations. These conversations are latency-tolerant and do not demand a high throughput since most of the delay is on the server side, where data processing and content generation take place.

The evolution of GenAI, from chatbots to real-time conversations with richer and more interactive content, will require more powerful servers that can process complex requests, do it faster, and send back high-resolution images or videos. In this context, minimizing downlink latency is crucial to maintaining natural conversations between users and GenAI agents. Additionally, download speed becomes essential for sending and rendering media content on users’ devices. In the future, the ease of creating realistic images or videos (or enhancing them) will encourage more people to become content creators and editors. This will enable new use cases that combine real and synthetic content generated on the fly and streamed live. In this scenario, good uplink latency and throughput metrics will be essential to provide viewers with a high-resolution and lag-free experience.

The reduction in the size of the LLM models, the increased optimization of GenAI applications, and the growing power of system-on-chips (SoCs) have given rise to a new generation of GenAI-powered smartphones. Devices such as the iPhone 16 with Apple Intelligence and Samsung Galaxy S24 with Galaxy AI are slowly transforming the way we interact with our phones and consume content, shifting from individual app usage to an ‘app-free’ experience with an intelligent, voice-controlled digital assistant that can draft emails, translate conversations, make a booking, or manage calendars. According to IDC, GenAI smartphones will be the fastest-growing smartphone segment through 2028, with over 234 million units expected to ship in 2024 alone. 

The proliferation of dedicated GenAI applications on smartphones will help to engage more users with GenAI and shift their usage from static to nomadic. While basic tasks will continue to be handled on the device for responsiveness and data privacy, more complex tasks like generating a picture or a video clip will be partially offloaded to a server (similar to Apple’s Private Cloud Compute). In this case, download speed and latency could be paramount to ensuring a good customer experience. Even for offline GenAI usage or when using on a portable processing unit (for example, embedded in a wireless router), mobile users may need to regularly download and update large GenAI models, highlighting the importance of robust cellular download speeds.

There is much uncertainty about GenAI’s overall impact on traffic profiles. The widespread adoption of GenAI media content could significantly disrupt consumer usage patterns, potentially leading to a surge in data traffic that could strain existing network infrastructure. On the other hand, advancements in data compression and prompting techniques could mitigate demand pressures, leading to less severe outcomes than anticipated.

Service providers will need to monitor developments and be prepared to make further investments to upgrade bandwidth and capacity to keep pace with demand and maintain performance levels. If GenAI leads to substantial additional traffic, then operators may ask GenAI service providers to pay their ‘fair share’ use of the infrastructure. 

GenAI drives the need for new service quality metrics and facilitates network optimization for operators

The rise of GenAI is transforming the way we create and consume content. This transformation necessitates new means of collecting network data and methodologies to capture the performance of AI-powered services accurately. Traditional metrics like throughput, packet loss, and jitter remain essential, but new KPIs need to be defined to reflect the ‘perceived quality’ of generated (or enhanced) content. Some aspects that could be considered include the relevance and accuracy of the generated content and the responsiveness of interactive GenAI applications.

The infrastructure required to support the growth of GenAI is likely to be different from traditional content delivery networks (CDNs). Some use cases will support near real-time interaction and require robust data privacy, necessitating edge computing deployment to offload the central cloud-based GenAI servers, reduce latency by processing data closer to the user, and minimize the amount of traffic exchanged between the user and the central server. Therefore, future network measurement approaches should factor in the distributed nature of these edge compute units and redefine latency to include edge processing and content generation time at the edge. 

Operators can use GenAI tools to analyze measurement data, identify patterns, and predict potential issues based on recent incidents. According to a 2023 GSMA Intelligence survey, two-thirds of operators considered troubleshooting and maintenance to be the biggest impact of GenAI. For example, GenAI tools can provide network engineers with automated, tailored network analytics and upgrade recommendations. Ookla is using GenAI in Downdetector Explorer™ to provide network maintenance teams with up-to-date situation reports following service disruptions.

GenAI promises more immersive experiences for consumers while driving network measurement innovation

We are at the beginning of the GenAI revolution, and its potential is undeniable. GenAI promises consumers a more immersive experience, increased engagement, and greater creativity and content sharing. Combining AI-generated text, code, audio, and video could enable highly personalized and hyper-realistic immersive experiences beyond what is currently available. In the medium term, real-time services like video conferencing or gaming could be enhanced with AI-generated content, such as virtual avatars and personalized game characters, pushing network requirements to new heights. This shift will also significantly impact content distribution infrastructure and demand new approaches to moderating content, measuring network performance, and assessing customer quality of experience.

As GenAI content becomes more sophisticated and indistinguishable from human-produced content, concerns around data misuse, the spread of misinformation, and mistrust will grow. Addressing these issues will be crucial for fostering consumer and business confidence, ultimately driving the widespread adoption of AI technologies. In the meantime, GenAI will continue to be deployed with human supervision while we improve their understanding of its risks and limitations and establish regulatory and technical safeguards for its use.

We will continue to follow the development of GenAI and its adoption by consumers and telecom operators to gauge its impact on network performance and optimization. Furthermore, Ookla is actively using GenAI to transform its data into actionable insights and predictive analytics that help providers understand how to deliver better real-world connectivity experiences to consumers.