Fact-Checking the Complaints: Is Malaysia’s mobile network really as bad as netizens claim? 

CelcomDigi’s modernised 4G network

Celcom and Digi completed their merger in 2022 and now operate under CelcomDigi. Following the merger, the company began consolidating and modernising the two previously separate radio access networks.

CelcomDigi currently operates a substantial LTE spectrum portfolio across several frequency bands, including:

1. Band 1: 20 MHz

2. Band 3: 2x 20 MHz carriers

3. Band 7: 2x 10 MHz carriers

4. Band 8: 10 MHz

When calculating the usable bandwidth of paired FDD spectrum, I refer to the assigned channel bandwidth rather than counting the uplink and downlink portions twice.

CelcomDigi’s available spectrum allows the network to support multiple carrier aggregation combinations. The company has also worked with Huawei and ZTE to modernise its infrastructure through technologies such as 4×4 MIMO, 256QAM, active antenna units, Twin Beam solutions and indoor radio systems such as pRRUs.

These improvements are particularly noticeable at modernised sites with sufficient backhaul capacity.

Nevertheless, the configuration is NOT consistent across every region.

From my observations, many Huawei-based sites around Kuantan, Pahang DO NOT broadcast the 2x 10 MHz Band 7 carriers commonly observed on ZTE-based sites in Klang Valley. This means that users in different states may experience different carrier aggregation combinations even when using the same operator and device.

This regional variation is one reason why it is difficult to evaluate an operator based on only one speed test or one location.

Why a strong 4G network still matters

As more UE is set to prefer 5G, a portion of mobile traffic is shifted away from LTE and onto the respective 5G network.

For CelcomDigi, Maxis and Yes users, 5G traffic is generally carried through Digital Nasional Berhad’s network where coverage is available. This can reduce the load on the operators’ LTE networks and potentially improve the 4G experience for users who remain connected to LTE.

However, offloading traffic to 5G does NOT eliminate congestion. It can simply move the congestion from the operator’s 4G network to the shared 5G network.

This is particularly relevant in Malaysia because multiple operators may serve their users through the same DNB infrastructure within a particular area.

DNB’s expanding 5G capacity

Malaysia’s 5G landscape has become more complex following the development of the second 5G network.

U Mobile has migrated its customers to its own ULTRA5G network, while unifi mobile has also moved towards U Mobile’s wholesale 5G infrastructure. DNB continues to serve users from major operators including CelcomDigi, Maxis and Yes.

DNB has access to a large amount of mid-band spectrum, including up to 200 MHz of n78 spectrum between 3.3 GHz and 3.5 GHz, alongside low-band 700 MHz spectrum used for wider coverage.

However, not every DNB site currently broadcasts the full 200 MHz n78 allocation.

From my observations, DNB’s network can generally be divided into several deployment configurations:

1. Outdoor sites with newer or capable active antenna units, broadcasting up to 200 MHz of n78 together with 700 MHz coverage.

2. Outdoor sites using older antenna equipment, continuing to broadcast approximately 100 MHz of n78 together with 700 MHz.

3. Indoor deployments using ZTE QCell systems, which commonly broadcast 100 MHz of n78 and 700 MHz. I have not personally observed a 200 MHz indoor QCell deployment so far.

Several sites in Klang Valley have already been upgraded to broadcast 200 MHz of n78. At these locations, the improvement can be significant, with substantially higher peak throughput and greater capacity

Should devices be moved back to 4G during severe 5G congestion?

This raises an interesting network-management question.

When a 5G cell becomes severely congested, should operators implement more aggressive policies to move users back to LTE where sufficient capacity is available?

From a user-experience perspective, the objective should not simply be to keep the device connected to 5G. The objective should be to provide the most stable and responsive connection available.

If a 5G cell is delivering only 1–2 Mbps while the LTE network has spare capacity, remaining on 5G may not be the most efficient outcome for either the user or the network.

Smarter mobility, load-balancing and traffic-steering policies could potentially improve the experience by considering actual cell utilisation and service quality rather than prioritising the network generation displayed on the phone.

This is especially important because most consumers DO NOT KNOW whether their operator’s 5G service is provided by DNB or U Mobile.

When the service performs poorly, users are likely to blame the brand displayed on their SIM card rather than the underlying wholesale network. From the customer’s perspective, this is understandable because they subscribe to and pay their chosen mobile operator.

The technical distinction between the mobile operator and the wholesale 5G infrastructure is largely invisible to the average consumer.

U Mobile ULTRA5G deployment observations

U Mobile’s ULTRA5G network currently serves U Mobile users, unifi mobile users and supported MVNO customers.

I've previously written an article about the deployment of ULTRA5G in Chinese.

Because U Mobile is developing an independent 5G infrastructure, its traffic profile differs from DNB’s shared network. At present, DNB carries a significantly larger combined user base from multiple established operators, while ULTRA5G serves a comparatively smaller number of users.

This difference in user load must be considered when comparing speed-test results.

A less congested network may produce higher speeds, but this does not automatically mean that its architecture or equipment is superior. If a large proportion of Malaysia’s mobile users migrated to U Mobile’s network, ULTRA5G would also experience greater loading, while DNB’s performance would likely improve as traffic moved away.

Therefore, network comparisons should consider:

• The number of users connected to each network

• Spectrum bandwidth

• Site density

• Backhaul capacity

• Indoor and outdoor deployment strategies

• Device capability

• Time of day

• Traffic demand within the coverage area

From observations shared between my friend and me across Peninsular Malaysia, U Mobile appears to configure some of its N78 active antenna units and N28 coverage with relatively high transmission power.

The network frequently uses Huawei MetaAAU or BladeAAU equipment. In some locations, the N78 signal strength can appear stronger than U Mobile’s Band 3 LTE, which is an interesting characteristic of the deployment.

U Mobile also appears to be deploying Huawei LampSite solutions aggressively for indoor coverage. Huawei pRRUs can now be observed in several major shopping malls around Klang Valley, helping to improve indoor signal availability and network capacity.

Indoor coverage is an important part of the overall network experience. A network may have excellent outdoor macro coverage but still perform poorly inside shopping malls, offices, universities, hospitals and residential buildings without a properly designed indoor system.

Peak speed is only one part of network quality

Malaysia’s mobile networks have demonstrated impressive peak results.

Modernised CelcomDigi LTE sites can deliver more than 600 Mbps through 5xCA  aggregation. Maxis can provide 4xCA LTE carrier aggregation using Bands 1, 3, 7 and 8. DNB’s 200 MHz n78 deployment can approach extremely high gigabit-class speeds under favourable conditions. U Mobile’s ULTRA5G network has also demonstrated very strong performance, including results exceeding 1 Gbps.

However, peak speed does not represent the complete customer experience.

A high-performing speed test requires several elements to work together:

• Sufficient radio spectrum

• Strong and clean signal conditions

• Suitable MIMO rank

• High-order modulation

• A capable smartphone modem

• Low cell utilisation

• Adequate transport and backhaul capacity

• Sufficient core network capacity

• A nearby and uncongested test server

If any part of that chain becomes a bottleneck, the user may experience poor performance even when the radio network itself supports advanced technology.

This is why adding more spectrum or installing a new active antenna unit does not automatically solve every problem. A site broadcasting 200 MHz of n78 can still underperform if its backhaul connection is insufficient or if too many users are competing for capacity.

Is Malaysia’s mobile network actually bad?

My personal answer is no—but the quality remains inconsistent.

Malaysia’s mobile infrastructure has improved considerably since the Movement Control Order period. Operators have modernised equipment, increased carrier aggregation capabilities, expanded indoor coverage and deployed more advanced antenna systems.

The collaboration between the government, regulators, infrastructure providers and mobile operators through programmes such as JENDELA has also contributed to broader mobile coverage and network improvements.

At the same time, there are still genuine problems.

Some locations suffer from severe congestion. Some sites have insufficient backhaul. Certain indoor areas remain poorly covered. Upgrades are not always deployed quickly enough in high-demand locations. Network configurations also vary significantly between regions.

These issues should NOT be ignored.

However, capacity limitations are not unique to Malaysia. Mobile operators around the world experience congestion, uneven coverage and backhaul constraints, including those operating in countries widely regarded as leaders in 5G deployment.

The more accurate conclusion is that Malaysia does not have a universally poor mobile network. Instead, Malaysia has a network that can perform exceptionally well in some locations and poorly in others.

The challenge is consistency.

Looking beyond the network-generation icon

I first started using 4G in Malaysia around 2015. Comparing that experience with the network available in 2026, the improvement in both coverage and speed is substantial.

Today, a modernised LTE site can deliver speeds that would previously have been associated only with early 5G marketing. Advanced LTE features such as carrier aggregation, 4×4 MIMO and 256QAM continue to make 4G an important and highly capable part of Malaysia’s mobile infrastructure.

Instead of evaluating a network based only on whether the phone displays “4G” or “5G,” users should consider the actual experience:

• Is the connection stable?

• Is latency acceptable?

• Can applications load consistently?

• Does the network remain usable during peak periods?

• Is indoor coverage available?

• Can the site handle the number of users in the area?

A reliable 100 Mbps LTE connection may be more valuable than a 1 Gbps 5G result that collapses to 1 Mbps when the network becomes busy.

Malaysia’s operators have made meaningful progress, particularly through network modernisation and more advanced radio deployments. The next major improvements will depend not only on spectrum, but also on site densification, backhaul expansion, indoor coverage, load balancing and faster upgrades at congested locations.

There is still considerable room for improvement, but the technical capabilities already present in Malaysia’s mobile networks should also be recognised.

Hats off to all the staff and engineers who make these networks possible!