Every new wireless generation arrives with the same headline claims: faster speeds, lower latency, more connected devices. We’ve heard it since the jump from 3G to 4G, and again with 5G. With every new generation, power, heat, and reliability are still the hard constraints, especially once the hardware leaves the lab.
On paper, the targets are ambitious. Terabit-per-second peak data rates and extremely low latency are now common talking points. But for anyone building embedded hardware, be it radio modules, base stations, edge devices, or industrial gateways, speed isn’t the hard part anymore. Control is.
6G isn’t shaping up to be a ‘bigger 5G’. It’s shaping up to be a network generation defined by how precisely we can manage energy, thermal limits, and system complexity at scale. And that is a very different kind of engineering problem.
Is the quest for performance becoming a liability?
Massive MIMO and mmWave spectrum provided the required bandwidth but forced designers to confront physical constraints that had often been ignored – antenna arrays grew, RF front ends became denser, and power budgets increased while thermal margins tightened.
Anyone who has sat through a design review where the thermal model barely fails in a few edge cases knows what usually happens next: those edge cases show up in the field, and they show up fast. 6G isn’t a reset. Think of it as travelling in the same direction, just with less room for error.
5G pushed into mmWave above 6 GHz, but early 6G roadmaps extend that further into the 7-24 GHz range and toward sub-terahertz operation, roughly 100 GHz up to 1 THz (Figure 1).
At those frequencies, propagation is no longer forgiving. Range drops sharply, penetration through common materials becomes unreliable, and the practical coverage footprint starts to look more like a short-range link than a traditional cellular cell. In many realistic environments, you’re dealing with tens of meters, not kilometers, and that has immediate consequences for deployment density and hardware design (Figure 2).
If your link budget is tight, you don’t get to hand-wave obstacles away - walls matter and weather matters. Even the way a device is held or mounted can matter. In other words, 6G pushes wireless closer to the world of optical links, except we’re still pretending it’s cellular.
Ultra-massive MIMO meets embedded reality
Much of the optimism around 6G depends on scaling MIMO even further. 5G popularized massive MIMO arrays with 16, 32, or 64 elements. 6G discussions go beyond that, describing ultra-massive or even giga-MIMO with hundreds or thousands of antenna elements (Figure 3).
Scaling MIMO does improve spatial multiplexing and beam control, but it also shifts the burden onto the hardware. An antenna element at this point is not a passive component; it includes converters, phase control, amplification, calibration circuitry, and interconnects needed to keep the array coherent.
As the array grows, the supporting electronics quickly dominate the design. Power consumption rises, heat density increases, and thermal behaviour starts to limit how long the system can operate at full performance. In deployed infrastructure, those limits show up as reliability and availability concerns rather than abstract performance trade-offs.
Effective thermal management is a key challenge as antenna arrays grow and hardware is pushed toward mast-head deployments. Thermal issues are often the first thing teams solve on paper and the last thing they solve in the field.
Beamforming: essential, not optional
Higher frequencies can deliver impressive bandwidth, but they also incur signal loss and limited penetration through obstacles. That is why beamforming is no longer a feature - it’s a requirement (Figure 4).
Beamforming is already a major factor in making mmWave viable in 5G, and 6G will push it further, especially as satellite integration becomes more prominent. Hybrid beamforming is a likely approach, balancing analog efficiency with digital flexibility.
Hybrid beamforming is an engineering compromise, acknowledging that pure digital beamforming scales poorly in cost and power. But hybrid approaches introduce calibration challenges, drift issues, and failure modes that are harder to diagnose. And, when a system is deployed on a tower or satellite link, ‘harder to diagnose’ becomes ‘harder to mend’.
AI isn’t magic – it’s orchestration
A 6G network is expected to make real-time operational decisions, such as beam selection, antenna activation, bandwidth scheduling, and power management, based on traffic demand and channel conditions. That kind of control loop becomes difficult to scale with predefined, fixed criteria, especially once the network is operating across mixed spectrum bands and heterogeneous infrastructure.
But, where does this intelligence actually live? If it sits close to the radio, the embedded platform needs sufficient compute and memory bandwidth to keep up, along with deterministic, low-latency interfaces to the RF chain. If the intelligence sits further upstream, there is greater reliance on the backhaul and a higher risk of new failure modes. Either way, it changes the hardware, thermal budget, and cost model. AI also introduces security concerns, because a more adaptive network may be an appealing target.
Energy becomes the hard limit
A subtle but important theme in early 6G work is the shift from performance-first to efficiency-first. Energy efficiency is no longer a secondary benefit – it’s a requirement.
Strategies such as ultra-massive MIMO, AI-based dynamic power adjustment, and renewable-powered infrastructure tick the sustainability and energy-reduction boxes. And, zero-energy devices that harvest energy from RF, solar, or kinetic sources reduce dependence on batteries. From an embedded standpoint, that’s where things get interesting.
Zero-energy devices sound futuristic, but they’re really just an extension of ultra-low-power design philosophy. The trade-offs are immediate: shorter duty cycles, burst-based communication, aggressive sleep states, and often a complete rethink of what ‘always connected’ means. In other words, the network might be 6G, but the endpoint design starts to resemble energy-harvesting sensor nodes that embedded engineers have been building for years. The difference is scale. 6G assumes millions (or billions) of these devices.
The deployment cost problem doesn’t go away
Here’s the part that rarely makes it into the hype cycle: economics. Higher frequencies mean shorter range. Shorter range means more infrastructure. More infrastructure means more cost. That maths doesn’t change because the technology is newer.
5G deployment costs were substantial, and the rollout of 6G will again require new infrastructure, including fiber, antennas, towers, and supporting equipment. To support broadband access in remote areas, satellite integration is likely to include Very Low Earth Orbit (VLEO) satellites at an altitude of approximately 350 km (Figure 5).
Satellite integration is promising, especially for rural coverage and resilience. But it doesn’t address the challenges of densely populated urban areas. Urban deployments still require dense terrestrial infrastructure. Engineers know this, and operators know too. The question is who pays for it, and what the business model looks like. If 5G taught us anything, it’s that ‘build it and they will come’ is not a deployment strategy.
What 6G success might actually look like
If 6G is successful, it probably won’t be because consumers suddenly start bragging about terabit-per-second speeds. Most people don’t run into bandwidth limits on a phone often enough to care, and when they do, it’s usually the backhaul or the app that’s the bottleneck.
Digital twins that update in near real time, city-scale sensing, medical wearables with continuous monitoring, immersive XR, and eventually more ambitious remote applications are all possible. But they won’t be unless the underlying RF and edge compute stack deliver consistently low latency and high uptime.
From an embedded engineering standpoint, the main takeaway is straightforward. 6G will be sold as a speed upgrade, but the bigger change is the way networks are expected to operate. Instead of operating as best-effort data pipes, they’re being pushed toward continuous optimization, switching resources on and off, dynamically shaping beams, managing energy consumption, and adapting to demand in real time.
So, the ‘6G challenge’ is not a single problem that can be solved in software. It’s a systems problem that pulls RF, compute, power delivery, packaging, thermals, security, and manufacturability together. Embedded engineers have faced those trade-offs for decades. 6G just drags them into the center of the network.
Murata is helping to lay the technological foundation for intelligent, resilient, and energy efficient networks, enabling 6G to move beyond connectivity and become a catalyst for innovation, societal progress, and a more connected digital future.
Explore more here how Murata is powering next-generation communication solutions and the technologies shaping what’s next.
Have questions or want to dive deeper? Connect directly with the team to continue the conversation.