Vladislav Zaimov is a distinguished telecommunications expert with an extensive background in managing high-stakes enterprise networks and securing vulnerable wireless infrastructures. As the industry pivots toward the next generation of connectivity, Zaimov’s insights into the shift from consumption-heavy 4G and 5G models to the production-oriented architecture of 6G are highly sought after. In this conversation, we explore the rise of “Physical AI,” the technical hurdles of the 6GHz spectrum, and the economic realities facing global operators.
The following discussion examines the transition toward uplink-centric networking, the potential of “giga MIMO” to solve propagation issues, and the strategic value of existing FDD bands versus new TDD spectrum. We also address the competition between cellular and Wi-Fi in indoor environments and the specific performance metrics required to make 6G a viable industrial tool.
Most mobile networks prioritize downloads, but physical AI applications like bedside robots often require an uplink-heavy traffic ratio of 70%. How will 6G architecture accommodate this reversal of data flow, and what specific modifications to radio units are necessary to support such massive uploads?
To handle a 70% uplink ratio, 6G must fundamentally decouple the uplink and downlink functions, moving away from the asymmetric designs of the past. In current 5G setups, the hardware is heavily biased toward the downlink to support streaming, but for a robot that needs to feed sensory data to the cloud in real-time, we need a complete separation of these paths at the radio level. This involves a shift in radio unit flexibility, where the hardware can dynamically allocate resources to the upload path without being bottlenecked by the download architecture. We are looking at a future where the radio side is much more modular, allowing for a software-defined approach to traffic flow that treats the upload as the primary mission rather than an afterthought.
The 6GHz “golden band” offers high capacity but faces significant challenges with signal propagation and indoor penetration. How can operators effectively use beamforming or “giga MIMO” to overcome these physical obstacles, and what are the practical limits of using these bands without massive site densification?
The 6GHz band, stretching up to 8.4GHz, acts much like a high-speed lane that is unfortunately blocked by every wall and window it encounters. To solve this, we are evolving from the Massive MIMO of 5G into what we call “giga MIMO,” which packs even larger arrays of antenna elements into the radio unit to steer signals like heat-seeking missiles directly at the user. While these “pencil beams” can extend the reach of the signal, they cannot entirely defeat the laws of physics regarding indoor penetration. Without densifying the network by building more sites, operators will find that 6G struggles to match the existing 5G grid coverage, potentially leaving “blind spots” in the very indoor environments where AI gadgets are most likely to be used.
While new spectrum is being eyed for 6G, many operators already possess unused uplink capacity in existing FDD bands. Why is there a push for new TDD-based spectrum instead of repurposing these “empty highway lanes,” and what hardware trade-offs does this choice entail for global roaming?
The push for TDD in the 6GHz “golden band” is driven by the sheer amount of contiguous bandwidth available, which is easier to manage for high-capacity bursts than the fragmented FDD lanes we currently have. Many operators are indeed sitting on “empty highways” in their FDD spectrum, but leveraging these requires a complex multiband strategy and new investments in FDD-based Massive MIMO hardware, which companies like Verizon are just beginning to explore. The trade-off is one of global standardization; TDD allows for more flexible allocation of uplink and downlink on the same frequency, but if different regions choose different chunks of the 6GHz band, the dream of seamless global roaming for AI robots becomes a much more expensive hardware challenge for manufacturers.
Network densification requires substantial capital, yet recent trends show a significant decline in radio access network revenues. If 6G is viewed as an evolutionary step rather than a total overhaul, how can telcos justify the investment, and which specific AI-driven efficiencies might offset these infrastructure costs?
It is a difficult sell when RAN revenues have plummeted from $45 billion in 2022 to an expected $35 billion by 2025, signaling a cautious market. However, the justification for 6G lies in its “evolutionary” nature, where telcos can implement AI-driven software upgrades to squeeze more spectral efficiency out of existing hardware without a total “rip and replace” scenario. By using AI to optimize how the network handles capacity and power, operators can delay the massive costs of densification while still offering improved performance. The goal is to prove that 6G can deliver significantly better results with only about half the number of new base stations compared to previous generations, relying on satellite integration and software intelligence to fill the gaps.
Data shows that roughly 80% of mobile device traffic currently occurs over indoor Wi-Fi rather than cellular networks. Given that many AI gadgets will be used inside homes or hospitals, how will 6G provide a more compelling value proposition than fiber-backed Wi-Fi for high-bandwidth uplink tasks?
The 80% figure for indoor Wi-Fi usage is a sobering reality for cellular operators, especially since physical AI like bedside assistants or industrial visors will operate primarily indoors. To compete, 6G must offer something Wi-Fi often lacks: guaranteed quality of service and seamless mobility across wide areas without the “handoff” drops common in local networks. If a robot moves from a hospital ward to an outdoor transport bay, 6G provides the continuous, high-reliability uplink that a fiber-backed Wi-Fi network cannot maintain once the device leaves the building. The value proposition is not just about raw speed, but about a managed, secure, and ubiquitous connection that treats a hospital or a factory as a single, uninterrupted data environment.
Uplink traffic is reportedly growing faster than downlink traffic for the majority of global operators. Beyond just speed, what latency and reliability metrics must 6G hit to support real-time industrial AI, and how will software-driven upgrades bridge the gap before the 2029 rollout?
We are seeing uplink traffic growth outpace downlink for 80% of operators, which suggests that the “production” side of the internet is finally waking up. For real-time industrial AI, 6G needs to provide ultra-low latency that is consistent, not just “best effort,” to ensure that a worker’s AI visor or a remote-controlled robot reacts instantly to environmental changes. Between now and the 2029 rollout, software-driven upgrades in the 5G Radio Access Network will be the bridge, allowing telcos to experiment with AI-driven resource management. These software layers can prioritize critical uplink data packets today, preparing the infrastructure for the massive, three-fold increase in upload demands we expect to see every five years starting in 2025.
What is your forecast for 6G?
I expect 6G to be the first generation where the “G” cycle as we know it—marked by massive hardware overhauls—begins to fade in favor of a continuous, AI-led evolution. We will see a “multiband” reality where the 6GHz spectrum provides the high-capacity heart of the network, while repurposed lower bands provide the reach, creating a hybrid environment specifically tuned for the 70% uplink demands of physical AI. While the initial rollout in 2029 may be measured, the shift toward a network that “listens” as much as it “broadcasts” will fundamentally change the economics of the industry, making cellular connectivity indispensable for industrial automation in ways that 5G only hinted at.
