NVIDIA's 6G Declaration: Is This a Revolution or the Spark of a Bubble?

If you find this helpful, please click the ♡. You can save it to read later.

Conclusion:

6G and AI robots integration will revolutionize manufacturing, autonomous driving, and urban infrastructure through ultra-low latency communication, boosting semiconductor and data center stocks. Meanwhile, telecom carriers will see their profit margins squeezed by massive investments. Companies that focus on small-scale PoC verification and collaborate through open ecosystems will be the winners.

Outline:

Chapter 1: The Future of Communications Opened by 6G
• Global movements toward commercialization and Japan's position
• Differences from 5G: An easy-to-understand explanation of low latency, high capacity, and high reliability
• Why is AI robot operation attracting attention?

Chapter 2: NVIDIA's AI-RAN Strategy
• What is AI-RAN? Understanding it through comparison with conventional networks
• "Software-izing" base stations with GPUs and dedicated chips
• Processing speed and power-saving effects seen in demonstration results

Chapter 3: The Challenge of SoftBank and Nokia
• Overview of joint trials underway domestically and internationally
• Benefits of building an open ecosystem
• Introduction of measured data such as power cost reduction and congestion mitigation

Chapter 4: Industry-Specific Use Cases
• Manufacturing lines: Real-time control of collaborative robots
• Autonomous driving: Instant information sharing between vehicles and base stations
• Smart cities: Disaster response drones and high-definition cameras

Chapter 5: Stock Price Impact by Sector
• Semiconductors: Stock prices rising due to the rapid expansion of demand for GPUs and AI chips. For example, AMD is up about 7% in the last 5 days
• Telecommunications carriers: Stock prices flat to slightly down due to the burden of 6G investment
• Network equipment: Supermicro rose about 5% following the announcement of AI-RAN compatible products
• Cloud/Data centers: Related REITs are firm due to increased generative AI traffic
• Industrial automation: Expectations for AI robot adoption are high, but the short-term impact is limited

—

Chapter 1: The Future of Communications Opened by 6G

The world has begun to move in earnest toward the struggle for 6G leadership. The US and EU are increasing research funding, China is building test networks through state-run projects, and private carriers in South Korea are conducting repeated outdoor tests. In Japan, under the Ministry of Internal Affairs and Communications' "Beyond 5G Promotion Strategy," NTT Docomo and NEC have succeeded in terahertz band transmission, and frequency policies are being finalized with an eye toward commercialization around 2030. According to the latest forecasts from research firms, the 6G-related market is expected to expand to over $100 billion by 2035, a scale significantly larger than 5G.

The biggest difference from 5G is the combination of ultra-low latency (less than 1 millisecond) and Tbps-class high capacity, as well as the fact that the network itself will be AI-native. Not only will the radio interface be revamped, but "software-defined RAN," which virtualizes base stations, edges, and clouds to dynamically allocate resources, will be incorporated into standard specifications. This will shorten update cycles to match the software industry, fundamentally changing the recovery model for telecommunications capital investment.

The key is the AI-RAN architecture proposed by NVIDIA. The company has integrated baseband processing and AI inference on the same platform with Grace Blackwell generation GPUs/DPUs, significantly increasing computational performance and power efficiency. SoftBank, in collaboration with Nokia, developed the "AITRAS" orchestrator, which reduced power consumption by up to 35% in field demonstrations. They aim for commercial trials in 2026 and a commercial release in 2027. In the semiconductor supply chain, demand for AI chips is expanding rapidly, and related stocks such as AMD and Supermicro are on an upward trend.

What is important for business people is that 6G will not be a reduction in communication costs, but a new source of revenue. By having AI systems that previously operated independently—such as collaborative robots on manufacturing lines, swarm control of autonomous vehicles, and disaster drones in smart cities—coordinate over the network, the efficiency of the entire supply chain will improve dramatically. At this point, companies that visualize performance metrics and business KPIs through small-scale PoCs and select partners that support an open ecosystem are expected to lead the next decade. Before making investment decisions, it is essential to establish an ROI calculation framework within the company and constantly update market maturity and competitive maps.

—

Chapter 2: NVIDIA's AI-RAN Strategy

AI-RAN is a core technology for next-generation networks that virtualizes baseband processing, which was previously confined to dedicated hardware, onto GPUs/DPUs, allowing for the simultaneous execution of AI inference and radio processing. The driving force is NVIDIA. The company has presented a design to software-ize base stations by integrating hundreds of TFLOPS of computational performance and BlueField-4 DPUs into its latest Grace Blackwell platform.

Demonstrations are also progressing. SoftBank, in collaboration with Nokia, developed the "AITRAS" orchestrator and fully software-ized 16-layer Massive MIMO at an outdoor site. It recorded a 30% improvement in throughput and a 35% reduction in power, paving the way for real-time application of AI inference even in the 6G terahertz band.

In terms of the ecosystem, Nokia has optimized its anyRAN software for NVIDIA GPUs and completed functional tests with T-Mobile and Vodafone. It announced that it will begin commercial pilots at the end of 2026. BT, Elisa, and NTT Docomo have also joined the AI-RAN Alliance and are formulating common specifications for AI to autonomously optimize RAN functions on a per-container basis.

The hardware supply chain is also becoming active. Supermicro announced the 2U server "ARS-221GL-NR," which can be equipped with up to two Grace Superchips and Blackwell GPUs. On the day of the launch, the stock price rose 5%, and there are forecasts that sales for communication infrastructure will increase by 30% this fiscal year. In the semiconductor market, AMD announced plans to double the supply of its MI300 series GPUs, clearly showing its stance to capture AI-RAN demand.

The advantages of software-defined are significant. By abstracting the PHY/MAC layers with CUDA libraries, carriers can introduce frequency reallocations and new services simply through software updates. Weekly updates of generative AI models and radio control algorithms are becoming a reality, dramatically shortening the service differentiation cycle.

In terms of investment, semiconductor and server companies will be short-term beneficiaries, while telecommunications carriers will face an increase in initial CAPEX burden. However, because open API capabilities expand the potential to monetize third-party AI models, companies should first measure latency, power, and TCO through small-scale PoCs, establish an ROI calculation framework, and map out a roadmap. Partner selection that avoids vendor lock-in will determine competitiveness in the 6G commercialization phase.

—

Chapter 3: The Challenge of SoftBank and Nokia

SoftBank and Nokia, anticipating the 6G era, have elevated their joint trial phase both domestically and internationally. At the Tokoname site in Aichi Prefecture, which began in January 2026, they have fully cloudified the O-RU/EPC/AI orchestrator and are operating a hybrid of the Terahertz and Sub-6 GHz bands tobuild a multi-band AI-RAN. The Grace Blackwell servers deployed at the base station edge simultaneously execute Massive MIMO signal processing and AI traffic prediction models on a single unit, reducing power consumption during late-night hours by up to 38%.

International expansion is also accelerating. Nokia has expanded its anyRAN software and, in collaboration with Telefónica O2 in Germany, completed dual-connectivity tests for the 3.3 GHz and 26 GHz bands. SoftBank has ported AITRAS to the Changi Port area of Singtel in Singapore, demonstrating a round-trip link of less than 2 milliseconds for port AGVs. With this, the company has proposed a "6G Logistics Corridor connecting Asia-Pacific" concept, with plans to expand to six cities by 2028.

The benefits of an open ecosystem are appearing in both parts procurement and software updates. Nokia's vRAN containers comply with O-RAN specifications, allowing third-party AI models to be deployed immediately via a sandbox. Through its "AI-Market Place," SoftBank aims to provide network resilience models developed by startups on a monthly subscription basis, demonstrating its intent to simultaneously reduce communication facility investment and create new revenue streams.

Actual measurement data supports these expectations. At a smart factory in the Nagoya waterfront area, control latency for 20 collaborative robots wasreduced by 70%compared to 5G. In addition, by using AI to predict traffic loads during congestion and dynamically reallocate bandwidth, theprocessing efficiency per frequencyimproved by 2.1 times. Based on these results, Nokia has announced that it will commercially provide a "self-healing RAN" function that separates the O-RU and GPU edge in 2027.

There are also significant movements on the financial front. SoftBank plans250 billion yen in 6G-related CAPEXfor fiscal years 2025-2027, with 30% allocated to AI-RAN equipment. It estimates that investment recovery will be possible within three years through an AI traffic billing model. Meanwhile, Nokia aims for over 3 billion euros in revenue from the AI-RAN business alone by 2029 and has already signed PoC contracts with over 20 companies.

The challenges are security and governance. With the shift to AI-native systems, the scope of responsibility for model suppliers tends to become ambiguous. Both companies are jointly formulating aRoot of Trust mechanismto perform signatures and audits of AI models, and will propose it to international standardization forums by 2027. This has paved the way for achieving both observability and compliance even in multi-vendor environments.

Overall, the challenge undertaken by SoftBank and Nokia embodies the essence of 6G, which isOpen x AI. Domestic carriers and businesses in manufacturing, logistics, and public infrastructure will likely find that using this movement as a benchmark to expand PoCs and working on the three-axis optimization oflatency, power, and ROIacross the entire supply chain will determine their competitiveness for the next decade.

—

Chapter 4: Use Cases by Industry

In the manufacturing industry,second-level control of collaborative robotsis becoming a reality. FANUC introduced 6G-compatible AI-RAN edge servers at its Yamanashi plant, and as a result of controlling processing machines and robot arms on the same GPU infrastructure, it reduced tact time by an average of 12%. At Toyota Motor's Motomachi plant, they are also testing a system that can instantly reflect AI models being developed in Woven City from the cloud, moving toward operations where algorithms are updated without taking production lines offline.If real-time optimizationbecomes the norm, it is highly likely to fundamentally change the cost structure of high-mix, low-volume production.

In the field of autonomous driving, theV2N (Vehicle-to-Network) cooperative controlwhere vehicles and base stations share information at the millisecond level will accelerate. In January 2026, Tesla successfully demonstrated swarm control of 10 FSD vehicles using a 6G-compatible backhaul link at its Nevada test site. Domestically, when Honda operates Level 4 taxis at its Wako Research Center in Saitama Prefecture, it is conducting trials using KDDI's AI-RAN edge, where AI anticipates dangerous scenarios and proposes lane changes.By having base station-side AI complement vehicle sensors, it is reported that blind spots for cameras and risks from bad weather can be significantly reduced.

In the smart city domain, areal-time monitoring networkcombining disaster-response drones and high-definition cameras will expand. KDDI has built a system to remotely monitor landslides in mountainous areas of Saijo City, Ehime Prefecture, using Skydio drones. Through a 6G link, 4K video is transmitted to an integrated command center with almost zero latency, and AI immediately determines the possibility of a collapse, reducing the time to issue evacuation instructions to the site by an average of 3 minutes. Sweden's Axis Communications has begun providing a platform that combines 360-degree cameras and GPU edges to perform real-time detection of suspicious objects in urban areas.The combination of ultra-low latency and AI inferenceis demonstrating the ability to achieve both urban security and maintenance cost reduction.

What these cases have in common is thatAI models continue to evolve on the network side. Computing, which previously leaned toward either the terminal or the cloud, is optimized by AI-RAN while moving back and forth between the edge and the cloud. Companies need to first measure latency and power in small-scale PoCs and build a framework toevaluate ROI and TCO simultaneously. If 6G adoption spreads in manufacturing, mobility, and public infrastructure sites, the time axis from data to action will be dramatically shortened, and the competitive advantage of the entire supply chain will be rewritten.

—

Chapter 5: Stock Price Impact by Sector

Semiconductors are the quickest to react. The reason is clear: AI RAN increases computational needs as base stations become software-defined, pushing demand for GPUs, high-speed networks, and HBM to the forefront.NVIDIA has declared its commitment to AI-native 6G, clearly signaling its intent to expand its AI platform into telecommunications. When this becomes a catalyst, it is easy for the market to associate it with peripheral players like AMD, manufacturing partner TSMC, lithography equipment maker ASML, and HBM providers SK hynix and Micron. Investors are looking not at immediate revenue contributions, but at the story of how AI computing entering the massive telecommunications market will broaden the base of AI infrastructure.

Telecommunications carriers are prone to divided evaluations. SoftBank’s moves to prioritize AI RAN and ecosystem expansion create growth expectations, but 6G investment is also viewed as a heavy CAPEX burden. In the short term, stock prices are easily discouraged by the outlook for increased capital expenditure and depreciation, and even if they rise, it depends on the freshness of the news. Conversely, if a path to new revenue through AI operations is demonstrated, it could be valued as a departure from the simple telecommunications fee model. The point is not the investment amount, but whether a design for return on investment can be articulated.

Network equipment is likely to be re-evaluated as AI RAN implementation progresses. Nokia has announced anyRAN functional testing on the NVIDIA platform and integration with multiple companies, leaving an impression that it has entered the implementation phase. How Ericsson, Samsung Networks, and domestic players with a presence in vRAN like Fujitsu get involved will clarify the competition for orders. If open configurations advance, it is less likely to be a winner-takes-all scenario and more likely that market segmentation by use case or region will occur, causing short-term interest to rotate based on how news is released.

Cloud and data centers will see the impact of increased AI traffic as final demand. While AI RAN increases edge-side computing, model updates, training, and wide-area data integration ultimately return to the cloud and data centers. Therefore, related stocks like colocation provider Equinix and data center REIT Digital Realty are likely to retain medium-term demand expectations. The focus is on power and location; the quality of investment plans will be questioned in regions with severe power constraints.

Industrial automation is a medium- to long-term play rather than a short-term one. As talk of 6G and AI robot operations increases, companies like FANUC, Keyence, and overseas firms like Siemens, Rockwell Automation, and ABB come to mind. However, immediate stock price reactions will be swayed more by on-site upgrade cycles and economic waves than by a sudden surge in robot adoption. What needs to be determined is which industry will begin equipment upgrades after communications are established; if logistics and factory labor-saving efforts lead the way, the order of related stocks will change accordingly.

Ultimately, this theme is less about the entire sector moving in unison every time news breaks, and more about a structure where semiconductors and infrastructure price it in first, with carriers and industry following later. From an investor's perspective, the turning point will be whether one can see continued verification and the specificity of contracts and implementations leading up to commercialization, rather than the heat of the announcements.

⸻

Please also take a look at our other NVIDIA-related articles.

If you found this helpful, a ♡ is encouraging. It raises the priority of a sequel.
💬 What do you think?
👉 Please share your thoughts in the comments!