Uncategorized
Analysis: Nvidia’s $2 billion investment in Marvell; NVLink Fusion ecosystem & RAN vendor silicon strategy
NVIDIA just announced a $2 billion investment in custom silicon developer Marvell Technology (NASDAQ:MRVL). This comes right on the heels of its $2 billion investments in Lumentum, Coherent, and $1 billion in Nokia.
- NVIDIA is also deepening its relationship with Marvell within its NVLink Fusion ecosystem. NVLink is NVIDIA’s proprietary scale-up networking system. Scale-up refers to connecting computing components within a rack rather than between racks.
- NVLink Fusion essentially allows customers to connect non-NVIDIA components to NVIDIA components within the same rack. Thus, customers can mix and match technologies from different vendors when they make a purchase. However, each platform does need to have at least one NVIDIA component.
- NVLink Fusion is in opposition to the UALink consortium, of which NVIDIA is not a member. Key NVIDIA competitors like Broadcom (NASDAQ:AVGO) and Advanced Micro Devices (NASDAQ:AMD) back UALink. Companies in this group have the same goal as NVIDIA does for NVLink Fusion: to allow customers to easily connect their devices together within racks.
- UALink’s goal is to reduce NVIDIA’s power by providing an alternative to NVLink Fusion. One of the key benefits to data center operators, the buyers of AI chips, is avoiding vendor lock-in. By being able to source components from a wide range of companies, there is greater competitive pressure, and thus more room to negotiate. Meanwhile, building AI infrastructure solely on NVLink grants NVIDIA massive bargaining power.
Photo Credit: Marvell Technology
Marvell has been a member of both NVLink and UALink, one of the few major chip companies that can make this claim. Now, NVIDIA is more formally recognizing Marvell’s place within NVLink, potentially expanding its ability to win customers. Meanwhile, Marvell strengthens its standing in the AI market. From Marvell’s perspective, the deal has significant benefits. Even though Marvell was already a part of NVLink Fusion, the company’s place within this ecosystem is now elevated. Not all companies in NVLink Fusion have received a multi-billion-dollar investment from NVIDIA or their own dedicated announcement.
These factors suggest that NVIDIA is particularly confident in Marvell’s solutions and that it will put in more effort to sell them to customers. NVIDIA now has 2 billion more reasons to do just that. This is particularly noteworthy, as MediaTek and Alchip Technologies are also in NVLink Fusion, and compete with Marvell in custom silicon.
In fact, Alchip has been the source of considerable volatility in Marvell shares over the recent past. This comes as some investors believed that the firm would siphon off much of the custom chip business that Marvell has built with Amazon. However, Marvell’s last earnings report helped to significantly quell these fears. Additionally, Marvell will add $2 billion to its balance sheet. That is very significant, as the company ended last quarter with cash and equivalents of just $2.64 billion, adding meaningful financial flexibility.
The announcement also outlines new products outside of custom silicon that NVIDIA and Marvell will collaborate on. This includes scale-up networking components, optical interconnect solutions, and silicon photonics. This comes just two months after Marvell completed its acquisition of Celestial AI, which it calls a “pioneer in optical interconnect technology for scale-up connectivity.”
Expanding the language of the partnership to include these very products suggests that NVLink Fusion could offer a significant pathway to grow this recently acquired business. Overall, between the $2 billion investment, the dedicated announcement, and the expanded partnership scope, Marvell now looks like the custom silicon provider of choice within NVLink Fusion.
NVIDIA is telling its customers that its technology and Marvell’s can integrate seamlessly. This could influence future customers who want to use both products to do so on NVLink. As a result, the firm could see revenue benefits by getting in on deals it otherwise might not have participated in. This logic extends to customers interested in Marvell’s optical interconnect and scale-up solutions.
……………………………………………………………………………………………………………………………………………………………………………………………………………………………..
RAN Silicon Strategy:
- Ericsson continues to prioritize its in-house custom ASICs, dismissing claims that the R&D required is unsustainable. Michael Begley, Ericsson’s Head of RAN Compute, noted at MWC that their 30-year legacy of iterative development creates a cost efficiency that a “blank sheet” competitor couldn’t match.. Despite this, Ericsson is diversifying through its Cloud RAN portfolio. Their vRAN software, currently optimized for Intel x86 architectures, is being architected for portability across AMD and Nvidia (Arm-based) platforms. While Ericsson frames this as addressing varied customer requirements, industry observers view it as a strategic hedge against the shifting hardware landscape.
- Conversely, Nokia is signaling a pivot away from proprietary hardware. Following a significant investment from Nvidia, Nokia’s leadership has articulated a shift toward general-purpose hardware and decoupled software models. Although Nokia remains contractually tied to Marvell for custom silicon through the mid-2030s, the potential to offload intensive Massive MIMO processing to Nvidia GPUs suggests a technical path toward phased-out reliance on traditional custom ASICs.
- Furthermore, Ericsson utilizing look-aside acceleration via custom ASICs while Nokia commits to in-line acceleration using SmartNICs for improved performance and TCO. Ericsson prioritizes in-house silicon for efficiency, whereas Nokia is shifting toward general-purpose hardware and Nvidia-backed GPU acceleration.
- Meanwhile, Samsung’s RAN silicon strategy has fundamentally inverted the traditional industry model: virtualized RAN (vRAN) is now its primary offering, with purpose-built custom silicon moved to the periphery. As of early 2026, Samsung is the global leader in vRAN deployments, surpassing 53,000 active sites. For vRAN, Samsung is aggressively diversifying its silicon partners beyond Intel to include AMD (x86) and NVIDIA (Arm-based Grace CPUs) to prevent hardware lock-in. The South Korean company is positioning its vRAN as the “AI-native” foundation for 5G-Advanced and 6G, aiming for fully autonomous “autopilot” networks by 2027 through its CognitiV NOS (Network Operations Suite).
- Samsung is leveraging its internal silicon foundry to develop 2nm Exynos chips and Silicon Photonics (“Dream Chip”), with a mass production target of 2028 to integrate optical transmission directly into AI accelerators and RAN modules. Its new Network in a Server (NIS) solution consolidates Core, vCU, vDU, and AI agents into a single 1U server, targeted at private 5G and enterprise edge use cases. While Samsung has pivoted heavily toward a vRAN-first strategy, it still maintains a portfolio of purpose-built (non-vRAN) equipment that utilizes custom ASICs. However, the role of these chips is shifting from a primary focus to a supporting role as the industry moves toward general-purpose hardware.
- Legacy and Specialized Support: Samsung continues to sell and support its traditional Baseband Units (BBUs), which are powered by proprietary silicon developed in partnership with companies like Marvell. These ASICs are still used for high-density, performance-critical deployments where standard CPUs are not yet chosen by the operator.
- Hardware Acceleration: In non-vRAN scenarios, custom ASICs handle the most computationally intensive Layer 1 (L1) tasks, such as beamforming for Massive MIMO and FEC (Forward Error Correction).
- Phased-Out Trajectory: Samsung executives have acknowledged that the era of proprietary hardware is likely nearing its end. Alok Shah, VP of Network Strategy at Samsung, has noted that while they still provide purpose-built BBUs, it is only a “matter of time” before vRAN becomes the universal standard.
……………………………………………………………………………………………………………………………………………………….
References:
https://www.lightreading.com/5g/nvidia-backed-marvell-pitches-one-chip-to-rule-the-ran
RAN Silicon Rethink- Part II; vRAN and General-Purpose Compute
RAN silicon rethink – from purpose built products & ASICs to general purpose processors or GPUs for vRAN & AI RAN
AI-RAN Reality Check: hype vs hesitation, shaky business case, no specific definition, no standards?
Ericsson goes with custom silicon (rather than Nvidia GPUs) for AI RAN
AT&T and Ericsson boost Cloud RAN performance with AI-native software running on Intel Xeon 6 SoC
Analysis: Rakuten Mobile and Intel partnership to embed AI directly into vRAN
Is the “far edge” a bridge to far to cross for AI inferencing? What about “Distributed AI Grids”?
Analysis: Edge AI and Qualcomm’s AI Program for Innovators 2026 – APAC for startups to lead in AI innovation
Will “AI at the Edge” transform telecom or be yet another telco monetization failure?
Will 2026 be the “Year of the AI Ontology” for telecoms?
Overview:
For the telecommunications industry, many pundits say 2026 will be the year of “AI Ontology [1.],” primarily because a standardized knowledge plane is now seen as the “ultimate driver” for reaching higher levels of network autonomy. Industry experts from companies like Telstra and Amdocs emphasize that for agentic AI to move from isolated pilots to enterprise-scale operations, it requires a structured, explainable, and typed world model—an ontology—to unify data across fragmented systems.
Note 1. An ontology in AI is a formal, machine-readable framework that defines the concepts, properties, and relationships within a specific domain to enable knowledge sharing, reasoning, and semantic understanding. It structures data into a network of “things” (classes) rather than just files, acting as a “Rosetta stone” that allows AI systems to understand context, infer conclusions, and act on data.
…………………………………………………………………………………………………
Several network providers are adopting a “standardized, ontology-driven knowledge plane” to enable agentic AI to operate across traditionally siloed network systems. This shift in 2026, is driven by the need for Level 4 and 5 network autonomy, where agents require a common language to reason about network states and business intents.
1. Mark Sanders, Telstra’s chief architect, talked about the emergence of a structured, explainable knowledge plane that removes silo barriers between agents, freeing them up to become the workhorses of network automation. “We think for the autonomous network to reach level four or five is going to require a standardized, ontology-driven approach on the knowledge plane,” said Sanders at a recent Ericsson conference, touting this approach as the ultimate driver in next-level autonomous networks.
2. For BT, agentic AI is already yielding tangible results in IT service desks, especially as organizations shift from assistance to execution, according to Girish Mahajan, senior leader for mobile AI data/automation. In particular, AI agents have reduced trouble ticket resolution times. “It has reduced the time of the manual effort, and it has also increased efficiency of the service desk,” he said. However, same autonomy that drives value also introduces unpredictability.
“The outcome of agentic AI is something unpredictable because it’s continuously adapting during execution,” he said, adding a call for better design principles. “We need reflection-based architecture, and we need better AI/human collaboration. AI agents should learn from their actions and should work along with humans in their day-to-day.”
3. For Vodafone, work has revolved around lighthouse projects: small-scale efforts to demonstrate the value of a larger business use case.
“It’s quite a mundane use case around energy cost recovery. So obviously, energy is a huge operational expense for our industry,” said Simon Norton, digital/OSS engineering director, Vodafone Group. “It’s very complex, especially when you’re working in that multi-market environment, to manually compare line by line with energy bills against your own data sets.”
Vodafone’s AI agents, therefore, have been automatically ingesting bills and comparing them to identify any tariff anomalies.
“It’s mundane but actually super valuable,” said Norton, who stressed operators should find a project with a clear value proposition and get it out into production quickly. “You build the credibility, you start to get the funding into the system, and it buys you the time to work on that longer-term strategy.”
………………………………………………………………………………………..
- From Assistant to Doer: AI is evolving from a “helper” that provides insights to a “doer” that autonomously observes, decides, and executes actions within governed boundaries.
- Multi-Agent Orchestration: 2026 will see the rise of coordinated multi-agent ecosystems. These systems require an ontology to ensure that a “planner agent” can accurately break down goals for specialized “worker agents” without semantic confusion.
- Intent-Based Orchestration: To ensure network stability, telcos are adopting intent-based orchestration layers. These layers use ontologies to provide the deterministic, model-driven framework necessary to ground agent actions in real-world business intent.
- Network Autonomy: CSPs are aiming for TM Forum Level 3 or 4 autonomy by late 2026, using agents to turn intent into outcomes in live networks.
- Operational Leverage: Rather than massive headcount cuts, agentic AI is providing “operational leverage,” allowing teams to manage growing network complexity with the same workforce.
- Measurable ROI: Investments are focusing on high-impact areas like autonomous incident handling (30-40% cost reduction) and predictive maintenance (up to 40% fewer outages).
- Structured Knowledge Plane: Operators are shifting toward a standardized, ontology-driven knowledge plane to remove silo barriers between agents. This allows multiple specialized agents to collaborate on “broader, bigger outcomes” like root cause analysis across billing, CRM, and network systems.
- Enabling Agentic Autonomy: While 2025 focused on “agentic AI” as a buzzword, 2026 is about the foundational infrastructure—specifically graph-based data systems and digital twins—that gives agents the “executable semantics” they need to plan and act safely.
- Unified Truth for Agents: Without a central ontology, horizontal AI platforms often suffer from “agent drift,” where different agents interpret the same business logic (e.g., “unlimited plan”) differently, leading to billing and provisioning errors.
Ericsson’s View:
Hassan Iftikhar, Ericsson’s head of product domain data & analytics, called for better hyperscaler collaboration on scale, foundational cloud, and AI capabilities.
“The AI tooling, the security framework, we use those to industrialize and put agents into production… It’s pretty much an ecosystem that works together,” he said. At the panel, the data head revealed the vendor’s role in the agentic ecosystem through the use case of one operator needing help with catalog management, as well as scarce developer skills.
“They wanted to take the pain out of product configuration. So we designed a multi-agentic system where it basically helps product managers and marketers to configure and publish new instances through an actual language. So very complex catalog engineering, which can take weeks, is reduced to hours where you can search for reuse and launch.”
Iftikhar also revealed an OSS tool to help one operator’s engineers to diagnose and resolve issues within their operational instances – resulting in an agent that was seemingly too autonomous for the client.
“We put this use case together, basically taking an intent from an operations engineer, such as data diagnostics, and into it, we built the ability to take remediation actions automatically. What we sort of decided from that was a bit of a step too far to just throw that to an operations department for it to autonomously take steps. So we actually had to go in and build guardrails to limit that capability to a human oversight capability.”
“I think what we learned is that we have to sort of build that confidence in the team step by step before we can actually go to fully autonomous operation. Our learning from adjusting that use case was to be practical and adapt very quickly to what the business really needs.”
…………………………………………………………………………………
References:
https://www.sdxcentral.com/analysis/has-telco-already-faced-the-year-of-ai-agents/
The Financial Trap of Autonomous Networks: Scaling Agentic AI in the Telecom Core
Telecom operators investing in Agentic AI while Self Organizing Network AI market set for rapid growth
Nokia to showcase agentic AI network slicing; Ericsson partners with Ookla to measure 5G network slicing performance
T-Mobile US announces new broadband wireless and fiber targets, 5G-A with agentic AI and live voice call translation
Ericsson integrates Agentic AI into its NetCloud platform for self healing and autonomous 5G private networks
Agentic AI and the Future of Communications for Autonomous Vehicles (V2X)
AWS to deploy AI inference chips from Cerebras in its data centers; Anapurna Labs/Amazon in-house AI silicon products
Analysis and Impact of Blockbuster FCC ban on foreign made WiFi routers
On March 23rd, the Federal Communications Commission (FCC) updated its Covered List to prohibit the sale of foreign made consumer-grade (WiFi) routers to be sold in the U.S. The FCC’s Covered List is a list of communications equipment and services that are deemed to pose an unacceptable risk to the national security of the U.S. or the safety and security of U.S. persons. This FCC decision follows a determination by an Executive Branch interagency body, which concluded those devices pose unacceptable risks to U.S. national security and the safety of its citizens. . The new FCC restriction applies strictly to new foreign made router models, meaning retailers can continue marketing previously approved units and consumers can operate their existing equipment without interruption.
Impact:
TP-Link, Netgear, and Asus are currently among the top-selling Wi-Fi router brands in the U.S. consumer market. Estimates for early 2026 indicate that TP-Link alone holds approximately 35% of the U.S. consumer router market share, while Netgear and Asus collectively account for another 25%. The TP-Link Archer AXE75 is frequently rated the best router for most users due to its Wi-Fi 6E speed and reasonable price.
AXE5400 Tri-Band Gigabit Wi-Fi 6E Router
…………………………………………………………………………………………………………………………………
Linksys and Ubiquiti are American-based companies, but their hardware is produced by contract manufacturers overseas in locations like China, Vietnam, and Taiwan. Similarly, Amazon eero and Google Nest mesh routers are not made in the U.S.
–>Hence, these companies ability to sell new WiFi router models in the U.S. is now facing strict regulatory hurdles.
Quotes:
FCC Chairman Brendan Carr said: “I welcome this Executive Branch national security determination, and I am pleased that the FCC has now added foreign-produced routers, which were found to pose an unacceptable national security risk, to the FCC’s Covered List. “Following President Trump’s leadership, the FCC will continue to do our part in making sure that US cyberspace, critical infrastructure, and supply chains are safe and secure.”
Bogdan Botezatu, director of Threat Research at cybersecurity firm Bitdefender, says this ban is a step to harden the cybersecurity readiness of U.S. households, given ongoing geopolitical tensions. “Consumer routers sit at the edge of every home network, which makes them an attractive target and a strategic risk if compromised at scale,” he says. Asked whether he thinks the risk is real, Botezatu says the risk is real, though there’s no easy way to prove intent. “[Internet of Things] devices, including routers, are a weak point across the internet.”
“Virtually all (WiFi) routers are made outside the United States, including those produced by US-based companies like TP-Link, which manufactures its products in Vietnam,” a spokesperson from TP-Link tells WIRED. “It appears that the entire router industry will be impacted by the FCC’s announcement concerning new devices not previously authorized by the FCC.”
- Reduced Product Availability: New, high-performance routers manufactured outside the U.S. will not receive the necessary approval to be imported or sold, restricting future consumer choices.
- Higher Costs: The, “This ruling has the potential to significantly disrupt the U.S. consumer router market,” according to, likely resulting in increased prices for consumers as companies grapple with new regulatory requirements.
- Shift in Manufacturing: Router manufacturers, including those targeting the U.S. market, will likely need to shift production to the U.S. to satisfy security concerns and bypass the ban, says PC Magazine.
- Security Focus: The ban targets vulnerabilities in foreign hardware and firmware.
- No Impact on Existing Devices: Consumers can continue to use routers they currently own
References:
https://www.wired.com/story/us-government-foreign-made-router-ban-explained/
U.S. Weighs Ban on Chinese made TP-Link router and China Telecom
China backed Volt Typhoon has “pre-positioned” malware to disrupt U.S. critical infrastructure networks “on a scale greater than ever before”
WSJ: T-Mobile hacked by cyber-espionage group linked to Chinese Intelligence agency
Trump and FCC crack down on China telecoms; supply chain security at risk
Semtech LoRa® PHY technology enables Amazon Sidewalk to expand while supporting fixed and mobile IoT endpoints
Introduction:
Semtech Corporation, a leading provider of high-performance semiconductor, Internet of Things (IoT) systems and cloud connectivity service solutions, is the creator and primary owner of the intellectual property (IP) for LoRa® technology, providing the Physical layer chips (PHY transceivers) used in LoRaWAN – the very popular Low Power Wide Area Network (LPWAN) for IoT endpoints.
The Camarillo, CA based company last week announced that LoRa® technology will continue to serve as the core radio modulation for Amazon Sidewalk across all markets in this year’s Sidewalk international expansion. Sidewalk’s global expansion officially begins in Canada and Mexico with further expansion to other international regions is scheduled for later in 2026. The network is projected to expand to over 30 new countries by year’s end.
Amazon Sidewalk is increasingly viewed as a commercial success in terms of infrastructure deployment and technical capability, transitioning from a niche smart home feature to a broad, LoRa-based Low Power Wide Area Network (LPWAN). While it faced initial skepticism regarding privacy and adoption, the network now boasts massive, passive coverage of over 95% of the U.S. population and is undergoing rapid international expansion.
Architectural role of LoRa in Sidewalk:
LoRa is the de facto wireless platform of LPWANs for IoT. Semtech’s LoRa chipsets connect sensors to the Cloud and enable real-time communication of data and analytics that can be utilized to enhance efficiency and productivity. LoRa devices enable smart IoT applications that solve some of the biggest challenges facing our planet: energy management, natural resource reduction, pollution control, and infrastructure efficiency.
Amazon Sidewalk aggregates spectrum in unlicensed bands and combines multiple physical layers, with Semtech’s LoRa modulation providing the long‑range, low‑power tier for neighborhood‑scale coverage beyond home Wi‑Fi and short‑range Personal Area Networks (PANs). By using ONLY LoRa as the core wide‑area PHY, Sidewalk evolves from a home‑centric LAN into a geographically distributed WAN that can support both fixed and mobile IoT endpoints across dense residential environments.
Network scale and coverage:
Sidewalk already covers roughly 95% of the U.S. population, making it one of the largest license‑free, consumer‑facing LPWA deployments, and the 2026 roadmap extends the footprint into Canada and Mexico first, followed by additional international markets later in the year. This expansion effectively turns Sidewalk into a multi‑continent overlay network, leveraging existing consumer premises equipment and LoRa‑enabled endpoints to provide persistent connectivity without requiring dedicated operator‑grade RAN build‑outs.
Technology differentiation vs other LPWAN options:
NB-IoT (included in ITU-R M.2150 IMT 2020 RIT/SRIT standard) holds the largest LPWAN share at roughly 54%–58% of total LPWAN connections, due to massive adoption in China which accounts for approximately 84% of all global NB-IoT connections. Outside of China, LoRaWAN is the clear market leader with a 41% share of connections. As of late 2025, there are over 125 million LoRaWAN end devices deployed globally, growing at a 25% annual rate. It is the preferred choice for private IoT networks, specifically in smart buildings, agriculture, and industrial asset tracking.
LoRa’s combination of long range, ultra‑low power operation, and mature ecosystem (silicon, gateways, and cloud stacks) gives Sidewalk a differentiated profile relative to alternatives such as narrowband cellular IoT and other unlicensed LPWAN modulation methods. For Amazon, anchoring Sidewalk on LoRa reduces RF and protocol fragmentation on the end‑device side while preserving flexibility to layer higher‑level Sidewalk services and security on top of the underlying LoRa/LoRaWAN protocol stack.
Market and ecosystem context:
Amazon Sidewalk now sits alongside large industrial and enterprise LoRaWAN networks, reinforcing LoRa’s position as the leading low‑power wide‑area connectivity technology in unlicensed spectrum. The LoRaWAN IoT connectivity market is forecast to grow from about 10.7 billion USD in 2025 to 44.8 billion USD by 2030 (33.1% CAGR), while LoRaWAN deployments have surpassed 125 million devices globally with a 25% CAGR, signaling a robust runway for Sidewalk‑class Massive IoT use cases.
Implications for device and service design:
For device OEMs and service providers, Amazon’s decision effectively de‑risks LoRa as a long‑term connectivity bet for consumer and prosumer IoT, given Sidewalk’s trajectory to tens of millions of active devices worldwide. Vendors integrating LoRa‑based designs can now target both traditional LoRaWAN operator networks and the Sidewalk ecosystem, enabling common hardware platforms to support smart home, safety, environmental monitoring, and asset‑tracking applications at neighborhood and city scale.
LoRa Enables Sidewalk’s Technical Evolution:
Chirp spread spectrum (CSS) modulation in LoRa technology provides the technical foundation enabling Amazon Sidewalk’s new capabilities:
- Enhanced Network Density: LoRa multi-spreading factor capability optimizes longer range and shorter time-on-air, supporting higher device concentrations in urban environments while maintaining reliable connectivity.
- Location-Based Services: Unique location accuracy service that combines the power of Wi-Fi, Bluetooth Low Energy (BLE) and GPS enables a new class of location aware devices that don’t need expensive cellular solutions for asset tracking applications.
- Hub-Less Deployments: Utilized for both out-of-band-diagnostics as well as signaling radio for battery-powered cameras, LoRa lowers the need for hubs/repeaters, reducing infrastructure complexity for consumers while extending effective coverage areas.
Proven Heritage of LoRa in Massive IoT Networks:
Semtech’s LoRa technology has been deployed by more than 170 major mobile network operators globally, with over 500 million connected devices across smart cities, utilities, logistics, unmanned aircraft systems, and industrial applications. This proven deployment heritage provides the technical foundation and ecosystem maturity required for Amazon Sidewalk’s global expansion.
The technology’s long-range capability, extending connectivity up to several kilometers from Sidewalk bridge devices, combined with its ability to penetrate buildings and operate in dense urban environments makes it uniquely suited for neighborhood-scale networks. LoRa provides free, long-range connectivity that consumers can rely on for years of battery-powered operation.
Building on CES 2026 Momentum:
Ring showcased its expanded product portfolio using LoRa at CES 2026, introducing comprehensive sensor families for security, safety and home automation. These products join the growing network of devices powered on Sidewalk, including water leak and freeze detection sensors, wearable devices and environmental monitoring solutions, all leveraging the connectivity advantages of LoRa.
The Sidewalk network’s architecture—combining LoRa for long-range communication with Bluetooth Low Energy for device setup—creates a robust, resilient IoT infrastructure that can scale to support millions of devices while maintaining the ultra-low power consumption critical for battery-operated sensors and cameras.
…………………………………………………………………………………………………………………………………………………………………..
About Semtech:
Semtech Corporation (Nasdaq: SMTC) is a leading provider of high-performance semiconductor, IoT systems and cloud connectivity service solutions dedicated to delivering high-quality technology solutions that enable a smarter, more connected and sustainable planet. Our global teams are committed to empowering solution architects and application developers to develop breakthrough products for the infrastructure, industrial and consumer markets.
Amazon’s Sidewalk Low Power WAN is a dud, despite four announced partnerships
From LPWAN to Hybrid Networks: Satellite and NTN as Enablers of Enterprise IoT – Part 2
Enterprise IoT and the Transformation of UK Telecom Business Models – Part 1
Empowering Low-Power Wide-Area Networks to Meet the IoT Challenge
NTT’s IoT Services for Sustainability supported by new LoRaWAN low power network
IoT for Smart Cities: LoRa with Semtech Silicon as a leading LPWAN
LoRa Alliance and OMS Group integrate smart metering standards which may result in massive scale for IoT
LoRaWAN gains momentum: NEC LoRaWAN server + New Zealand nationwide network
LoRaWAN and Sigfox lead LPWANs; Interoperability via Compression
New LoRaWAN™ network for IoT in Saudi Arabia from Actility and machinestalk
LPWAN to Application standardization within the IETF
Taara Lightbridge Pro: an ultra reliable wireless optical communications system for 5G mobile backhaul
Google moonshot factory X graduate Taara [1.] is launching Lightbridge Pro, a wireless optical communications system designed to deliver 99.999% (“five nines”) carrier-grade uptime for 20 Gbps backhaul, addressing weather-related reliability issues in Free Space Optical Communication (FSOC).
Lightbridge Pro is designed for seamless integration into carrier-grade networks, including mobile backhaul and mission-critical infrastructure. By integrating intelligent optical switching directly into the hardware, it automatically reroutes traffic to fiber or RF backups during, for example, heavy fog or rain.
Note 1. Tara says that for the last eight years, they have been developing novel technology that uses beams of light to deliver high-speed, secure connectivity where fiber and wireless can’t – bringing abundant access to everyone, everywhere.
“As demand for data soars, existing connectivity solutions are reaching their limits. What if we could harness the power of light to deliver a better, faster, more efficient connection, without the need for cables?” Mahesh Krishnaswamy, Founder and CEO.
Key Features and Impact:
Carrier-Grade Reliability: Lightbridge Pro is purpose-built for high-availability requirements of 5G mobile backhaul,, and city-wide network providers.
Intelligent Switching: The system ensures seamless, near-instantaneous, switching between optical and backup connections (like RF) to maintain continuity.
Performance: It delivers up to 20 Gbps full-duplex capacity, bridging gaps where fiber installation is too costly or difficult.
Global Application: Already deployed in over 20 countries, the technology is used in dense urban, rural, and disaster recovery scenarios.
Operational Efficiency: The system includes comprehensive Fault, Configuration, Accounting, Performance, and Security (FCAPS) management, suitable for integration with existing Operations and Business Support Systems (OSS/BSS).
Deployment and Use Cases: Tara’s platform is aimed at large-scale network operators and mission-critical communications, particularly in dense urban environments or rough terrain where laying fiber is not economically viable.
Current Partners: Taara Lightbridge is already deployed in more than 20 countries, from dense urban cores to remote terrain to disaster recovery scenarios. Carriers already using Taara’s technology include Airtel, T-Mobile, SoftBank, Digicel, and Liquid Intelligent Technologies. T-Mobile previously deployed Taara units for high-capacity backhaul at Coachella and the Albuquerque Balloon Festival.
………………………………………………………………………………………………………………………………………………………………………..
Taara is showcasing these new solutions at Mobile World Congress (MWC) Barcelona 2026 where the start-up will also be announcing a new photonics-based wireless optical system designed for even greater density and scalability.
………………………………………………………………………………………………………………………………………………………………………..
Editor’s Analysis:
Taara’s Lightbridge Pro looks like a serious, carrier-minded evolution of free-space optics (FSO) for 5G backhaul, but its real value will hinge on how well the “five nines” claim holds up under diverse atmospheric and operational conditions in the field.
Risks and open questions:
-
SLA realism: “Five nines” across mixed optical/RF paths is a strong claim; operators will want multi-year availability data by climate region, plus clear modeling of residual outage during extreme events where both optical and RF paths can degrade.
-
Operational complexity: Even with integrated switching and FCAPS, adding a new transport technology introduces planning, monitoring, and skillset overhead versus staying on homogeneous fiber/microwave.
-
Regulatory and spectrum: Where RF is the backup, spectrum licensing, interference management, and coordination with existing microwave/E‑band layers will affect total cost and deployment speed, and those aspects are not detailed in the product material.
Overall assessment:
For 5G mobile backhaul, Lightbridge Pro is best viewed as a targeted tool for high-value, hard-to-fiber routes, and for rapid-capacity or temporary deployments, rather than a universal replacement for fiber or microwave. If Taara’s integrated protection switching performs as advertised at scale, it meaningfully advances FSO from “interesting niche” to a credible part of a multi-layer transport strategy for carriers and city-scale operators.
…………………………………………………………………………………………………………………………………………………………………………………………
References:
https://www.taaraconnect.com/post/introducing-lightbridge-pro#
https://www.taaraconnect.com/product/lightbridge-pro
https://www.taaraconnect.com/about
Google X spin-out Taara and Digicomm International partner to offer high speed wireless communications
Dell’Oro: Mobile Core Networks +15% in 2025; Ookla: Global Reality Check on 5G SA and 5G Advanced in 2026
Palo Alto Networks and Google Cloud expand partnership with advanced AI infrastructure and cloud security
Google’s Project Suncatcher: a moonshot project to power ML/AI compute from space
Google Cloud announces TalayLink subsea cable and new connectivity hubs in Thailand and Australia
India 5G subscribers top 400 Million with rapid adoption continuing without 5Gi
With over 400 million 5G subscribers, India now ranks #2 globally (China is #1 [1.]). What’s even more remarkable is the speed of adoption after 5G spectrum auctions were repeatedly delayed. Jyotiraditya Scindia, India’s union minister for communications and development of the North Eastern Region, said the country is “setting new global benchmarks in scale, speed and digital transformation.”
According to figures cited by the minister, the country’s 5G subscriber base now exceeds that of other major markets, including the United States with around 350 million users, the European Union with 200 million, and Japan with 190 million. China remains the global leader, with more than 1.2 billion 5G connections.
Note 1. China has over 1.2 billion 5G subscribers as of late 2025, representing over 60% of all mobile connections in the country, driven by massive infrastructure rollout and strong adoption across major operators like China Mobile, China Telecom, and China Unicom, making it the global leader in 5G penetration.
……………………………………………………………………………………………………………………………………………………………………………………………………..
Bharti Airtel and Reliance Jio Infocomm spearheaded the launch, becoming the first carriers to operationalize 5G networks and achieving significant subscriber acquisition, with each surpassing the 50 million user milestone within their initial year of service [1]. Vodafone Idea subsequently entered the 5G market with launches in specific cities during 2025.
State-owned telecom provider BSNL is projected to launch 5G services within the current year [1]. This deployment is slated to exclusively utilize India’s indigenously developed telecom technology stack, a collaborative effort involving the Centre for Development of Telematics (C-DOT), Tejas Networks, and Tata Consultancy Services (TCS).
As of the close of 2025, the national 5G infrastructure comprised 518,854 operational base stations, marking a substantial increase from approximately 464,990 recorded at the start of the year [1]. The Department of Telecommunications (DoT) reported the deployment of 4,112 new 5G base transceiver stations (BTS) in December 2025 alone, contributing to the year-end cumulative total
- Infrastructure Growth: Rapid BTS deployment, exceeding 4,100 new installations in December 2025 alone, demonstrates intense network densification, with coverage now reaching most districts.
- Vodafone Idea’s Entry: Vi, after initial delays, commenced its phased 5G service introduction in select cities during 2025.
- BSNL’s Indigenous Strategy: The state-owned operator is slated to launch 5G using India’s homegrown stack (C-DOT, Tejas, TCS), showcasing self-reliance in telecom technology.
- Market Dynamics: The rapid expansion aims to unlock enterprise and consumer use cases, positioning India as a significant global 5G player, despite ongoing discussions about monetization and infrastructure investment.
Technical & Deployment Highlights:
- Architecture: A mix of 5G SA (Jio) and 5G NSA (Airtel) is prevalent, with SA offering lower latency and true 5G capabilities.
- Spectrum: Operators utilize various bands, including sub-6 GHz (3.3 GHz, 26 GHz) for broad coverage and capacity.
- Deployment Pace: Driven by ministerial targets, operators installed BTS at an accelerated pace, focusing initially on high-revenue urban centers.
This author deeply regrets that the Telecommunications Standards Development Society India (TSDSI)’s 5Gi RIT specification, included as part of the ITU-R M.2150 IMT 2020 RIT/SRIT standard, was not implemented in India. On January 25, 2022, TSDSI told the Telecommunication Engineering Center (TEC) under the DoT not to proceed with the adoption of 5Gi as a national 5G standard. TSDSI added that it “does not intend to further update 5Gi specifications.”
References:
OpenSignal: real world 5G deployment in India, market status & what happened to 5Gi?
Nokia Executive: India to Have Fastest 5G Rollout in the World; 5Gi/LMLC Missing!
LightCounting & TÉRAL RESEARCH: India RAN market is buoyant with 5G rolling out at a fast pace
Nokia’s Bell Labs to use adapted 4G and 5G access technologies for Indian space missions
Reliance Jio in talks with Tesla to deploy private 5G network for the latter’s manufacturing plant in India
Communications Minister: India to be major telecom technology exporter in 3 years with its 4G/5G technology stack
India to set up 100 labs for developing 5G apps, business models and use-cases
Adani Group to launch private 5G network services in India this year
Adani Group planning to enter India’s long delayed 5G spectrum auction
At long last: India enters 5G era as carriers spend $ billions but don’t support 5Gi
5G in India to be launched in 2023; air traffic safety a concern; 5G for agricultural monitoring to be very useful
Bharti Airtel to launch 5G services in India this August; Reliance Jio to follow
5G Made in India: Bharti Airtel and Tata Group partner to implement 5G in India
India government wants “home-grown” 5G; BSNL and MTNL will emerge as healthy
5G in India dependent on fiber backhaul investments
Hindu businessline: Indian telcos deployed 33,000 5G base stations in 2022
Arm Holdings unveils “Physical AI” business unit to focus on robotics and automotive
- Cloud and AI: Focused on data center and AI infrastructure solutions.
- Edge: Encompassing mobile devices, personal computing, and related technologies.
- Physical AI: Integrating its automotive business with robotics initiatives.
- Market Opportunity: Acknowledged the significant growth potential in robotics, from industrial automation to humanoid robots, driven by AI advancements.
- Synergy with Automotive: Combined robotics and automotive within the unit due to shared technical needs, such as power efficiency, safety, and sensor technology.
- Strategic Reorganization: Positioned Physical AI as a third core business line, alongside Cloud & AI and Edge (mobile/PC), to better focus resources and expertise.
- Customer Demand: Responding to existing customers (like automakers and robotics firms such as Boston Dynamics) who are integrating more AI into physical devices.
- Enhancing Real-World Impact: Aims to deliver solutions that fundamentally improve labor, productivity, and potentially GDP, moving AI from data centers to physical interactions
2026 Consumer Electronics Show Preview: smartphones, AI in devices/appliances and advanced semiconductor chips
Marvell shrinking share of the RAN custom silicon market & acquisition of XConn Technologies for AI data center connectivity
Groq and Nvidia in non-exclusive AI Inference technology licensing agreement; top Groq execs joining Nvidia
The Internet of Things (IoT) explained along with ARM’s role in making it happen!
Sebastian Barros: Using telecom networks for weather sensing requires a strategic telco shift from connectivity providers to ecosystem orchestrators
Telecom networks for weather sensing can be facilitated by using existing microwave links and 4G/5G signals as virtual sensors, detecting changes in signal strength and timing caused by rain, humidity, and temperature, effectively turning vast infrastructure into a dense, real-time atmospheric monitoring system for improved forecasting and disaster alerts, notes Sebastian Barros on Substack. By analyzing signal attenuation, telecom networks create high-resolution weather maps, complementing traditional methods like radar. Yet very few network operators or vendors have attempted to use telecommunications infrastructure for dense atmospheric sensing. The data exists but is rarely activated, processed, or disclosed.
Global Navigation Satellite System (GNSS) [1.] based atmospheric estimation, rain attenuation on microwave links, and radio refractivity effects have been studied for more than 20 years. The physics is well understood and already embedded in network planning and synchronization systems. There are eight million radio sites span cities, roads, ports, factories, and borders. Every site has power, compute, backhaul, antennas, timing, and regulatory protection. Today, the network only provides connectivity. Integrated Sensing and Communication (ISAC) starts to change that. It repurposes radio waves for radar-like sensing, including presence detection, velocity, Doppler shift, and range.
……………………………………………………………………………………………………………………………………………………………………………
Note 1. Global Navigation Satellite System (GNSS) is the umbrella term for satellite constellations like the U.S.’s GPS, Russia’s GLONASS, the EU’s Galileo, and China’s BeiDou, which provide global positioning, navigation, and timing (PNT) services. GNSS receivers use signals from these orbiting satellites to calculate precise locations on Earth, offering increased accuracy and reliability compared to relying on a single system, enabling applications from smartphone navigation to autonomous vehicles and precision agriculture.
……………………………………………………………………………………………………………………………………………………………………………
- Receiver: Your end point receiving device (phone, car navigation, etc.) picks up signals from several of these satellites.
- Calculation: By measuring the time it takes for signals to arrive from at least four satellites, the receiver performs complex calculations (trilateration) to pinpoint your exact position.
Telecom networks can expose not just connectivity but structured awareness: object motion, crowd flow, anomalies, risks, and environmental state. In real time, across every continent. The infrastructure already exists. What’s missing is the architecture and the will to build the sensing layer on top of it. With dense, real-time sensing already in place, telecom can expose environmental intelligence through Open Gateway APIs, just as it exposes location or quality today. No new hardware. No new towers. Just activation, inference at the edge, and exposure.
- Data Siloing: The data produced by ISAC currently remains locked within the physical (PHY) and Media Access Control (MAC) layers of the network. It is primarily used for internal network optimization and is not exposed to external applications or platforms.
- Lack of Abstraction and APIs: There are no standard abstraction layers or Application Programming Interfaces (APIs) that would allow external systems (e.g., weather services, autonomous navigation systems, urban infrastructure management) to access and interpret the raw sensing data.
- Absence of Data Fusion Standards: There is no standard methodology to fuse the output from ISAC with data from other sensing modalities (e.g., vision, audio, thermal sensors). This prevents the creation of a comprehensive, multimodal sensing mesh.
- Missing Marketplace: The lack of standardized access and integration means there is no marketplace for this valuable data, which stifles innovation and collaboration across different industries that could benefit from real-world awareness information.
References:
https://sebastianbarros.substack.com/p/telecom-built-the-worlds-best-weather
https://sebastianbarros.substack.com/p/telco-network-as-a-sensor-is-a-huge
https://www.linkedin.com/feed/update/urn:li:activity:7413260481743769600/
https://www.euspa.europa.eu/eu-space-programme/galileo/what-gnss
2025 Year End Review: Integration of Telecom and ICT; What to Expect in 2026
Smart electromagnetic surfaces/RIS: an optimal low-cost design for integrated communications, sensing and powering
Deutsche Telekom: successful completion of the 6G-TakeOff project with “3D networks”
Smart electromagnetic surfaces/RIS: an optimal low-cost design for integrated communications, sensing and powering
Researchers at Xidian University in China have pioneered a smart electromagnetic surface that converts ambient electromagnetic waves into electrical power, marking a potential leap in stealth and wireless technologies. This meta-surface innovation merges advanced electromagnetic engineering with communication principles, enabling self-powered systems for demanding applications. The self-sustaining electronic system integrates wireless information transfer and energy harvesting and has the potential to upend the dynamics of electronic warfare.
The surface facilitates simultaneous energy harvesting and data transmission, drawing power from radar or environmental signals without traditional batteries. Xidian’s team highlights its role in “electromagnetic cooperative stealth,” where networked platforms collaboratively minimize radar cross-sections and sensor detectability. Prototypes demonstrate viability for real-time wave manipulation, building on metasurface designs that dynamically adjust phase and amplitude.
The researchers said this included investigating “electromagnetic cooperative stealth,” where multiple entities work together to reduce their visibility to radar and electromagnetic sensors. In electronic warfare, the technology flips the script on radar threats: stealth aircraft could harvest enemy beams for propulsion or comms, reducing logistical vulnerabilities. This cooperative approach extends to multi-asset formations, enhancing collective invisibility across spectra.
Early tests align with broader Reconfigurable Intelligent Surfaces (RIS), a two-dimensional reflecting surface. RIS advancements facilitate beam steering up to ±45° with low side lobes. According to a paper published in the IEEE Internet of Things magazine last year, RIS could also be used in anti-jamming technology, unmanned aerial vehicle communication and radio surveillance – all of which are difficult to do using older optimization tools.
Reconfigurable intelligent surfaces can also be configured to create intentional radio “dead zones” to mitigate interference and reduce the risk of eavesdropping, according to German electronics manufacturer Rohde & Schwarz. The European Space Agency has further highlighted RIS as a candidate technology for satellite-to-ground communications, where controllable reflection and redirection of signals could help route links around physical obstacles.
For telecommunications, the surface promises 6G breakthroughs like integrated sensing and powering satellites or base stations. China’s lead here could accelerate reconfigurable networks, improving coverage in non-line-of-sight scenarios. Ongoing refinements target complex interactions for higher precision. By including sensing, communication and power into one hardware platform, the device could allow for a range of advanced applications while reducing eavesdropping and interference.
“Ultimately, it is expected to have a broad impact on 6G communications, the Internet of Things, intelligent stealth and other related fields,” the team said in a paper published in the peer-reviewed journal National Science Review last month. Many scientists say that a key area for next-generation wireless communications will be the transmission channel.
Researchers from Fudan University, the University of Sydney and the Commonwealth Scientific and Industrial Research Organization note that, when combined with artificial intelligence (AI), this technology could significantly enhance the security of air-to-ground Internet of Things (IoT) links.
In their latest publication, the Xidian University team describes RIS as a “powerful solution” for future wireless networks, citing its low cost, high programmability and ease of deployment. However, for 6G systems, RIS must support both communication and sensing on a unified hardware platform by integrating data transmission and radar-like functionality to lower cost and optimize spectrum and hardware resource utilization.
Addressing this requirement will demand architectures that can jointly manipulate both scattered electromagnetic waves and actively radiated signals. The researchers propose that an electromagnetic all-in-one radiation–scattering RIS architecture could provide a viable path to meeting this dual-control challenge. “This achieves significant savings in physical space and cost while ensuring multifunctionality across diverse application scenarios,” the team said. The RIS system could also work in a receiver mode to harvest wireless energy to be used to power the meta-surface itself or charge other electronic devices, the paper added.
It could be used for line-of-sight wireless communication, where there is a direct, unobstructed path between a transmitter and receiver, as well as non-line-of-sight wireless communication, in which there is no direct visual link due to physical barriers like buildings.
The proposed RIS “stands out as the optimal low-cost design” for integrated communication and sensing. “In the future, this architecture could enable the development of environment-adaptive integrated sensing and communication systems, micro base stations and relay integrated systems, as well as self-powered sensing systems,” the team said.
………………………………………………………………………………………………………………………………………………………………………………………
References:
https://ieeexplore.ieee.org/document/10907868
Nokia sees new types of 6G connected devices facilitated by a “3 layer technology stack”
Electromagnetic Signal & Information Theory (ESIT): From Fundamentals to Standardization-Part I.
Electromagnetic Signal and Information Theory (ESIT): From Fundamentals to Standardization-Part II.
IMT Vision – Framework and overall objectives of the future development of IMT for 2030 and beyond
ITU-R WP5D: Studies on technical feasibility of IMT in bands above 100 GHz
Summary of ITU-R Workshop on “IMT for 2030 and beyond” (aka “6G”)
Excerpts of ITU-R preliminary draft new Report: FUTURE TECHNOLOGY TRENDS OF TERRESTRIAL IMT SYSTEMS TOWARDS 2030 AND BEYOND
Juniper Research: Global 6G Connections to be 290M in 1st 2 years of service, but network interference problem looms large
Groq and Nvidia in non-exclusive AI Inference technology licensing agreement; top Groq execs joining Nvidia
AI chip startup Groq [1.] today announced that it has entered into a non-exclusive licensing agreement with Nvidia for Groq’s AI inference technology [2.]. The agreement reflects a shared focus on expanding access to high-performance, low cost inference. As part of this agreement, Jonathan Ross, Groq’s Founder, Sunny Madra, Groq’s President, and other members of the Groq team will join Nvidia to help advance and scale the licensed technology. Groq will continue to operate as an independent company with Simon Edwards stepping into the role of Chief Executive Officer. GroqCloud will continue to operate without interruption. It remains to be seen how Groq’s new collaboration with Nvidia will effect its recent partnership with IBM.
Note 1. Founded in 2016, Groq specializes in what is known as inference, where artificial intelligence (AI) models that have already been trained respond to requests from users. While Nvidia dominates the market for training AI models (see Note 2.), it faces much more competition in inference, where traditional rivals such as Advanced Micro Devices have aimed to challenge it as well as startups such as Groq and Cerebras Systems.
Note 2. Training AI models (used by Nvidia GPUs) involves teaching a model to learn patterns from large amounts of data, while AI “inferencing” refers to using that trained model to generate outputs. Both processes demand massive computing power from AI chips.
…………………………………………………………………………………………………………………………………………………………………………………………………….
Groq has achieved a significant financial milestone, elevating its post-money valuation to $6.9 billion from $2.8 billion following a successful $750 million funding round in September. The company distinguishes itself in the competitive AI chip landscape by employing a unique architectural approach that does not rely on external high-bandwidth memory (HBM) chips. This design choice, leveraging on-chip static random-access memory (SRAM), mitigates the supply chain constraints currently impacting the global HBM market.
The LPU (Language Processing Unit) architecture, while enhancing inference speed for applications like chatbots, currently presents limitations regarding the maximum size of AI models that can be efficiently served. Groq’s primary competitor utilizing a similar architectural philosophy is Cerebras Systems, which has reportedly commenced preparations for an initial public offering (IPO) as early as next year. Both companies have strategically secured substantial contracts in the Middle Eastern market.
Nvidia’s investments in AI firms span the entire AI ecosystem, ranging from large language model developers such as OpenAI and xAI to “neoclouds” like Lambda and CoreWeave, which specialize in AI services and compete with its Big Tech customers. Nvidia has also invested in chipmakers Intel and Enfabrica. The company made a failed attempt around 2020 to acquire British chip architecture designer Arm Ltd. Nvidia’s wide-ranging investments — many of them in its own customers — have led to accusations that it’s involved in circular financing schemes reminiscent of the dot-com bubble. The company has vehemently denied those claims.
…………………………………………………………………………………………………………………………………………………………………..
This deal follows a familiar pattern in recent years where the world’s biggest technology firms pay large sums in deals with promising startups to take their technology and talent but stop short of formally acquiring the target.
- A great example of that was Meta which in June invested ~$14.3 billion in Scale AI for a 49% stake in the company. That move valued the startup at around $29 billion. As part of that deal, 28 year old Alexandr Wang resigned as CEO of Scale AI to become Meta’s first-ever Chief AI Officer. He will remain on Scale AI’s board of directors. Wang’s team at the new “superintelligence” lab is tasked with building advanced AI systems that can surpass human-level intelligence.
- In a similar but smaller scale deal, Microsoft agreed to pay AI startup Inflection about $650 million in cash in an unusual arrangement that would allow Microsoft to use Inflection’s models and hire most of the startup’s staff including its co-founders, a person familiar with the matter told Reuters. The high-profile AI startup’s models will be available on Microsoft’s Azure cloud service, the source said. Inflection is using the licensing fee to pay Greylock, Dragoneer and some other investors, the source added, saying the investors will get a return of 1.5 times what they invested.
Bernstein analyst Stacy Rasgon wrote in a note to clients on Wednesday after Groq’s announcement:
“The Nvidia-Groq deal appears strategic in nature for Nvidia as they leverage their increasingly powerful balance sheet to maintain dominance in key areas…..Antitrust would seem to be the primary risk here, though structuring the deal as a non-exclusive license may keep the fiction of competition alive (even as Groq’s leadership and, we would presume, technical talent move over to Nvidia). Nvidia CEO Jensen Huang’s “relationship with the Trump administration appears among the strongest of the key U.S. tech companies.”
…………………………………………………………………………………………………………………………………………………………………………………………………………….
Nvidia CEO Jensen Huang recently dedicated a significant portion of his 2025 GTC 2025 conference keynote speech to emphasize that Nvidia intends to maintain its market dominance as the AI sector increasingly transitions its focus from model training to inference workloads. Huang delivered two major GTC keynotes in 2025: The primary annual conference held in San Jose, California, in March 2025 and a second GTC in Washington, D.C., in October 2025. At these events, he emphasized the rise of “reasoning AI” and “agentic AI” as the drivers for an unprecedented 100x surge in demand for inference computing in just a couple of years. Huang announced that the new Blackwell system, designed as a “thinking machine” for reasoning, was in full production and optimized for both training and large-scale inference workloads.
Huang shared a vision of moving from traditional data centers to “AI factories“—ultra-high-performance computing environments designed specifically to generate intelligence at scale, positioning investment in Nvidia’s infrastructure as an economic necessity.
…………………………………………………………………………………………………………………………………………………………………………………………………………….
References:
https://www.arm.com/glossary/ai-inference
IBM and Groq Partner to Accelerate Enterprise AI Inference Capabilities
Sovereign AI infrastructure for telecom companies: implementation and challenges
AI wireless and fiber optic network technologies; IMT 2030 “native AI” concept
Custom AI Chips: Powering the next wave of Intelligent Computing
OpenAI and Broadcom in $10B deal to make custom AI chips
Reuters & Bloomberg: OpenAI to design “inference AI” chip with Broadcom and TSMC
Analysis: OpenAI and Deutsche Telekom launch multi-year AI collaboration
Expose: AI is more than a bubble; it’s a data center debt bomb
Will the wave of AI generated user-to/from-network traffic increase spectacularly as Cisco and Nokia predict?
