Edge, IoT and private 5G gain as smart factories, industrial AI and sensing drive shift to AI infrastructure plays

Executive Summary

• Sentiment around edge computing and IoT platforms this week is moderately positive, driven by private 5G build‑out, AI infrastructure, and industrial AI initiatives, but with lingering uncertainty around capex-heavy mega‑projects.
• Capital flows continue into infrastructure and industrial use‑cases rather than consumer IoT, including funding for smart factories and sensing platforms, and strategic moves in 5G wholesale networks and industrial AI.
• Key risks remain regulatory and execution risk in telecom, long capex payback periods, and competitive pressure in private 5G and industrial platforms; for value investors, the risk is overpaying for long‑dated AI/edge narratives with limited current cash flow.
• Near‑term catalysts include telco ownership shifts enabling faster 5G/edge deployment, new industrial AI partnerships, and incremental signs that institutional capital is rotating into AI infrastructure, edge networks, and sensing rather than front‑end applications.

1. Key Value Signals

• Private 5G as the “plumbing” for edge and IoT

  • Ericsson is doubling down on its channel strategy for private 5G and explicitly linking it to “physical AI” and industrial use cases.
  • Malaysian telcos are taking full private control of Digital Nasional Berhad (DNB), the 5G wholesale operator, implying a stronger incentive to monetize 5G via edge and IoT services.
  • These moves may indicate a medium‑term revenue and margin uplift for vendors and operators that provide managed edge, private networks, and IoT platforms on top of 5G.

• Industrial AI and sensing as early monetization layers

  • Huawei is pushing enterprises to adopt industrial AI solutions, likely including edge‑deployed inference and IoT analytics.
  • Openreach is preparing its network to support acoustic sensing for water leakage, a concrete example of telecom infrastructure being monetized with IoT‑style “sensing as a service”.
  • These are recurring‑revenue use cases that could support higher ROE and FCF for the strongest platforms.

• Smart factory and manufacturing digitization as a demand driver

  • UK startup Isembard raised a $50m Series A to roll out 25 smart factories by 2026, strongly implying reliance on edge computing, IoT platforms, and industrial automation.
  • This reinforces demand for low‑latency compute, private 5G or advanced Wi‑Fi, and integrated IoT stacks; equipment and platform providers with established moats could see durable growth.

• Institutional pivot toward AI infrastructure

  • Fulloop highlights that institutional capital is increasingly targeting AI infrastructure rather than just applications.
  • In practice this often includes edge compute nodes, data transport (5G/fiber), and IoT data pipelines—areas where current listed incumbents may trade at more reasonable P/E and P/B than AI application darlings.

2. Stocks or Startups to Watch — With Value Lens

Public-market fundamental data below uses latest available trailing metrics as of early March 2026 and should be independently verified before use.

2.1 Ericsson (NASDAQ: ERIC; OMX: ERIC-B)

Rationale:
Core supplier of mobile infrastructure, now explicitly positioning private 5G as an enabler for “physical AI” and industrial edge solutions. If private 5G adoption accelerates, Ericsson participates via hardware, software, and potentially recurring network features.

This week’s signal:
Ericsson is “drawing the channel closer” for private 5G, emphasizing partner enablement and industrial use‑cases for “physical AI” at MWC and beyond, indicating continued strategic emphasis on enterprise and edge deployments rather than only public macro networks.
Source: Ericsson draws channel closer – as private 5G jumps on physical AI.

Key metrics (approximate):

• P/E: ~14–16x
• P/B: ~1.3–1.6x
• Debt-to-Equity: ~0.3–0.5
• Free Cash Flow: Positive; multi‑year pattern of solid FCF, though cyclical with operator capex cycles
• PEG: Around 1–1.5 based on modest mid‑single‑digit earnings growth expectations

Value angle:

• Trading more like a mature cyclical than an AI beneficiary despite potentially strong leverage to edge computing and private 5G.
• Reasonable P/B and moderate leverage combine with a plausible long‑term moat in radio and core networks.
• Key risk: pricing pressure from Huawei, Nokia, and open RAN alternatives; operator capex constraints.

2.2 Malaysian Telcos with DNB Exposure

(CelcomDigi Berhad, Maxis Berhad, YTL Power / YES, U Mobile – all on Bursa Malaysia; U Mobile remains private)

Rationale:
Telcos buying out the Ministry of Finance’s stake in Digital Nasional Berhad (DNB) align shareholder incentives toward monetizing 5G, including edge and IoT services on top of the wholesale network.

This week’s signal:
CelcomDigi and Maxis confirmed purchase of MOF Inc’s shares, moving DNB toward full private ownership. Each operator will have invested more than MYR 677.5 million in DNB after latest share purchases and prior injections. This occurs as U Mobile pursues a competing wholesale 5G network, introducing competition at the infrastructure layer.
Sources:

• DNB heads to private status after CelcomDigi and Maxis buy MOF shares
• Malaysian telcos take full control of DNB

Key metrics (indicative, CelcomDigi & Maxis):

• P/E: typically in the mid‑ to high‑teens
• P/B: ~3–5x (telecoms in Malaysia often trade rich to book but backed by high dividend yields)
• Debt-to-Equity: moderate to high (0.8–1.5) given spectrum and network capex
• Free Cash Flow: Positive; historically strong cash generators with high payout ratios
• PEG: likely around 1.5–2.5 due to slow earnings growth and mature markets

Value angle:

• These are more dividend-income plays than deep value, yet may see incremental upside if 5G‑enabled IoT and edge services generate new revenue streams.
• DNB’s privatization may reduce political uncertainty but introduces competitive risk via multiple wholesale networks.

2.3 Huawei Ecosystem (Private) – Industrial AI and Edge

Rationale:
While Huawei is private / state‑linked and not investable for most, its strategic moves shape the competitive environment for many public companies in industrial AI and edge.

This week’s signal:
Huawei is “driving enterprises to ACT on industrial AI,” with specific focus on AI for Industry 4.0, energy (Eskom mentioned), and enterprise AI solutions. Industrial AI typically relies on edge computing nodes, IoT sensors, and private networks.
Source: Huawei driving enterprises to ACT on industrial AI.

Financial metrics:

• Not publicly available in standard market terms (P/E, P/B, PEG unavailable).
• Known to invest heavily in R&D and maintain a broad global portfolio across telecom gear, cloud, and enterprise IT.

Strategic relevance:

• Huawei’s push will likely pressure margins and pricing for industrial AI and private 5G solutions worldwide, especially in emerging markets.
• Public companies competing with Huawei (Ericsson, Nokia, some industrial automation vendors) may be forced to differentiate on software, integration, and compliance rather than price alone.

2.4 Openreach / BT Group plc (LSE: BT.A) – Acoustic Sensing on Fiber Network

Rationale:
Openreach is part of BT Group and operates the UK’s fixed access network. Using its network for acoustic sensing to detect water leaks is a classic IoT/edge play that monetizes existing infrastructure with higher‑margin analytics services.

This week’s signal:
Openreach’s network will support acoustic sensing for leaky water pipes, leveraging existing fiber for distributed sensing and data backhaul. This moves Openreach from pure connectivity toward infrastructure‑enabled sensing and analytics.
Source: Openreach’s network to support acoustic sensing for leaky water pipes.

Key BT Group metrics (approximate):

• P/E: ~8–10x
• P/B: ~0.9–1.2x (often trades near or below book)
• Debt-to-Equity: elevated, often >1.5, reflecting network build‑out and pension obligations
• Free Cash Flow: Positive but constrained by fiber and 5G capex; improving as legacy capex rolls off
• PEG: Around 1–1.5 based on low single‑digit growth expectations

Value angle:

• BT often trades as a distressed utility‑like asset. Moves like acoustic sensing suggest incremental, relatively asset‑light revenue on top of sunk capex.
• If such IoT overlays scale (water leaks, traffic monitoring, structural health), they could modestly re‑rate BT’s growth profile and return on capital.

2.5 Isembard (Private, UK smart manufacturing startup)

Rationale:
A pure‑play on smart factories and industrial digitization which inherently requires edge computing, sensors, and automation platforms. Potential demand‑side signal for listed providers of industrial IoT and edge stacks.

This week’s signal:
Isembard raised $50m Series A funding and plans to open 25 smart factories by the end of 2026. Smart factories typically combine robotics, local (on‑prem) edge compute, IoT sensors, and often private 5G or advanced wireless networks for low‑latency control.
Source: UK manufacturing startup Isembard nets $50m Series A.

Financial details:

• Funding stage: Series A
• Last known valuation: Not disclosed in the article
• Revenue model: Likely a mix of manufacturing-as-a-service, smart‑factory build‑operate‑transfer projects, and possibly licensing of factory software / platforms. Exact mix not detailed.
• Strategic relevance:
  • Represents demand growth for vendors of edge computing (industrial PCs, gateways, small data centers), private 5G, industrial IoT platforms, and robotics.
  • For value investors, the key is not owning Isembard (private) but identifying equipment and software vendors that will supply similar factories globally.

Metrics:

• P/E, P/B, PEG, Debt-to-Equity, FCF: Not available; private company with limited disclosed financials.

2.6 Fulloop (Private – AI infrastructure investment platform / research)

Rationale:
Fulloop itself may not be directly investable yet, but its commentary highlights institutional capital flows toward AI infrastructure, which frequently includes edge computing, data centers, and connectivity.

This week’s signal:
Fulloop notes that institutional investors are increasingly targeting the infrastructure layer of AI—compute, storage, networking—rather than front‑end applications. While not explicitly detailed, the infrastructure stack dorsally includes edge nodes where latency, data sovereignty, and bandwidth constraints matter.
Source: Fulloop Highlights Institutional Shift Toward AI Infrastructure Investment.

Financial details:

• Funding stage / valuation: Not disclosed.
• Revenue model: Likely advisory, data platform, or investment management fees focused on AI infrastructure themes.
• Strategic relevance:
  • Reinforces a narrative in which edge computing and IoT platforms are seen as fundamental rails for AI, not just optional add‑ons.
  • Suggests that large pools of capital are scouting for less crowded, more infrastructural AI plays, which can help support valuations for under‑appreciated incumbents.

2.7 Broader AI and Grid Infrastructure – Google, Tesla, Oracle

Some articles this week touch on broader infrastructure themes that are adjacent to edge/IoT:

• Google and Tesla on grid management

  • Article discusses how Google and Tesla believe the grid is being managed “all wrong,” hinting at distributed energy resources, real‑time data, and edge control systems at the grid edge.
  • This indirectly validates the thesis that edge computing at substations, buildings, and vehicles will be critical.
  • Source: Google and Tesla think we’re managing the electrical grid all wrong.

• Oracle’s Stargate project

  • Discussion around whether Oracle’s massive $500 billion Stargate AI infrastructure project is in trouble reflects execution and financing risk in ultra‑large‑scale cloud and AI infrastructure.
  • While mostly a hyperscale data center story, some architectures could extend to regional and near‑edge nodes.
  • Source: Is Oracle’s Massive $500 Billion Stargate Project in Trouble?.

For a value investor, these megaproject narratives highlight the importance of discipline in capex and return on invested capital, and the risk of over‑promising growth built on expensive infrastructure.

3. What Smart Money Might Be Acting On

• Rotation from AI front‑end to infrastructure and industrial layers

  • Fulloop’s commentary points to institutional capital increasingly targeting AI infrastructure, which likely includes edge compute, telecom, and sensing infrastructure.
  • This may mean more patient capital entering telecom and industrial automation names that historically traded as low‑growth cyclical utilities.

• Private 5G as a key pillar of “physical AI”

  • Ericsson’s push and the DNB ownership shift indicate that enterprise 5G is still in the early stages of monetization.
  • Channel‑focused strategies, such as Ericsson’s, suggest a broad ecosystem of integrators and VARs will be critical. Smart money may be scanning for:
    • Telecom equipment vendors with software revenue and services.
    • System integrators specializing in private 5G + industrial IoT + edge compute.

• Infrastructure reutilization for incremental IoT revenue

  • Openreach’s move into acoustic sensing shows that existing fiber and telecom assets can be repurposed for high‑margin data services, without equivalent incremental capex.
  • Smart money might look at fixed‑line or tower infrastructure providers globally and ask which ones can add sensing or analytics layers.

• Industrial AI in emerging markets

  • Huawei’s industrial AI programs and references to partners like Eskom imply strong adoption in power, mining, and heavy industry, especially outside the US/EU.
  • Public companies supplying sensors, ruggedized edge compute, and industrial software to these regions could benefit indirectly from Huawei’s ecosystem build‑out, similar to how Android boosted app and chipset ecosystems.

4. Signals and Analysis (With Sources)

Ericsson – Private 5G and “Physical AI”

• What happened: Ericsson is expanding its channel strategy for private 5G and explicitly linking it to “physical AI” – AI applications that interact with the physical world via robots, sensors, and industrial systems.
• Why it matters:
  • Enterprise 5G and private networks are gateways to recurring software and services revenues (network slicing, security, analytics) beyond initial hardware sales.
  • These revenues can lift gross margins and partially de‑cyclicize earnings versus reliance on macro‑RAN deployments.
  • If adoption follows through, Ericsson’s current mid‑teens P/E may not fully reflect a higher‑quality revenue mix.
• Source: Ericsson draws channel closer – as private 5G jumps on physical AI.

Malaysian DNB Privatization – Aligning 5G Incentives

• What happened: CelcomDigi and Maxis bought additional shares from the Ministry of Finance’s holding in Digital Nasional Berhad, each committing >MYR 677.5m alongside previous capital. DNB is transitioning from a state‑dominated wholesale 5G operator to a private, telco‑controlled entity, while U Mobile develops a competing wholesale network.
• Why it matters:
  • Private sector control usually tightens focus on ROI and monetization, potentially accelerating roll‑out of 5G‑enabled services including IoT, smart city applications, and edge compute.
  • However, competing wholesale networks risk duplicative capex and margin pressure.
  • For CelcomDigi and Maxis, success means durable cash‑generating IoT revenues layered on stable consumer mobile income; failure means elongated payback periods on large 5G investments.
• Sources:
  • DNB heads to private status after CelcomDigi and Maxis buy MOF shares
  • Malaysian telcos take full control of DNB

Openreach – Acoustic Sensing for Water Leaks

• What happened: Openreach will enable acoustic sensing on its fiber network to detect leaking water pipes, providing a data service to utilities.
• Why it matters:
  • This is an archetypal edge/IoT platform: existing fiber acts as a sensor array, edge nodes process the data locally, and analytics enable predictive maintenance.
  • The incremental revenue is largely software and service margin on top of already‑deployed CAPEX, supporting better return on assets for BT Group.
  • If replicated for other verticals (rail, highways, structural monitoring), this could form a small but growing, higher‑margin segment within a generally low‑growth telco.
• Source: Openreach’s network to support acoustic sensing for leaky water pipes.

Huawei – Industrial AI and Edge

• What happened: Huawei is promoting an initiative to push enterprises to ACT (a branded framework) on industrial AI, targeting Industry 4.0, utilities (such as Eskom), and other heavy industries.
• Why it matters:
  • Industrial AI deployments in factories and utilities often require on‑prem or near‑edge compute, ruggedized IoT sensors, and reliable private networks.
  • Huawei’s push suggests strong demand in emerging markets, but also heightens competitive intensity in industrial AI and private 5G, pressuring peers to innovate and potentially compressing hardware margins.
  • Sovereignty and security concerns may steer some customers toward non‑Huawei alternatives, benefiting listed Western and Japanese vendors with credible industrial IoT stacks.
• Source: Huawei driving enterprises to ACT on industrial AI.

Isembard – Smart Factories as Edge and IoT Demand Drivers

• What happened: Isembard raised a $50m Series A and plans to open 25 smart factories by end‑2026.
• Why it matters:
  • Each factory is a node of intensive edge computing, robotics, sensors, and connectivity, representing multi‑layer demand for both hardware and software platforms.
  • This validates the business case for industrial IoT platforms and low‑latency compute at the factory edge.
  • For current public‑market names in industrial automation, this trend may sustain multi‑year order books with potential for high‑teens ROE and robust FCF.
• Source: UK manufacturing startup Isembard nets $50m Series A.

Fulloop – Institutional Shift Toward AI Infrastructure

• What happened: Fulloop highlighted an institutional capital shift toward AI infrastructure, including compute, storage, and enabling layers, instead of just AI applications.
• Why it matters:
  • Many AI infrastructure assets – from data centers to connectivity and edge platforms – are currently held by mature incumbents trading on reasonable multiples.
  • Incoming institutional capital often looks for scale, cash generation, and moats (e.g., spectrum licenses, long‑term customer relationships, geographic dominance).
  • This supports the thesis that telecommunications, data‑center REITs, and industrial automation players may see multiple expansion or at least more stable valuation support as they are re‑rated as AI infrastructure plays.
• Source: Fulloop Highlights Institutional Shift Toward AI Infrastructure Investment.

Google, Tesla, Oracle – Broader Infrastructure Context

• What happened:

  • Google and Tesla criticized current grid management approaches, hinting at a future of distributed, data‑rich grid operations where EVs, homes, and micro‑grids participate via edge‑connected systems.
  • Questions are being raised about the viability of Oracle’s enormous Stargate AI infrastructure project.

• Why it matters:

  • Grid modernisation is inherently an edge computing and IoT challenge: managing bi‑directional power flows, battery storage, and EV fleets.
  • Overly ambitious, capital‑intensive infrastructure projects like Stargate highlight the dangers of over‑levered bets on AI infrastructure without clear payback.
  • More disciplined infrastructure players with clear use cases and incremental ROI may offer better risk‑adjusted returns.

• Sources:

  • Google and Tesla think we’re managing the electrical grid all wrong
  • Is Oracle’s Massive $500 Billion Stargate Project in Trouble?

5. Investment Hypothesis

Overall Stance: Watch for Under‑appreciated Infrastructure and Industrial Plays

• The current environment around edge computing and IoT platforms suggests an emerging infrastructure upcycle underpinned by AI, smart factories, and grid modernization, but the cash flows are still uneven and not fully visible.
• Public equities most directly exposed (telecom equipment, telcos, industrial automation) generally trade at moderate P/E and P/B multiples, reflecting skepticism about growth and long‑dated capex paybacks.

Risk/Reward Assessment

• Upside potential:

  • If private 5G and industrial AI adoption follow through, companies like Ericsson, select telcos, and infrastructure‑rich incumbents (e.g., BT) may see:
    • Incremental high‑margin revenues from software and services layered on fixed infrastructure.
    • Better utilization of sunk capex, lifting ROIC and FCF.
    • Gradual multiple expansion from “ex‑growth telecom” to “critical AI infrastructure” narratives.

• Key risks:

  • Execution risk: Private 5G and IoT deployments can stall due to integration complexity, standards fragmentation, or lack of clear ROI for enterprise customers.
  • Competitive pressure: Players like Huawei, regional vendors, and open‑source or open‑RAN initiatives can compress hardware margins.
  • Regulatory and political risk: Particularly acute in telecom and cross‑border industrial infrastructure; spectrum and security rules can shift.
  • Capex burden: Over‑investment in large projects (e.g., Oracle’s Stargate) without matching demand could reduce shareholder returns for years.

Signals and Themes to Monitor

1. Concrete industrial wins rather than pilots

   • Signed multi‑year contracts for private 5G, smart factories, and industrial AI deployments.
   • Evidence of recurring SaaS or managed‑service revenue in segment disclosures.

2. Improving FCF and capital discipline

   • Telecom and infrastructure providers showing declining capex‑to‑sales ratios while still growing edge/IoT revenue.
   • Explicit capital allocation frameworks targeting ROIC above WACC.

3. Edge monetization of existing networks

   • Replication of Openreach‑style sensing services in other regions and verticals.
   • Telcos or neutral‑host providers launching standardized “sensing as a service” or “edge analytics” products.

4. Institutional re‑rating of infrastructure

   • More reports like Fulloop’s, but backed by measurable flows into telecom, industrial, and infrastructure funds that explicitly cite edge computing and IoT as a thesis.

Synthesis

• The week’s news collectively suggests that edge computing and IoT platforms are maturing from concept to infrastructure layer for AI and industrial digitization, but remain in the early innings of monetization.
• From a value perspective, the more compelling opportunities appear in under‑loved incumbents with real infrastructure and improving service overlays, rather than in richly valued pure‑play AI names.
• Maintaining a watch stance on key enablers such as Ericsson and select telcos, while tracking concrete evidence of recurring IoT/edge revenue and improved capital efficiency, may help identify when the market underestimation of their role in the AI stack begins to narrow.