Cost, Security, and Infrastructure
Apiro Data
Transformation · Fractional Leadership · AI, IoT, Operational Advisory for organisations serious about change.
Hello and welcome,
What if the biggest AI story this month wasn't about intelligence, but about cost?
While TIME names the "Architects of AI" as 2025 Person of the Year, a different narrative is taking shape. Microsoft pours billions into Canadian data centers while OpenAI warns its next models could create "high" cybersecurity risks. Meanwhile, cellular IoT connections are forecast to hit 5.9 billion by 2035, and 5G RedCap devices finally moved from trials to commercial deployment.
The technology sector is confronting a reality check where capability matters less than deployment economics, security implications, and real-world infrastructure constraints.
"The convergence of AI risks, infrastructure investments, and connectivity standards shows an industry shifting from raw capability demonstrations toward addressing the practical challenges of scale, security, and sustainability."
AI Labs Acknowledge Cybersecurity Threats From Next Models
OpenAI publicly stated that its next generation of AI models could create "high" cybersecurity risks, including the potential to develop working zero-day exploits or assist in complex industrial intrusions. The company explained that future systems may craft remote exploits against well-defended targets and help plan operations aimed at real-world disruption. This candid admission marks a departure from the typical promotional rhetoric surrounding AI releases.
The company is responding with substantial investments in defensive applications such as automated code auditing and vulnerability patching, while implementing tighter access controls, monitoring systems, and infrastructure security measures. OpenAI's acknowledgment comes amid growing concerns about AI safety, particularly as models become more capable. TIME magazine's recognition of "Architects of AI" as 2025 Person of the Year highlighted both the transformative benefits and significant risks—from misinformation to deepening inequality, associated with AI's rapid advancement across industries.
What makes this development significant is the explicit recognition from a leading AI lab that their technology poses genuine threats rather than only offering benefits. Other companies have faced similar questions. Pharmaceutical research firm Inotiv confirmed a ransomware attack by the Qilin group that exposed data of more than 9,000 individuals, demonstrating how valuable research data attracts sophisticated attackers. As pharma and biotech firms increase AI use for drug discovery and clinical analytics, their data estates become more attractive ransomware targets. The industry faces a paradox where the same technology creating risks must also defend against them.
5G RedCap Finally Transitions From Trials to Deployment
After years of anticipation, 2025-2026 marks the shift of 5G RedCap (Reduced Capability) from experimental trials to early commercial deployments. Positioned between low-power LTE-M/NB-IoT and high-performance 5G NR, RedCap delivers mid-tier throughput with significantly better energy efficiency. The technology offers data rates in the "tens of Mbps" range, creating a practical middle tier in the cellular infrastructure for devices that outgrow LPWAN performance but don't require full 5G capabilities.
The adoption milestone came when Apple integrated RedCap technology into its latest Watch range, signaling the starting gun for broader deployment. Omdia forecasts cellular IoT connections will surge to 5.9 billion by 2035, with RedCap serving as a primary growth driver alongside 5G Massive IoT and 4G LTE Cat-1bis modules. Industrial systems show the strongest early traction, where machines, robots, and automated guided vehicles increasingly require moderate uplink performance, predictable latency, and long-term network reliability. RedCap also appears relevant for smart city infrastructure, high-resolution security sensors, wearables needing continuous telemetry, and capable fleet management systems.
The path forward involves practical challenges the ecosystem is now confronting. RedCap coverage remains uneven, forcing device makers to plan dual-mode architectures or maintain separate product variants for different markets. Module launches for eRedCap are expected to begin in 2026, further expanding mid-tier connectivity options. The automotive segment will drive substantial growth, rising from 500 million connections to around 1.2 billion over the next decade, increasing market share from 13% to 21%. Most of this connection growth will come from software-defined vehicles requiring reliable cellular connectivity for over-the-air updates, telematics, and safety features.
First Commercial 5G Satellite IoT Network Goes Live
Myriota launched HyperPulse, the first commercial 5G non-terrestrial network (NTN) for IoT, available December 15 in the United States, Mexico, Brazil, Australia, and Saudi Arabia. The network combines 5G NTN architecture with L-band capacity leased from Viasat, featuring a unique optimization layer that adjusts connectivity performance dynamically based on customer demand or environmental conditions. Europe, Southeast Asia, and additional Latin American countries follow in early 2026.
What distinguishes HyperPulse from traditional satellite IoT is its foundation on 3GPP 5G NTN standards, providing seamless interoperability with a growing number of chipsets and devices. The technology delivers lower latency and higher daily data allowances than older satellite systems, enabling applications requiring detailed reporting and rich sensing. Asset tracking for heavy equipment, containers, rail cars, and trailers becomes practical at global scale. Smart metering for utilities, environmental sensing for weather stations, and water quality monitoring gain coverage where terrestrial networks don't reach. Animal management including virtual fencing and remote monitoring works in locations previously impossible to connect.
Built on proven satellite infrastructure rather than experimental constellations, HyperPulse represents satellite IoT transitioning from niche applications to mainstream connectivity. Myriota already offers UltraLite service focused on extreme energy efficiency for the most power-constrained devices. HyperPulse complements this by serving applications where more bandwidth and lower latency justify slightly higher power consumption. The convergence of terrestrial 5G, satellite 5G NTN, and edge computing creates truly global IoT coverage where devices can roam between networks based on location and requirements. Companies no longer choose between terrestrial and satellite connectivity, they deploy solutions using both, switching automatically based on availability and cost.
December 16, 1947
December 16, 1947 American physicists John Bardeen, Walter Brattain, and William Shockley invented the first transistor at Bell Laboratories, launching the electronics revolution that would transform modern technology.
The team had been working on semiconductor materials as an alternative to vacuum tubes, which were bulky, generated excessive heat, and consumed significant power. Their breakthrough came when they successfully amplified an electrical signal using a germanium semiconductor with two gold contacts placed extremely close together. This point-contact transistor demonstrated that solid-state devices could perform the same functions as vacuum tubes while being smaller, more reliable, and more energy-efficient. The invention emerged from fundamental research into semiconductors rather than a specific application goal.
The transistor's impact was gradual at first. Bell Labs announced the invention publicly in June 1948, but early transistors were expensive and unreliable compared to mature vacuum tube technology. The first commercial transistor radio appeared in 1954, priced at $49.95, expensive for the era but demonstrating portability impossible with vacuum tubes. The technology improved rapidly as manufacturing processes advanced. Texas Instruments and other companies drove costs down while improving performance. Transistors enabled the miniaturization that led to integrated circuits in the 1960s, microprocessors in the 1970s, and eventually the billions of transistors found in modern smartphone processors.
Bardeen, Brattain, and Shockley received the 1956 Nobel Prize in Physics for the transistor, though their collaboration ended acrimoniously over credit disputes. Shockley left Bell Labs to found Shockley Semiconductor in Silicon Valley, which proved unsuccessful but seeded the region with talent that created companies like Fairchild Semiconductor and Intel. Bardeen became the only person to win the Nobel Prize in Physics twice, receiving it again in 1972 for his work on superconductivity. The transistor remains the fundamental building block of all digital electronics, with an estimated 13 sextillion transistors manufactured in 2025 alone, more than one million for every grain of sand on Earth.
Recommended by LinkedIn
Did you know?
Microsoft committed more than C$7.5 billion (about $5.4 billion) to expand AI infrastructure in Canada over the next two years, adding several new data centers in Quebec and other provinces. This represents part of a broader trend where hyperscale cloud and AI leaders race to build compute capacity near major enterprise and government customers. The investments address surging demand for AI workloads while also tackling data residency and sovereignty concerns that matter to governments and regulated industries.
BlackRock's chief Asia-Pacific strategist Ben Powell argues that AI infrastructure spending shows "no signs of slowing," framing the build-out as a long-duration capital cycle rather than a temporary bubble. He suggests many countries and corporations remain in early innings despite investor nervousness about whether AI-linked stocks and capital expenditure plans have overshot reality. Infrastructure investors and sovereign capital are doubling down on AI as a long-term asset class, viewing power, land, and grid capacity as strategic choke points for future competitiveness.
The counterargument comes from Oracle's recent earnings miss, which sent fresh warnings across the AI infrastructure trade. Markets are questioning whether massive data center investment clears the hurdle of believable economics. The tension reflects uncertainty about when, and whether, AI deployments will generate revenue that justifies the extraordinary capital commitments. OpenAI reports ChatGPT Enterprise usage surged eightfold with workers saving 40-60 minutes daily, providing evidence of productivity gains. Yet skeptics note that AI labs are incentivized to report optimistic metrics. The resolution will likely take years, with real adoption curves and revenue growth determining which view proves correct. What's clear is that billions in spending commitments are already locked in, making this one of the largest technology infrastructure build-outs in history.
The 5G infrastructure market is entering dramatic expansion, with total market size expected to surge from $12.25 billion in 2024 to $161.61 billion by 2034, representing a 29.5% compound annual growth rate. This encompasses the hardware, software, and network systems required to deploy next-generation wireless connectivity, including base stations, antennas, backhaul systems, macro and small-cell architectures, core network upgrades, and edge computing integration.
Several factors fuel this surge. The proliferation of IoT devices from smart home systems to industrial sensors and connected vehicles creates massive demand for reliable, high-capacity networks. As connected device numbers climb, 5G's high throughput and low latency become essential to sustain growth. Enterprises increasingly adopt private 5G networks for IoT, automation, cloud services, and real-time data processing. Commercial and industrial sectors rank among the fastest-growing segments. Regionally, North America retains a leading share driven by strong telecom infrastructure investment, while Asia-Pacific emerges as a powerhouse fueled by rapid urbanization, government initiatives, and massive demand for mobile connectivity.
The implications for stakeholders are substantial. Telecom equipment vendors and RAN providers should expect rising orders for small-cell and macro-cell base stations, antennas, and associated hardware. Enterprise IT and solutions providers building private 5G networks for industrial IoT, smart factories, logistics, and smart city infrastructure will see growing demand. Strong growth rates and broad adoption make 5G infrastructure a compelling investment target, especially in regions with nascent but fast-growing telecom infrastructure. The technology is transforming sectors from smart cities with real-time traffic management to industrial automation with massive IoT sensor deployments, connected mobility with autonomous systems, and enhanced mobile experiences. The scale of investment reflects confidence that 5G will fundamentally reshape how industries operate and how people connect.
That's all for now folks
We examined AI labs openly acknowledging cybersecurity risks from next-gen models, 5G RedCap moving from trials to commercial deployment, and the first satellite 5G IoT network going live globally. Three infrastructure stories about technology confronting practical deployment realities.
Till next time,
stay connected,
Iliana & the Apiro Data team.
