From Experiment to Infrastructure
Apiro Data
Transformation · Fractional Leadership · AI, IoT, Operational Advisory for organisations serious about change.
Hello and welcome,
Is AI heading toward accountability, or are we witnessing expensive experimentation disguised as progress?
Atos just earned recognition as a leader in advanced analytics and AI services, while India prepares to double its data center capacity with a $6 billion investment. Meanwhile, the IoT market is expected to hit $1.3 trillion by year-end, and neuromorphic computing chips are moving from research labs to commercial launches. The technology sector is experiencing a pivot from "what can AI do?" toward "how do we deploy AI responsibly at scale while managing the infrastructure costs?"
"The convergence of Edge AI, 5G expansion, and post-quantum cryptography signals an industry transition from experimental pilots to production deployments where reliability, security, and economic viability determine success."
Edge AI Becomes Infrastructure Rather Than Innovation
The shift from cloud-based AI to Edge AI is accelerating faster than predicted, with 2026 marking the inflection point when manufacturers scale from early pilots to broad portfolio refreshes. Atos received recognition as a leader in the ISG Provider Lens Advanced Analytics and AI Services 2025 report across Europe and the United States, highlighting how AI integration is moving into regulated, complex sectors with customized frameworks that meet compliance needs.
What makes Edge AI different from traditional cloud processing is latency elimination and local decision-making capability. When controlling robotic arms or monitoring high-speed manufacturing equipment, waiting milliseconds for cloud responses becomes unacceptable. Industrial IoT represents the fastest-growing segment of connected devices, with IIoT expected to account for significant market growth by year-end. The IoT predictive maintenance market has grown from $1.5 billion to $6.5 billion since 2016 and is projected to reach $28 billion this year, with leading implementations demonstrating 25-30% maintenance cost reductions and 20-25% asset life extensions.
The practical applications are transforming operations across sectors. Nokia's Cognitive Digital Mining platform, featuring the Nokia Black Box with integrated Telit Cinterion modules and NVIDIA GPUs, enables real-time edge intelligence in mining operations. Telit Cinterion's deviceWISE Visual Intelligence leverages NVIDIA AI infrastructure including the Cosmos Reason vision language model for edge deployments. These aren't experimental demonstrations, they're production systems running in harsh industrial environments where reliability matters more than impressive benchmarks. The technology has matured from promising concept to operational infrastructure.
India Doubles Data Center Capacity with $6 Billion Push
India's data center capacity is expected to double by year-end with a $6 billion infrastructure investment, driven by explosive growth in 5G, AI, and IoT deployments. This expansion builds the core infrastructure for a digital economy designed to benefit everyone, with government and industry partnerships linking cloud, edge, and network capabilities. The India AI mission is helping the country skip older technology models and focus on shared computing that can support global AI needs.
The investment reflects India's hyperconnected infrastructure strategy, using edge computing for practical applications like smart factories, logistics hubs, automated retail, telecom networks, and smart cities. These depend on rising numbers of IoT devices generating enormous amounts of detailed data. Success hinges on how well companies combine computing power, connectivity, and smart technology into scalable AI systems that can work with existing setups while solving compatibility issues.
What distinguishes this from typical infrastructure buildouts is the emphasis on sovereign and hyperconnected digital ecosystems. Major infrastructure spending could cross $100 billion by 2027, signaling that India is pushing forward with purpose rather than just meeting demand. The cloud is no longer just storage, it has become the center for decision-making and innovation, supporting AI applications that transform personalized education, telemedicine, manufacturing, and public services. Organizations that succeed will be those seeing integration as strategy rather than task, building connected ecosystems where cloud, connectivity, and AI work together as parts of a single system.
IoT Market Approaches $1.3 Trillion with New Technology Drivers
The IoT market continues its expansion trajectory, projected to reach approximately $1.3 trillion by year-end with strong growth across industrial, consumer, and enterprise segments. Juniper Research released its Top 10 Emerging Tech Trends for 2026, identifying technologies like neuromorphic computing, physical AI, and post-quantum cryptography as poised to reshape how businesses plan and invest over the next few years.
The report positions 2026 as an important juncture for adoption and deployment. Post-quantum cryptography is shifting from theory to hybrid IoT solutions in the real world, with IoT vendors preparing for long-term changes in security standards and encryption. Neuromorphic computing commercial chipsets are designed to address AI bottlenecks, with several launches expected this year. Physical AI and humanoid robotics technology will advance over the next three years with improvements in autonomy and dexterity, particularly benefiting logistics and industrial applications.
What makes this moment significant is the move toward scalability and maturity in an industry becoming more resilient. The shift to long-term scaling and provision is particularly evident in businesses deploying IoT in infrastructure where reliability and security are mandatory. As Juniper Research senior analyst Molly Gatford notes, "Across security, compute, energy, and infrastructure, organizations are being forced to make real deployment decisions on technologies that were theoretical only a few years ago." Technologies once speculative have moved out of research status and are developing quickly, often faster than companies can implement them. The challenge facing enterprises will be how quickly they can adopt these technologies without increasing risk or complexity, and whether they have organizational readiness to do so at scale.
January 20, 1885
January 20, 1885 LaMarcus Adna Thompson received a patent for the Gravity Pleasure Switchback Railway, the first American-designed amusement roller coaster that would transform entertainment and engineering.
Thompson's Switchback Railway had opened at Coney Island in Brooklyn, New York the previous June. For just 5 cents, riders climbed to the top of a 50-foot platform and rode bench-like cars down a 600-foot track at 6 miles per hour, reaching another tower where the vehicle was switched to a return track for the trip back. The simple design proved wildly popular, earning $600 daily during peak season, an enormous sum in 1884 and sparking immediate competition from imitators.
Thompson didn't invent the roller coaster concept. The history traces back to at least the 17th century with Russian ice slides, where riders would slide down specially constructed wooden hills covered in ice at speeds approaching 50 mph. French entrepreneurs brought the idea to France in 1817 with Les Montagues Russes a Bellevilles, the first ride locking sled wheels into a track. What Thompson did was patent and commercialize the technology for mass entertainment, accumulating nearly 30 patents related to roller coaster technologies over his lifetime.
The patent filing on April 3, 1884 and approval on January 20, 1885 gave Thompson legal protection just as competitors flooded Coney Island with copies of his design. The Switchback's success launched the first golden age of roller coasters, with hundreds appearing across America by 1900. Thompson continued innovating, patenting designs that included dark tunnels with painted scenery to create "Scenic Railways" that became standard features in amusement parks nationwide. The Great Depression ended this first golden age, with roller coasters experiencing decline until 1972 when The Racer opened at Kings Island in Ohio, triggering a second golden age continuing today. Thompson's mechanical engineering breakthrough transformed entertainment, giving millions of people controlled thrills and establishing the roller coaster as the premier attraction at amusement parks for over 140 years.
Recommended by LinkedIn
Did you know?
The 5G and IoT Convergence Accelerates
Global forecasts predict nearly 5 billion 5G subscriptions worldwide by year-end, enabling new IoT use cases previously impossible due to latency, bandwidth, or reliability constraints. The expansion of private 5G networks has emerged as a trend providing enhanced control, security, and customization for business-critical industrial applications. 5G networks introduce five transformative capabilities that enable the next generation of IoT applications.
Ultra-low latency achieves response times under 1 millisecond, necessary for autonomous vehicles and robotics where tiny delays could cause accidents, at 100km/h, a 1ms delay means 2.7cm of uncontrolled travel. Massive device connectivity supports up to one million devices per square kilometer, transforming dense urban deployments. Enhanced mobile broadband delivers peak data rates exceeding 10 Gbps, supporting bandwidth-intensive applications like 4K video streaming and AR/VR experiences. Network slicing creates virtual networks optimized for specific use cases, allowing a single physical infrastructure to simultaneously support consumer smartphones, industrial sensors, and autonomous vehicles with different performance requirements.
Energy efficiency improvements of up to 90% compared to previous generations mean battery-powered IoT sensors can operate for years without replacement, making large-scale sensor deployments economically viable. Together, these five capabilities transform 5G from simply a faster phone network into foundational infrastructure for autonomous systems, smart cities, and industrial automation. The smart cities segment demonstrates particularly robust growth, with the global market projected to reach $312 billion this year at a compound annual growth rate of 19%. This expansion reflects increasing urbanization and the imperative for data-driven urban management.
The rollout of 5G New Radio will reach significant maturity this year, reaching its full potential for IoT applications. While initial 5G deployments focused on enhanced mobile broadband, the transformative capabilities for IoT lie in Ultra-Reliable Low-Latency Communication (URLLC) and Massive Machine-Type Communications (mMTC). URLLC promises latencies as low as 1 millisecond and reliability levels exceeding 99.999%, not an incremental improvement but a paradigm shift enabling mission-critical applications demanding instantaneous responses and unwavering dependability.
Think of remote surgery where a surgeon's hand movements must translate to robotic instruments with zero perceptible delay. Factory automation with collaborative robots requires split-second coordination to prevent accidents when humans and machines share workspace. Vehicle-to-everything communication for autonomous driving depends on cars, infrastructure, and pedestrian devices exchanging information faster than human reaction times. Smart grid control needs real-time load balancing across distributed energy sources to prevent blackouts. These applications were theoretically possible but practically impossible until URLLC matured.
The convergence is happening across specialized edge AI processors and IoT gateways becoming standard equipment, designed for energy efficiency and robust operation in harsh environments. These enable complex analytics and AI model inference directly at the source rather than round-tripping to the cloud. We're seeing Industrial IoT evolve from Industry 4.0 into Industry 5.0, emphasizing human-machine collaboration and sustainability alongside automation. Predictive maintenance has reached version 2.0, where AI and digital twins provide hyper-accurate predictions of machinery failures, enabling just-in-time maintenance and minimizing downtime with significant operational cost savings. Production lines are becoming reconfigurable on the fly, with IoT-enabled machines adapting to changing demands and producing customized products with mass-production efficiency.
That's all for now folks
We examined Edge AI transitioning from innovation to infrastructure, India doubling data center capacity with a $6 billion investment, and the IoT market approaching $1.3 trillion with new technology drivers emerging. Three infrastructure stories about technologies moving from experimental to operational at scale.
Till next time,
stay connected,
Iliana & the Apiro Data team.
Apiro Data / Iliana Hanewich This resonates! Edge inference solves latency, but the hard part is everything after the pilot: versioning, rollback, drift detection, and patching fleets in messy real-world environments. I’d love to see more discussion on the “boring infrastructure” layer: policy controls, observability, and operational readiness. That’s the difference between a clever demo and something you can trust in regulated, safety-critical workflows.