IT Brief New Zealand - Technology news for CIOs & IT decision-makers
Story image
Internet of Things software market evolving – report
Fri, 9th Jun 2023

IoT enterprise spending reached $201 billion in 2022, significantly lower than many had previously predicted, according to new analysis from IoT Analytics.

IoT adopters will focus on building the IoT software backbone, developing IoT applications, and infusing AI in the coming five years. Some 47% of IoT applications are expected to have an AI element by 2027, while the five things vendors focus on as they evolve their IoT strategies include ecosystem, solutions; AI vision; acquisitions; and open standards.

IoT Analytics has published insights on the enterprise IoT market, including a new market evaluation, spending outlook, and a discussion of common IoT vendor strategies.

"The IoT is evolving. Software for example is becoming a more important part of the IoT stack, more specifically IoT applications," says Knud Lasse Lueth, CEO at IoT Analytics.

"AI is also starting to play a bigger role. We are observing that IoT vendors deploy 5 distinct strategies for "winning" in this changing IoT market. The partner ecosystem for example is becoming a crucial element in any IoT vendors' go-to-market and IoT delivery."

In 2017, BCG published a much-cited and discussed analysis, asking: Where are the growth opportunities in enterprise IoT? The authors discussed the growth opportunities along the IoT tech stack from 2015 to 2020. The conclusion: The top layers of the tech stack (services, applications, and analytics) will grow the fastest between 2015 and 2020 and will see the highest spending.

Today, six years later, we wonder: How accurate was the assessment by various industry analysts back then, and what does it take to win in a changing enterprise IoT market between 2023 and 2027?

Evaluating the IoT market vs. past predictions
According to IoT Analytics' January 2023 IoT market update, IoT enterprise spending reached $201 billion in 2022, up from below $100 billion in 2018. For comparison, this represents roughly 5% of the global IT market in 2022. 

While IoT Analytics had to lower the outlook for the forecast period of 2023-2027 in its January 2023 update, the market is still expected to grow by +19% annually and reach $483 billion in spending by 2027. The IoT market is expected to reach $483 billion by 2027.

The number of global active IoT connections grew by 18% in 2022 to 14.4 billion active IoT endpoints. In 2023, IoT Analytics expects the global number of connected IoT devices to grow another 16% to 16.7 billion. While 2023 growth is forecasted to be slightly lower than it was in 2022, IoT device connections are expected to continue to grow for many years to come.

How accurate were industry predictions in 2017?

IoT Analytics compared its market tracker that is based on actual revenues of 170+ companies in the IoT market to industry predictions of several third-party data providers that were cited by BCG in its 2017 article.

Its verdict: Industry predictions were directionally correct but too bullish on the pace of IoT adoption and on how fast different technology categories would shift.

Total market size Industry prediction for 2020: $250 billion 
IoT Analytics IoT spending tracker: $136.6 billion. According to estimates, the IoT market came out 45% lower than estimated. The market grew substantially in the 2015-2020 time frame but did not explode as had been projected by many (some segments, such as services, had been forecasted to grow 40%+ every year). 

Tech stack changes: Industry prediction for 2020: Services, applications, data analytics, and software platforms would substantially outgrow other parts of the tech stack. 
IoT Analytics IoT spending tracker: According to data, this did happen. However, data analytics, applications, and services captured 39% of the market in 2020 and not 60% as forecasted by some data providers. 

Segment changes: Industry prediction for 2020: Discrete manufacturing, utilities, and logistics would make up the largest share of the market. 
IoT Analytics IoT spending tracker: Data also show these three segments as being among the largest, but IoT Analytics' view is that discrete manufacturing is even bigger than predicted.

"Although we believe that general industry predictions (including our own predictions) were too bullish, we do see strong evidence that the market is growing further and now believe that the projected 2020 IoT market size of $250 billion will be surpassed at some point in 2024," says Lueth.

The IoT market in 2023 and beyond
IoT technology in 2023 adds value to organisations in all verticals, from manufacturing (e.g., factories) to retail (e.g., warehouses) and transport (e.g., cars). The IoT market has moved past headlines such as the infamous 3/4 of all IoT projects fail (Cisco 2017).

87% of all IoT projects meet or exceed expectations
In 2023, 87% of all IoT projects met or exceeded expectations, based on a 2023 survey of 300 IoT decision-makers that will be published in an upcoming IoT Analytics adoption report. Some companies have connected millions of connected IoT devices (e.g., Walmart, Tesla, and Hapag-Lloyd) and are looking to expand with more sophisticated software tools. 

Despite several key challenges remaining related to interoperability, skills and know-how, and chipset supply, companies do not question if they should do IoT but rather how it will be scaling from here.

IoT technology maturity framework
To understand how the IoT tech stack is changing and where the growth opportunities in IoT are going forward, one needs to consider a typical technology-focused maturity curve an IoT adopter goes through:

Stage 1: Enabling the asset
In the first stage, whether it is a smart washing machine, a heavy asset in a factory, or a ship at sea, companies need to invest in sensors and local controllers/gateways to be able to process IoT data. Despite a renewed edge computing investment cycle that is seeing many companies invest into more powerful and flexible hardware, many companies at this point have passed the first hurdle of ensuring they have basic IoT data to work with. We expect the spending for IoT hardware/devices to be the lowest growth category at 14% until 2027.

Stage 2: Establishing connectivity
In the second stage, enterprise end users establish and simplify the connectivity to their IoT hardware. While some technologies used have been around for decades (e.g., certain field buses), companies in recent years have invested heavily into higher bandwidth connectivity (like ethernet), wireless connections (e.g., 4G/5G and LPWAN), and more modern and lightweight protocols (e.g., OPC-UA and MQTT). Spending on connectivity is expected to grow by 18% until 2027.

Stage 3: Creating the software backbone
Data normalisation and analysis are key to the third stage of IoT maturity. Companies invest in the software backbone that allows them to access various IoT data sources and build valuable services, e.g., using cloud storage and platform services, centralized data lakes, containerisation, and modern databases. Many companies are currently in a major investment phase in this part of the maturity curve. That is why we expected spending related to IoT Platforms and middleware to grow 30% and 34% for Infrastructure as a Service (IaaS) until 2027.

Stage 4: Building value adding IoT applications
In the fourth stage of IoT maturity, IoT end users build cloud native or edge-based applications that make use of IoT data at scale. The ability to connect to any asset (stage 1) in a standardised fashion (stage 2) and having those data easily accessible (stage 3) enables a number of IoT use cases. Some of the early innovators (e.g., several automotive OEMs) have reached this stage and are building all kinds of internal (e.g., for their factories) and external (e.g., for their cars) IoT applications. We expect more companies to reach this stage of maturity in the coming years, which is why we expect the spending on IoT applications to exhibit a CAGR of 29% until 2027.

Stage 5: AIoT = infusing AI into IoT
Enabling business with AI is the fifth stage of IoT maturity. This is where companies explore ways to augment existing applications and build new applications by embedding AI. Machine vision and predictive maintenance are two of the most common AI-enabled IoT use cases today. Recent breakthroughs in generative AI may add a new dimension and are a driver for companies to rapidly adopt AI further.

The shift toward AIoT: Nearly half of all IoT applications to be AI-infused by 2027
Although some companies are still struggling to enable their assets (stage 1) and establish connectivity (stage 2), the future of IoT is the convergence of AI and IoT, also often referred to as AIoT.

The future of IoT is the convergence of AI and IoT 
In 2022, IoT Analytics conducted a survey of 500 senior IT and engineering staff at manufacturing companies. Fifteen percent of those companies indicated having fully implemented their AI strategy.

Two key use cases with an AI component that were cited were machine vision and predictive maintenance, both scoring a consistently high ROI for these companies.

"Based on the data we have and the discussions we lead with various organisations, we estimate that approximately 6% of IIoT applications today are AI-based, meaning that AI algorithms play a core role in enabling the use case (i.e., when using machine learning for vision detection in quality control)," says Lueth.

"Another 11% of applications are classified as AI-augmented, meaning that AI does play a role for the use case, but the use case is not dependent on that AI functionality (e.g., AI-augmented AGV/AMR routing instead of using rules decided by operators).

"We see a strong push toward embedding AI into IoT applications (e.g., by moving from condition monitoring to predictive maintenance) and expect a strong increase in the next years," he says.

"By 2027, we forecast nearly half of all IIoT applications to have some AI element with the share of AI-augmented applications tripling in that time frame (from 11% to 34%). 

"Many software vendors are currently undergoing large internal exercises on how they can augment their existing software offerings with the use of AI. In May 2023, for example, SAP announced 15 new business AI capabilities across the existing ERP product portfolio, including nine new generative AI scenarios."

Not everything is required to be AI-based, however. Many legacy non-AI IoT applications that are already deployed may not be touched. Many rule-based applications will remain sufficient for some clients and uses cases, and not every dashboard will need to be AI-augmented.

The hype around ChatGPT and Generative AI is likely to impact the AI strategy of companies in the coming years. It is still early to tell if and how quickly the implementation of generative AI in IoT will lead to new large-scale use cases. Generative AI models predominantly focus on text and images, with only a few models centred around sensor data at this point.

"We see a strong push now to bring AI into the business landscape, with the expectation that AI will reengineer enterprises as completely as enterprise software did three decades ago," Leuth says.

IoT continues to offer tremendous opportunities, even if the market is not advancing as fast as some had predicted in the past. IoT enterprise spending is expected to increase to $483 billion by 2027. In a maturing IoT market, the biggest growth opportunity lies in the software, especially with (AI) applications.

"The biggest IoT growth opportunity lies in software Nearly half of all applications in 2027 will be AI driven," says Leuth. 

"Leading IoT vendors, such as hyperscalers, industrial automation vendors, and connectivity providers, and also many start-ups position themselves to stay relevant in an AI-first IoT market," he says. 

"With a clear IoT strategy, companies have an opportunity to position themselves in what remains one of the biggest market opportunities of our generation."