Artificial Intelligence (Global Market)

Main article: Artificial Intelligence

• Artificial Intelligence Market Trends

• Artificial Intelligence (Russian market)
• Artificial intelligence in the United States
• Artificial intelligence (Ukrainian market)
• Artificial Intelligence (China Market)
• Artificial intelligence in India
• Artificial Intelligence (Uzbekistan Market)

2024

Billions of dollars wasted: More than 80% of AI projects in the world fail

At the end of August 2024, RAND Corporation published a study according to which more than 80% of AI projects fail. This is a very high figure, twice the usual failure rate for non-AI technology startups.

According to the study, the main reason for the failure of AI startups is the disproportionate goals of key project participants. Management often has no idea what exactly AI can and should achieve, so field professionals simply don't have the resources and time they need to achieve their goals. And the engineers themselves are often distracted by the latest developments in the field of AI and implement them into their projects without thinking about what benefits they should bring. Scientists and engineers are ready to hunt for new AI technologies like magpies, focusing only on a brilliant description, but they should think about whether these innovations will really help solve the problem or only further confuse the situation.

More than 80% of AI projects fail

Other reasons for the failure of projects are highlighted, including the lack of properly prepared data sets, inadequate infrastructure and the inapplicability of AI to solve tasks. In addition, the researchers note that these problems are not limited to the technology sector: even in academia, it is noted that scientists often focus on describing AI instead of looking for a real application.

Industry experts also agree with these estimates: for example, Baidu CEO Robin Li Yanhong notes that too many large language models are being developed in China at once, which waste precious resources with minimal practical use. And if AI projects fail to meet expectations for several years, the entire industry could burst like a huge economic bubble.^[1]

Named the most important AI technologies in the world

Investment in artificial intelligence applications and systems continues to increase, and organizations around the world are actively implementing generative services. The focus is increasingly on risk management and data security. Analytical company Gartner on June 17, 2024 named the most important AI technologies on a global scale.

Autonomous systems

These platforms provide a level of adaptability, flexibility, and agility that cannot be achieved with traditional AI techniques alone. The flexibility of autonomous systems is especially important in situations where the operating environment is unpredictable and real-time monitoring and control is impractical.

Quantum AI

We are talking about the nascent field at the junction of quantum technologies and neural networks. Research in this area in the future can lead to the emergence of qualitatively new AI algorithms designed to work in quantum systems.

Multi-Agent Systems

AI platforms of this type consist of several independent agents, each of which is able to perceive the environment and perform certain functions. Agents can be AI models, programs, robots and other elements. Several agents are able to work towards a common goal beyond the capabilities of each of the agents individually.

Neuro-symbolic AI

This is a type of artificial intelligence that combines machine learning methods and symbolic systems, for example, knowledge graphs. This approach makes it possible to create more reliable and trustworthy AI models. Neurosymbolic AI eliminates limitations in traditional AI systems, such as incorrect inferences and failure to explain the steps that led to a specific outcome.

Composite AI

This type of system provides for a combination of several artificial intelligence methods for deeper interpretation of data and solving a wide range of business problems. The goal is to create AI solutions that require less data and energy to learn. Composite AI provides greater opportunities for those organizations that do not have access to large amounts of historical or labeled data.

General AI

We are talking about creating software with intelligence similar to human. Such systems will be able to self-learn and perform tasks for which they were not originally intended. General AI without human intervention will be able to solve various problems, including those that arise in humans. In general, the concept provides for the development of AI systems with autonomous self-control, a sufficient degree of self-awareness and the ability to master new skills.

Sovereign AI

This is an attempt by various states to independently develop and implement AI with less dependence on the commercial market. The approach embodies political and cultural differences to achieve sovereign goals. Sovereign AI aims to maximize the value of technology while reducing risks.

AI on the periphery

The concept involves the use of AI in IoT devices, gateways, edge servers, etc. AI algorithms can be integrated into mobile gadgets, vehicles, medical diagnostic equipment, streaming video analytics systems, etc.

Generative AI

These technologies have a significant impact on business operations. GeneAI is able to stimulate innovation, automate creative tasks and provide personalized interaction with customers. Many companies see Genia as a powerful tool for creating content and solving complex problems.^[2]

Putin: When humanity realizes the threat from AI, the world will negotiate at the interstate level

As in the case of nuclear weapons, when humanity realizes the threat from the unlimited and uncontrolled development of artificial intelligence, then there will be a period to agree at the interstate level on how to regulate this, President Vladimir Putin said on February 9 in an interview with American journalist Tucker Carlson^[3].

When there is an understanding that the unlimited and uncontrolled development of artificial intelligence, either genetics, or some other modern directions that cannot be stopped, all the same, these studies will be, just as it was impossible to hide what gunpowder is from humanity, and it is impossible to stop research in this or that area, these studies will still be. But when humanity feels a threat to itself, humanity as a whole, then, it seems to me, there will be a period to agree at the interstate level on how we will regulate this, "said Vladimir Putin.

Tucker Carlson's interview with Vladimir Putin lasted more than two hours

Humanity should think about what will happen to it in connection with the development of these latest research and technologies in genetics or in artificial intelligence, the president said. And, he believes, it is possible to predict approximately what will happen.

Therefore, when humanity felt a threat to its existence from nuclear weapons, all owners of nuclear weapons began to negotiate among themselves, because they understood that their careless use could lead to complete, total destruction, "Vladimir Putin explained.

He also noted that there are studies in the field of genetics that can create a superman, a special person: a man-warrior, a man-scientist, a man-athlete.

Now they say that in the United States , Elon Musk has already introduced a chip into the brain of a person. I think that it is impossible to stop Musk - he will still do what he sees fit. But we need to somehow negotiate with him, we need to look for some ways to convince him. It seems to me that he is a smart person, that is, I am sure that he is a smart person. We must somehow agree with him that this process needs to be canonized, subordinated to some rules, the president concluded.

2023

The global market for AI platforms soared by 44.4% over the year and reached $27.9 billion

At the end of 2023, the global market for artificial intelligence platforms reached $27.9 billion, which is a new record. This is 44.4% more than in the previous year, when costs were estimated at $19.3 billion. The industry is rapidly developing, which is due to the active introduction of AI in various fields. Market trends are addressed in the IDC study, which TAdviser reviewed in early November 2024.

The AI platform market shows no signs of slowing down. Rapid innovation in generative AI (Genia) is changing how companies develop and implement applications and how they use technology to rethink their business models for competitive advantage. IDC expects growth to continue to accelerate amid the emergence of unified platforms for predictive and generative AI, says Ritu Jyoti, group vice president and general manager of IDC.

The global market for AI platforms soared by 44.4% over the year

It is noted that AI platforms make it easier to develop and deploy models and all kinds of applications, such as intelligent assistants, that can imitate a person's cognitive abilities. The technology components of AI platforms include machine learning, deep learning, generative functions, natural language processing (NLP), text analytics, multimedia analytics, tagging, search, categorization, clustering, hypothesis generation, question answers, visualization, filtering, alerts, and navigation. The AI platform market consists of three main segments: AI lifecycle software, services and software for finding and gaining knowledge.

The study authors say AI has the potential to improve user experience, optimize advertising, personalize content and improve data analytics. These advances can drive growth in online activity and entrepreneurship, with positive economic consequences. On the other hand, AI carries significant risks. Organizations need to monitor patterns and anomalies, identify potential risks and fix problems before they become widespread. Although not all risks can be eliminated completely, their reduction can be achieved with the help of AI itself. There are some other difficulties.

Organizations have faced the harsh realities of implementing AI, including high costs, a shortage of highly skilled professionals, and the growing difficulty of integrating these systems into existing workflows. The excitement around AI has reached an all-time high, but the real test has been whether companies can overcome existing serious challenges and reap tangible long-term benefits from their investment in AI. Those who have failed to accept these challenges risk falling behind in an increasingly competitive environment, emphasizes Raghunandhan Kuppuswamy, IDC research manager.

In 2023 Microsoft , it became the leader in the AI platform market, which can be explained by the reliable and comprehensive proposals of the Redmond Corporation, as well as its extensive ecosystem of products and services. AI solutions are deeply integrated into Microsoft's popular software and cloud services, providing ample opportunities for both enterprises and developers alike. The top five players in the industry also include,, and Palantir OpenAI. Google Amazon Web Services (AWS)

IDC analysts believe that in the future, the CAGR in the market under consideration will be 40.6%. Thus, by 2028, costs on a global scale, according to the presented estimates, could reach $153 billion. IDC believes cloud deployments of AI platforms will grow at a faster rate - with a CAGR of 50.9%.^[4]

OpenAI and Microsoft dominate AI market

The global market for AI solutions for media and entertainment for the year grew to $16.4 billion

In 2023, the volume of the global market for solutions with support for artificial intelligence for media and entertainment reached $16.4 billion. For comparison, a year earlier, expenses in this area were estimated at $13.3 billion. Thus, growth was recorded at 23%. This is stated in a study by Market Research Future, the results of which were published in early August 2024.

Analysts cite several factors contributing to the rapid development of the industry. In particular, AI helps in recognizing fake publications and detecting plagiarism. Using machine learning algorithms and natural language processing, market participants are able to efficiently analyze huge amounts of information. In addition, AI makes it possible to generate various content, including high-quality images. AI-based predictive analytics enables you to meet marketing challenges. One of the most important trends in the field under consideration is the use of neural networks for voice and speech recognition.

From a geographical point of view, the largest region for the implementation of AI solutions for media and entertainment is North America, which in 2022 accounted for $2.7 billion. Europe is in second place, and the Asia-Pacific region closes the top three, where the highest growth rates are observed. Among the significant market players are Amazon Web Services (AWS), Gearhouse South Africa PTY, Gravity Media, IBM, Matchroom Sport, Production Resource Group, Synthesia, Valossa Labs, Veritone, Sportway AB, EVS Broadcast Equipment, etc.

Analysts believe that in the future the industry will develop rapidly. The CAGR (compound percentage CAGR) is expected to be 31.9%. As a result, by 2032, costs will increase to $198.4 billion.^[5]

The global market for equipment, software and services in the field of artificial intelligence for the year reached $9.54 billion

At the end of 2023, the volume of the global market for embedded solutions with support for artificial intelligence reached $9.54 billion. For comparison, a year earlier, the costs in this area were estimated at $8.3 billion. Thus, an increase of about 15% was recorded. Such data are reflected in the Market Research Future study, the results of which were published in mid-June 2024.

Analysts take into account costs in the field of hardware and software, as well as related services. The first group includes network devices, processors, memory modules, sensors and other electronic components. It is noted that in the total volume of the market in 2022, the largest revenue was brought by hardware products. This is followed by software and services.

It is said that the development of the global industry of intelligent embedded solutions is stimulated by integration with cloud platforms. Cloud computing, the study said, provides a number of benefits for AI. This is, in particular, the scalability and availability of various resources, depending on the needs. Different pre-trained models can be provided within the cloud to perform certain tasks. This allows developers to quickly integrate AI features into embedded devices without creating a model from scratch. In addition, tools for monitoring and monitoring the state of equipment are available through cloud platforms.

AI services easily integrate with applications for common tasks such as personalized recommendations, contact center upgrades, improved security and security, and increased customer engagement. Common cloud scenarios include image and video recognition, natural language processing (NLP), and predictive analytics. In general, the combination of embedded AI devices and clouds allows you to form a flexible infrastructure in which complex tasks and model training are carried out in data centers, while peripheral devices process information at the edge of the network in real time with low latency.

Geographically, the largest share in the total volume of equipment ON and services in the field of AI provides. In North America 2022, the region accounted for approximately $3.35 billion. North America has a developed ecosystem of technology corporations, startups and scientific organizations that contribute to the development of the industry. In particular, Silicon Valley serves as the center of advanced research and development in the field of artificial intelligence. In second place in terms of spending is, and the Asia-Pacific Europe region closes the top three, where the highest growth rates are recorded. This is primarily due to the spread of all kinds of gadgets, rapid urbanization and the development of Internet infrastructure. Among the leading players in the industry under consideration are named Microsoft Google IBM Siemens,,,,,,,,,,,, AWS Nvidia Intel Qualcomm STMicroelectronics Oracle Salesforce

NXP Lattice HPE, , Octonion и. Competition in the market is due to various factors, including prices, product quality, delivery times and the ability of manufacturers to offer customers individual solutions.

Market Research Future analysts believe that in the future, costs for embedded solutions with support for artificial intelligence will continue to increase rapidly. The CAGR (compound percentage CAGR) is expected to be 20.9%. As a result, by 2032 the volume of the industry will reach $43.44 billion.^[6]

AI startups around the world increased the volume of attracted investments by 9% - up to $50 billion

At the end of 2023, the volume of investments in startups in the field of artificial intelligence on a global scale reached almost $50 billion. This is about 9% more than the result for the previous year, when investments in such projects were estimated at $45.8 billion. Relevant data are provided in the Crunchbase study, the results of which were published on January 4, 2024.

The report states that global investment in startups in all segments in 2023 amounted to about $285 billion. For comparison: in 2022, the figure reached $462 billion. Thus, a fall of 38% on an annualized basis was recorded. And against this background, the growth of the AI segment is especially distinguished, which is associated both with the rapid development of this area as a whole, and with the boom in generative services. Analysts Crunchbase note that only, and OpenAI Anthropic Inflection AI together managed to attract about $18 billion during 2023.

For a while - especially in the middle of [2023] - it seemed like investors were spending nine-figure sums every week on a new and better AI start-up. Sometimes there were so many large-scale rounds that they were difficult to count, - emphasizes Crunchbase.

Analysts also note that many major rounds of funding for AI startups involved the world's leading IT corporations - Microsoft, Amazon, Google and Nvidia. Moreover, Nvidia is strengthening its position in the AI market thanks to its GPU-based accelerators, which were in short supply due to high demand from companies that create various AI services and teach large language models. It is noted that the largest corporations in Silicon Valley are displacing traditional technology investors in the AI segment.^[7]

AI startups in the world raised a record $27 billion over the year

At the end of 2023, startups in the field of artificial intelligence on a global scale attracted approximately $27 billion for development, which is a record result. The previous maximum at $11 billion was recorded in 2021. Such figures are given in a study by PitchBook, the results of which were published at the end of December 2023.

The report said that of the $27 billion, about two-thirds were provided by corporations, and Microsoft Google. In Amazon particular, Microsoft invested $10 billion OpenAI in the developer company, and neuronets ChatGPT also participated in a round of financing for the AI startup Inflection AI in the amount of $1.3 billion. In turn, Google and Amazon jointly allocated $6 billion to support a young firm Anthropic that specializes in developing large language models.

Thus, as noted, technology corporations significantly outstripped venture capital funds in terms of investment in AI startups in 2023. This is explained by several reasons. Young companies prefer to partner with large IT corporations, since they, in addition to financial assistance, can provide cloud infrastructure and access to the most powerful hardware components that are necessary for AI training. At the same time, venture capital funds are forced to cut costs, adapting to the difficult macroeconomic situation.

Even the world's leading venture capital funds, which manage tens of billions of dollars, cannot fight to preserve the independence of these AI companies. Against the backdrop of the rapid development of AI platforms, tech giants seem to have taken control of most startups with exceptional potential, says Patrick Murphy, partner and founder of Tapestry VC, a venture capital fund.[8]

Business is waiting for an explosive growth in the use of AI. Research

As part of the latest study by Comindware and PEX Network, significant penetration of artificial intelligence (AI) into various aspects of operational efficiency and digital business transformation was identified. The report for 2024 demonstrates that companies in various industries have begun to actively introduce AI into their business processes, rely on the widespread use of generative neural networks. Comindware announced this on December 19, 2023.

In particular, the PEX 2024 study, the Russian localization of which was carried out by the Comindware analytical group, showed that many companies are already at different stages of AI implementation.

There has been a trend of increased investment in artificial intelligence, reflecting its growing importance. In 2024, 35% of organizations plan to invest in AI, which is a significant indicator of its impact on operational efficiency and business transformation processes, "said Igor Prostokvashin, leading analyst at Comindware.

Prostokvashin clarified that when considering methodologies for optimizing operating costs, artificial intelligence occupies a significant share - 20% of organizations already use it. This confirms that AI is becoming a key tool in change management, business analysis, business process reengineering, and digital transformation.

According to the expert, the PEX report emphasizes that AI finds application in a wide range of tasks - from optimizing operational processes to improving customer interaction and analyzing large amounts of data.

We see that organizations are at different stages of using AI. Some are just discussing its potential application, others are developing pilot projects or have already launched one or more successful projects. It is interesting that there are also companies that systematically apply AI in their processes of change, - said Igor Prostokvashin.

The 2024 Comindware and PEX Network report clearly demonstrates that AI is not just the technology of the future, but is already actively shaping the modern business landscape. Organizations that integrate AI into their operations are opening up new horizons of opportunity and efficiency. The AI market, which was valued at $136.6 billion in 2022, is showing significant growth and expansion in practical applications of the technology, including content creation and autonomous cars, according to the study authors' findings.

Artificial intelligence is actively used in various areas of organizations, including production activities, customer support services and call centers, as well as in data processing and management. This confirms the multifunctionality and adaptability of AI to various aspects of the business, analysts say.

It is noteworthy that a significant demand has already formed in Russian business for the use of AI to automate processes. Additional data and conclusions of analysts regarding the Russian market have already been published in the localized PEX report on the Comindware website.

How artificial intelligence is changing business. 10 trends

Artificial intelligence is changing the IT industry and business operations practices. Worldwide, there has been an explosive increase in interest in generative AI (GenAI) technologies, which allow the creation of text, images and diverse content based on data used to train models. In 2027, spending on various AI solutions is expected to exceed $500 billion, according to an IDC study published on October 26, 2023. Analysts call 10 trends that will change the global business ecosystem.

1. Transforming the IT Industry

IDC expects that the rapid shift in IT spending towards artificial intelligence will affect virtually all industries and applications. By 2025, the world's 2,000 largest companies (G2000) will direct more than 40% of their core IT spending to AI-related initiatives.

2. Tipping Point in the IT Industry

The IT industry will feel the influence of AI more than any other area, as almost every company seeks to present products and services based on neural networks, machine learning and large language models. In addition, organizations are actively helping their customers implement AI.

3. Infrastructure turbulence

The level of AI spending for many enterprises will be limited until 2025 due to the transformation of business processes and the redistribution of workloads in enterprise systems and cloud platforms. The current macroeconomic situation will have an impact on the industry.

4. Data arrays

Information is a critical asset in the AI world. Model learning efficiency and application functionality depend on datasets. Developers realize this, and therefore increase investment in data collection platforms in order to gain a competitive advantage.

5. IT Skills Mismatch

A shortage of qualified specialists in artificial intelligence, cloud, data processing and security will negatively affect the attempts of enterprises to succeed in the relevant segments of the IT market.

6. Service Industry Transformation

Generative AI will provoke change in many areas. IDC believes that by 2025, 40% of services, such as risk assessment and IT operations, will use GenAI tools in one form or another. These tools enable the creation of virtual assistants that generate humanoid responses, develop video games with dynamic and evolving content, and even generate synthetic data to train other AI models.

7. Unified control

One of the most challenging challenges for IT teams is the development of management platforms, which should cover many different areas: infrastructure operations, business applications and processes, AI systems and data

8. Convergent AI

Analysts believe that organizations should plan, test and implement fully convergent AI solutions that will allow them to develop new services to meet customer needs while saving costs.

9. Peripheral Computing

The introduction of generative AI will allow companies to improve peripheral computing platforms by aligning the results with customer expectations. And this will help to increase the effectiveness of business operations.

10. Satellite systems

Satellite internet constellations will provide broadband everywhere, helping to bridge the digital divide and opening up many new opportunities and business models. By 2028, 80% of enterprises will begin to use satellite communications, creating a single structure of digital services and services.^[9]

Named 6 main AI trends in education

In August 2023, the team of the Intersectoral Center for Technology Transfer and the Research Center in the Field of IIUnnivalence Innopolis prepared an open patent and marketing report "Application of artificial intelligence in priority sectors of the economy." In particular, experts have identified several AI trends in the educational sector.

Help with routine tasks and processing documents/tasks/forms

Here we are talking about the automation of processes: from checking tasks and processing forms to issuing ratings, including based on the results of written work.

Evaluation of works

AI has been used for years to support student learning and assessment. Using works that are first evaluated by faculty, AI grading systems examine how grading criteria are applied and then apply these criteria at scale. According to experts, only 15 of the noted works can lead to the fact that the AI system will be able to accurately estimate 10,000 works. AI systems can do this for various forms of evaluations, from solving mathematical problems to short or long essays and multiple choices. These tools can also evaluate the same task in different languages without considering external factors such as handwriting, culture, or the language the student chose.

Coaching and Teacher Professional Development

Artificial intelligence can provide feedback on a lesson's success, track a student's progress and warn when performance issues may arise, and identify areas where a teacher can improve teaching.

Digital assistants

Experts pointed to the possibilities of automating processes using digital assistants. The latter, according to experts, will not completely replace the teacher, but will do what the teacher does not have enough time for. Thanks to digital tools, the evidence-based capabilities of pedagogy increase many times over. First of all, this is a constant included measurement of the student - an operational assessment of the work that he performs, which is very important to increase his motivation. After all, the teacher does not have time to check all the notebooks, because often this is a boring, routine activity, and because of this, the student loses interest in studying, because he does not feel attention to himself or sees that he was assessed according to a template.

Another feature of the digital platforms with which schools begin to work is the generalization of feedback data for the teacher. For example, they show exactly where there are laggards, what topics they did not understand, and the teacher has a different information base for assessment.

Individual training

Artificial intelligence helps to create adaptive training programs. According to the researchers, AI will achieve a level of personalization that is impossible for teachers managing 30 students in the classroom by 2023.

Chat boats

With the chatbot, the teacher can collect statistical information about the student. There is no need to personally write to a student and ask his parents' phone number - all this information is already accumulated in the artificial intelligence system. The chatbot works as a teaching tool. He can offer the student additional tasks, links to the necessary resources. For example, the foreign language learning service Duolingo was one of the first in its field to use chatbots in its application. They understand spoken language and support audio message format in multiple languages. If at some point in the conversation the user has problems expressing his thought, the bot will give him several possible phrases to choose from.

The use of artificial intelligence in priority sectors of the economy

Named the leading platforms for the development of AI solutions

The leading cloud platforms for the development of technologies and services based on artificial intelligence the global scale are, and Amazon. Google Microsoft Such data are contained in the study, the Gartner results of which were released at the end of June 2023.

By services for AI developers, Gartner means cloud or container services that allow software creators who are not experts in the field of data processing to use AI models through APIs, software development kits (SDKs) or applications. The main opportunities in this market include automated machine learning (ML), data preparation, model building, etc. A number of additional features are also highlighted, such as language and visual tools, image creation, etc.

The Gartner report examines 11 cloud service providers for AI developers: companies, Alibaba Cloud,, Amazon Web Services (AWS) Baidu Clarifai,, Google H2O.ai,,, Huawei Cloud IBM Microsoft, and. Oracle Tencent It is noted that the platforms of these providers help developers in creating new applications and services, including generative AI, and also speed up the launch of products to the commercial market.

Amazon Web Services

The AWS platform is named one of the industry leaders. Gartner emphasizes that the company has an extensive global cloud infrastructure, and availability zones provide a wider coverage than any of the competitors. Thanks to this, AWS allows customers to easily deploy their AI models in several geographical regions.

Microsoft

The Redmond giant offers flexible uses: a free version of Azure AI, pay-as-you-go access for smaller teams and discounted subscription plans for major developers. Microsoft provides its machine learning services for free, charging only for computing resources and storage.

Google

The company offers pre-configured automated machine learning services for language, structured data, and images and videos.

IBM

Watson Discovery and Watson Natural Language Understanding provide ample opportunities to answer questions and generate descriptions. In addition, IBM provides advanced video content analysis features.

Alibaba Cloud

Alibaba's ModelScope platform offers many off-the-shelf models that developers can easily modify by optimizing them for their own projects.

Baidu

The company offers training courses, an open source platform and free computing power for developers. The cost of services is often lower compared to competitors.

H2O.ai

The company creates products together with customers, offering its technologies, experience in the field of AI, assistance in bringing to market, etc., while customers share data.

Tencent

The provider offers flexible pricing options, including free trial periods, pay-as-you-go, consumption-based subscriptions, and more.

Clarifai

The company provides developers with various functions for creating AI-enabled business applications, including generative AI, APIs, configurable pre-trained models, etc.

Oracle

The advantage of this platform is comprehensive support for various AI models - both pre-configured and industry-specific.

Huawei Cloud

The advantage of cloud services of this company is called the highest level of reliability - with a declared availability of at least 99.9%.^[10]

AI venture funding significantly exceeds investment in cryptocurrencies

According to PitchBook, in the second quarter of 2023, the volume of venture financing of cryptocurrencies fell to the lowest level since 2020, significantly inferior to investments in artificial intelligence.

2022

Global spending on applied AI exceeded $100 billion. Where it is most often used

At the end of 2022, global spending on applied artificial intelligence reached $104 billion. At the same time, the demand for specialists in the relevant field is growing rapidly: the number of vacancies has more than tripled from 2018 to 2022. Such data are given in a study by McKinsey, the results of which were released on July 20, 2023.

Analysts note that against the background of a difficult macroeconomic situation and a high level of inflation, costs for applied AI are decreasing. So, the peak was recorded in 2021, when the market volume reached $146.8 billion. Nevertheless, in 2022, spending continues to outstrip the levels of 2018-2020, when the average cost was $73.5 billion.

With application AI capabilities, companies in all industries can leverage data and gain insights to automate processes, add or expand capabilities, and make better decisions. McKinsey estimates that the potential economic benefit of using AI ranges from $17 to $26 trillion. In addition, the introduction of AI can reduce the burden on employees when performing routine tasks.

Among the key areas of application of applied AI, analysts call machine learning, computer vision, natural language processing and deep learning with reinforcement. AI models can be used to solve classification, prediction and management problems. The additional potential of applied AI can be unleashed by combining it with new technologies.

At the same time, a number of factors are distinguished that have a negative impact on the field of applied AI. These are a lack of available resources (such as skilled professionals and funding), cybersecurity and privacy issues, legislative regulatory issues and ethical considerations.^[11]

Global investment in AI technology has declined for the first time in 10 years

In 2022, global investment in artificial intelligence declined for the first time in 10 years. This is reported in a study that analysts of the NTI Competence Center in the direction of "Artificial Intelligence" on the basis of MIPT presented in July 2023.

According to experts, the decline in investment in AI can characterize the transition of investments from startups to mature companies implementing artificial intelligence. According to the report, the global venture capital market in 2022 amounted to $415 billion, showing a fall of 33% compared to 2021. The segment of artificial intelligence is 11% in 2022, and the share from 8 to 12% has been preserved for 6 years, the report noted. The venture capital market of AI in the world fell in 2022 by 31% and amounted to $45.8 billion. The number of transactions increased by 4% compared to 2021 and reached 2956. Compared to 2020, in 2022, the volume of investments in AI projects decreased by 47% in the volume of transactions and by 36% in terms of the number of transactions.

The leader in both volume and number of transactions is the United States, where funds for 2022 invested $26.7 billion in 1,150 transactions. Asian investors made 1,013 transactions, but the volume of investments amounted to $11.5 billion, which is 2.3 times less than in the United States.

The deals under consideration in the venture capital market include not only investments in the company's capital (transactions of rounds A, B, C and D), but also mergers and acquisitions M&A (), initial offerings on the stock exchange () IPO and transactions with special structuring (). SPAC

The number of M&A deals in AI fell 57% globally compared to 2021. A total of 259 transactions were made, 12% of them were in the field of health. In 2022, 19 IPOs (-18%) and 15 SPACs (-50%) took place - a company created specifically for merging with another private company that wants to go public, bypassing the IPO procedure.

In 2022, another 41 AI companies reached an estimate of more than $1 billion, becoming "unicorns." At the end of 2022, there were 166 AI companies in the world with an estimate of more than $1 billion. By the end of 2022, 258 unicorn companies appeared in the world, 16% of which are AI companies. The United States remains the leader in the number of unicorn companies. At the end of 2022, there were 105 unicorns in the United States, and 41 in Asia.

The expected dramatic fall occurred on the Russian market in 2022. Only 21 transactions worth $51 million were invested against 77 transactions worth $226 million in 2021, a drop of 78%.

Among the deals in 2022, experts highlight the purchase of a stake in Motive NT by Kaspersky Lab. The amount of the transaction was not disclosed.

According to the researchers, the structure of investments in the context of startup rounds changed very much in 2022. Many major rounds took place in 2021, including apparently delayed deals from 2019-2020. In 2022, investments fell 5 times, large investors paused, almost all transactions were in the earliest stages - seed and round A. 72% of all transactions in 2022 were in seed stages and round A.

In their analytical work, experts listed several trends in the global artificial intelligence market. Among them are large language models and ChatGPT, signs of the end of the era of Open Source technologies, the use of image-generating models in everyday life and work, as well as the emergence of new AI professions, the development of ethical artificial intelligence and the "new Internet" (when information can be searched not only in traditional search engines, but also in neural networks like ChatGPT).

Almanac "Artificial Intelligence 2022"

Global investment in artificial intelligence sank 27%

At the end of 2022, private investment in artificial intelligence on a global scale amounted to approximately $91.9 billion. This is 26.7% less than in 2021. The corresponding figures are reflected in the report of Stanford University, published on April 3, 2023.

It is noted that world costs in the AI sphere in 2022 decreased for the first time in a decade. A year earlier, record investments from private investors were recorded - approximately $125.4 billion. In 2022, private investment in AI was 18 times more than in 2013.

Impressive technological advances in AI have attracted the attention of politicians and industry leaders as well as the public. At the same time, the opinions of citizens regarding AI vary greatly depending on the country. On the one hand, the report says, neural network-based products and services help save time and money when solving complex problems and processing huge amounts of data from disparate sources. Artificial intelligence models help accelerate scientific progress, including in important areas such as the synthesis of new materials and the development of drugs. On the other hand, AI tools open up qualitatively new opportunities in terms of organizing surveillance of citizens, creating deepfakes, replacing human labor, etc.

The share of companies introducing AI stabilized by the end of 2022. Enterprises that apply such solutions have achieved reduced costs and increased revenues. The key applications of AI in the business sphere are process automation, computer vision and virtual agents. Overall, the share of companies implementing AI in 2022 more than doubled from 2017, according to the annual survey.

The report says that large language models that have led to significant progress in the field of AI are becoming larger and more expensive. For example, Google PaLM, one of the flagship models released in 2022, cost 160 times more and was 360 times larger than GPT-2, one of the first large language models released in 2019. Generative models such as Dall-E 2, Stable Diffusion and ChatGPT continue to develop actively.

The authors of the study say that until 2014, the most significant machine learning models were created by academia. The IT industry has since taken over. So, in 2022, 32 significant models were developed by various companies, while academic institutions produced only three. The fact is that modern AI systems require huge amounts of data, as well as large computing and financial resources. Therefore, often non-profit and scientific organizations cannot afford to develop advanced AI models due to financial difficulties.

In 2022, the US became the leader in private investment in AI technologies with $47.4 billion. In second place is China with costs of $13.4 billion, and Britain closes the top three with $4.4 billion. In terms of investment, such areas as medicine and healthcare ($6.1 billion), data processing and cloud technologies ($5.9 billion), fintech ($5.5 billion) are leading.

An analysis of indicators for 127 countries around the world suggests that the number of bills containing the phrase "artificial intelligence," which were adopted as laws, increased from one in 2016 to 37 in 2022. These documents cover a variety of areas - from reducing the risks of automation using neural networks to forming weather forecasts.

Artificial Intelligence Index Report 2023

McKinsey named 2 top trends in artificial intelligence

On August 25, 2022, a study was published, according to McKinsey which the application artificial intelligence and implementation of machine learning are the two most significant technological trends in the AI market.

Implementation of application AI

Applied AI, which McKinsey believes is based on proven and developed technologies, has a viable application in more industries and is closer to a state of mass adoption than other trends.

In McKinsey's global survey on the state of AI conducted in 2021, 56% of respondents said their organizations had implemented AI, up from 50% in the 2020 survey. Technology industries are leading the way in implementing AI, according to a 2022 report, and product development and service operations are the business features that have benefited the most from using AI.

We see that everything is moving away from advanced analytics... using machine learning to work with large data sets to solve complex problems in a new way.

This trend is reflected in the rapid growth of AI publications, and not only because AI scientists publish more work, but also because people in a wide variety of fields use AI in their research and advance the application of AI into the future, he explained.

Indeed, there is a shift from science to technical development and scaling, "he said. We see that development in the field of AI is moving quite quickly along this path, and I am very pleased with the fact that more and more things are moving from science to large-scale application.

However, McKinsey's report also highlights a number of key uncertainties that could impact the future of applied AI, including the availability of talent and funding, cybersecurity concerns and stakeholder questions about the responsible and reliable use of AI.

Machine Learning Implementation

According to the McKinsey report, the introduction of machine learning (ML) "implies the creation of a functionally compatible stack of technical tools to automate ML and expand its use so that organizations can fully realize its potential." McKinsey expects ML adoption to spread as more companies look to use AI for a growing number of applications, the report noted.

More broadly, ML includes a view of a technology stack that promotes scaling, which can go as far as innovating at the level of changing microprocessor structures, Roberts said. You see many new possibilities in silicon that support the acceleration of certain types of AI tasks, and these innovations will be used more widely, providing faster and more efficient scaling both in terms of computing resources and in terms of their stability.

The trend also includes integrated hardware and heterogeneous computing used in ML workflows.

We will begin to see more and more venture capital activity and corporate investment as we build a tool system for this new class of software and new class products in the form of specialized services, "he explained.[12]

2021

Spending on AI solutions in the world increased by more than 20%

On September 12, 2022, IDC analysts published a study according to which the global market for artificial intelligence (AI), including software and hardware, as well as services, amounted to $383.3 billion in 2021, which is 20.7% more than in 2020.

IDC predicts that the artificial intelligence market will be approximately $450 billion in 2022 and will maintain growth dynamics throughout the projected five-year period.

{{quote "In all sectors and segments, the end user is beginning to realize the benefits of AI technology, as increasingly efficient AI solutions allow for more informed decisions and greater performance," says Rasmus Endsbjerg, IDC Assistant Vice President for Data and Analytics. The realities are such that AI can offer solutions to all the challenges we currently face. AI can be used as a resource to accelerate digital transformation, provide cost savings during a period of staggering inflation, as well as support automation efforts during a period of labor shortages. }} In 2021, AI software still accounts for the largest share of the overall market. Together, four categories of AI software - the preparation and deployment of AI solutions, AI applications, AI system infrastructure software and artificial intelligence platforms - provided more than $340 billion in market value in 2021, with AI applications accounting for almost half of the total. Artificial intelligence platforms saw the highest growth for the year at 36.6%.

In the AI applications category, AI solutions for customer relationship management (CRM) and enterprise resource management (ERM) accounted for about 16% of the total category. The rest came from many other software AI solutions available on the market. With about 300 companies vying to get their share, the market for AI solutions remains highly competitive.

IDC AI Tracker demonstrates that AI-oriented applications, in which AI technology is central and critical to ensuring application health, continue to slowly increase their share of the AI-software market. In 2021, AI-oriented applications accounted for 12.9% of the market, which is 29.3% more compared to 2020. The share of the remaining market was in non-AI applications, in which AI technologies form an integral part of certain application workflows, but if these technologies are removed, the application may continue to function.

Similarly, the deployment of AI-software in the cloud continues to show steady growth. In 2021, 47.3% of purchased AI software was deployed in the public cloud, up 4% from 2020 and 8.4% from 2019. IDC expects that in 2022 the cloud deployment of the acquired AI software will exceed the volume of deployments in local networks.

The total cost of the AI-related services market increased by 22.4% on an annualized basis and amounted to $24 billion in 2021. Customer demand for specialists in the development of production AI solutions contributed to the growth of the AI category in the IT services sector by 21.9% on an annualized basis to $18.8 billion. The AI business services category grew 24.2% year-on-year as companies sought help on AI management, business processes and talent acquisition strategies.

Hardware AI was both the smallest ($18.8 billion) and fastest growing (38.9% growth per year) segment of the AI market. The hardware growth was driven by efforts to build specialized AI systems capable of meeting the increased computing and storage needs of AI models and data arrays. Although both servers with built-in AI and AI storage systems showed significant growth in 2021 - 39.1% and 32.9%, respectively - server purchases were significantly higher - $15.6 billion.^[13]

AI Software Market Up 14%

The volume of the global market software using algorithms in artificial intelligence 2021 will reach $51.5 billion, an increase of 21.3% compared to 2020. Such data in November 2021 was published in a research company. Gartner

The market under consideration at Gartner includes applications with built-in AI capabilities, for example, machine vision programs, as well as software for AI systems.

The largest segment of the market, analysts call virtual assistants - spending on them on a global scale by the end of 2021 will amount to $6.21 billion, which is 12% more than a year ago.

Higher growth rates (+ 13.7%) are expected in the software segments for self-driving cars and digital workplaces. Software sales here will be measured at $5.7 billion and $3.59 billion, respectively.

The leader in dynamics will be the segment of solutions in the field of knowledge management. We are talking about systematic processes for the creation, collection, accumulation, preservation, distribution and application of knowledge in companies. Such assets may include databases, documents, policies, procedures, as well as the knowledge and experience of individual employees that have not been previously recorded. In 2021, sales of such software, according to Gartner, will grow by 17.6%, to $5.47 billion.

The AI software market is accelerating, but its long-term trajectory depends on how much enterprises develop in the field of AI, says Alys Woodward, senior research director at Gartner. - Successful results of using artificial intelligence for business purposes will depend on careful selection of projects. Those AI use scenarios that bring significant business value but can be scaled to reduce risk are critical to demonstrating the potential for investment in artificial intelligence to have an impact.

Taking into account the volume of projected revenue, analysts included in the top 5 the following scenarios for using such software: knowledge management (KM), virtual assistants, unmanned vehicles, digital workspace and collective data collection.

Sales of software that provides collective data collection in 2021 will rise by 13.6% compared to 2020 and exceed $3.48 billion.

Gartner data suggests high interest in AI from enterprises. In a survey conducted by analysts, 48% of organization leaders said they are already introducing artificial intelligence and machine learning technologies or plan to do so in the next 12 months.

However, in reality, the use of AI remains limited. Organizations are experimenting with AI but face difficulties trying to make the technology part of their standard operations, a Gartner study has found.

Experts believe that only by 2025 half of organizations in the world will be able to reach the stage of AI implementation, which Gartner calls the "stabilization stage," indicating that maturity has been reached in this area.

Maturity in the field of AI will lead to revenue growth in the market under consideration, since it will contribute to an increase in the costs of enterprises for AI software, according to Gartner. On the contrary, lagging maturity, caused by a reluctance to use AI technologies, a lack of trust in them and difficulties in obtaining commercial benefits, will negatively affect organizations' spending on artificial intelligence and revenue in the market, analysts said.^[14]

Britain plans to become an AI superpower within 10 years

The British government has developed a plan according to which within ten years the country must bypass the United States and China in the field of AI. This became known on September 23, 2021.

The document "National Strategy for the Development of Artificial Intelligence" was developed with the aim of promoting the use of AI in business, attracting international investment in British companies in the field of AI and nurturing a generation of British talents in the technological field.

The project includes a number of programs, reports and initiatives. Among them is a national AI development and research programme designed to improve coordination and collaboration between UK researchers.

The government also intends to increase investment in industries not yet fully using AI, such as power and agriculture. In addition, an analysis of the availability and performance of computing power for British researchers and organizations will be carried out, and consultations on copyright and patents for AI will be aimed at assessing what benefits patented technologies can bring to the country.

Cambridge University specialist Beth Singler, who specializes in AI and robots, said the UK is making increasing efforts to find professional areas in which it could compete with larger states after Brexit.

Stanford University Research Center specialist Seb Krier said there are several "very promising" aspects of the National Strategy for Artificial Intelligence. According to the chief executive of DeepMind, Lila Ibrahim, it is good to see a clear focus on effective technology management, and it is simply vital^[15] gain public and business confidence in AI].

Gartner: The AI market is still at the "stage of development"

In early September 2021, the analytical company Gartner published a study stating that artificial intelligence technologies are still at the "stage of development" and the formation of a full-fledged developed market is still far away.

According to experts, many companies want to use AI to solve problems for which artificial intelligence is still adapted. These companies focus on implementation, risk management and ethics, while they scale different AI initiatives. Data leaders, however, do not always prioritize investment in AI technology, threatening the ability to extract any data from these initiatives in the first place.

Gartner went on to clarify that responsible AI is critical to stakeholders as they expect more trust, transparency, fairness and the ability to verify. Therefore, by 2023, all experts hired to work on AI will have to prove that they are capable of working on responsible AI. What exactly this may entail, it is not yet clear, most likely, at the selection stages, when closing a vacancy, tests will be introduced that will help the hiring company assess the mental state of a technical specialist at the selection stages, and possibly annual checks with the provision of a separate certificate.

The COVID-19 coronavirus pandemic has played an important role in the development of technology. Startups working on AI products raised more than $73 billion in the 4th quarter of 2020, which is $15 billion more than in 2019.

Studies have shown that only half of AI projects go from the pilot stage to the production stage. The transition period takes an average of nine months. Innovations such as AI orchestration and automation platforms and model operationalization provide reusability, scalability and management, accelerating AI adoption and growth, says Gartner Vice President of Research Svetlana Sikular.

Even though AI is far from perfect, its impact on the global economy since early 2018 has been transformative. Technologies such as computer vision, intelligent decision-making, machine learning have already radically changed various industries around the world, but this is only a small part of the overall capabilities associated with AI.^[16]

Australian court allows AI to obtain patents

An Australian court allowed AI to obtain patents. This became known on August 3, 2021.

Australian laws nowhere say a patent applicant must necessarily be human, a judge has said. Read more here.

What AI costs are built from and how to reduce them

In May 2021, Aikanysh Orozbaeva, head of the department for work with partners and clients at Embedika, spoke about what the costs of AI are built from and how to reduce them. Read more here.

Gartner: 5 AI market forecasts

At the end of January 2021, Gartner released a series of research forecasts for 2021. Including analysts told how the artificial intelligence market will develop until 2025, and described serious and far-reaching ethical and social problems associated with AI. In the article "Forecasts for 2021: Artificial Intelligence and its Impact on People and Society," five Gartner analysts presented their considerations, paying special attention to the unintended consequences of the development of new technologies.

The report looks at five different forecasts for the artificial intelligence market and provides guidance on how companies can address emerging challenges and adapt to the future:

• by 2025, pre-trained AI models will be mostly concentrated in the hands of 1% of suppliers;
• in 2023, 20% of successful account-grabbing attacks will use deepfakes;
• by 2024, 60% of AI providers will include measures in their software to prevent its potentially harmful/misuse;
• by 2025, 10% of governments will avoid privacy and security violations by using selected populations to train AI;
• by 2025, 75% of workplace conversations will be recorded and analyzed to enhance organizational value and risk assessment.

Together, these predictions paint a picture of a bleak future with multiple ethical challenges, potential abuse of AI and loss of privacy in the workplace.

The capabilities of artificial intelligence already allow you to generate hyper-realistic content. Soon people will forget forever what it means to believe their own eyes, says the report.

Generative AI, for example, is already capable of creating surprisingly realistic photos of people and objects that do not actually exist, and in the near future a new generation of the same software will be used to create deepfakes. Privacy violations and unreliability of news are sure to become a major problem in the coming years, and companies should be ready to quickly adapt to these conditions, analysts say.

Among other things, they recommend that companies that rely on AI services set up ethics committees. This advice is especially relevant for enterprises that plan to record and analyze conversations in the workplace: for example, analysts believe that by 2023, large US corporations will use analysis of employee conversations to determine the amount of remuneration. Gartner believes they should first set up a committee with employee representation to ensure data is used fairly. Gartner also recommends that companies establish criteria for responsible use of AI and prioritize providers that are "responsible for the development of AI and provide solutions to relevant social problems."

Regarding security concerns related to deepfakes and generative AI, Gartner recommends that organizations schedule deepfake training.

We are moving to a world of zero trust. No one and nothing in this world can be trusted without seeing the confirmation in the form of a cryptographic digital signature, the report says.

Analysts believe that in the long term, the best software for detecting deepfakes will reach a 50% level of identification, which means that companies will have to protect themselves^[17]

2020

An increase in the number of scientific publications on AI in Russia by 2.6 times, to 4.8 thousand, in the world - by 1.3 times

In the period from 2016 to 2020, Russia accounted for about 4.8 thousand scientific publications in the field of artificial intelligence technologies. In comparison with 2011-2015. the indicator increased 2.6 times, and in the world - 1.3 times. This was reported at the Institute for Statistical Research and Economics of Knowledge (ISIEZ) of the Higher School of Economics with reference to the database of scientific citation Web of Science Core Collection.

According to experts, a significant part of Russian research and development in the field of AI is concentrated in Moscow. Among 36 world cities, the capital of Russia is in 17th position in terms of the number of publications in this area (2,003). For comparison: the leaders of Beijing and Shanghai have 19,602 and 7,161, respectively. However, in terms of the average number of citations, Moscow is in the penultimate place (2.9 citations per publication).

In Moscow, the largest number of publications in the field of AI in 2016-2020 was prepared by researchers from the Higher School of Economics (271), MIPT (232), MEPhI (226), Moscow State University (195) and the Federal Research Center "Informatics and Management" of the Russian Academy of Sciences (193). Researchers from Rostov-on-Don, Kazan, Tomsk, Volgograd, Yekaterinburg and several universities in St. Petersburg are also published. Representatives of large IT companies are also producing more and more works. So, in the corporate ranking in 2016-2020, the top 100 in terms of the number of publications included Microsoft (1455) publications, of which 175 were in Q1, Google (1312 and 76) and IBM (989 and 54). In Russia, Yandex (42), the Center for Speech Technologies (39) and Roscosmos (33) have the most publications.

In terms of the number of publications in the field of artificial intelligence for 2016-2020, Russia took 17th place. In the first position is China (76.3 thousand publications), in the second - the United States (44.4 thousand), in the third - India (27 thousand).^[18]

China spends 350 times more on AI than Russia

China spends 350 times more on AI than Russia. This became known on April 16, 2021. Read more here.

Countries have developed AI strategies

Growth in investment in AI by 40%, to $67.9 billion

In 2020, the volume of investments in developments based on artificial intelligence technologies increased by 40%, reaching $67.9 billion. This is evidenced by data from the AI Index Report 2021 from Stanford University researchers.

The share of private investment in the development of artificial intelligence technologies in 2020 showed an increase of 9.3%, exceeding $40 billion. Among the leaders in the amount of funds raised: technology companies in the development of new medicines using AI solutions, including for the fight against cancer, as well as companies that specialize in research in the field of molecular medicine. Over $13.8 billion was invested in these developments, which is 4.5 times more than in 2019. Also among the leaders of the company, which develop unmanned vehicles ($4.5 billion) and companies developing educational products based on AI ($4.1 billion).

Mergers and acquisitions accounted for most of total investment in 2020, up 121.7 per cent on 2019, report says

The demand for specialists in the field of artificial intelligence has grown significantly over the past seven years. The US is the only country of six reviewed by researchers to record a decline in the proportion of AI job adverts from 2019 to 2020. This may be due to the COVID-19 coronavirus pandemic or the development of the AI labor market. Singapore, Brazil, India, Canada and South Africa have become the countries where demand for employees in this area has increased the most over the past 4 years.

Despite growing calls to pay attention to ethical issues related to the use of AI, companies show limited attention to the accuracy and fairness of decisions made using AI, according to a McKinsey survey. ^[19]]

Coronavirus will lead to explosive growth in AI spending

The spread of the coronavirus COVID-19 will lead to an explosive increase in spending on artificial intelligence in the world, according to IDC analysts, who announced their forecast on April 9, 2020.

According to experts, after the outbreak of COVID-19, many industries, such as transport and consumer services, will have to reduce their investment in technology.

On the other hand, artificial intelligence is a technology that can play an important role in helping companies and society cope with and resolve large-scale problems caused by quarantine restrictions, says Andrea Minonne, senior analyst at IDC Customer Insights & Analysis. - In all industries, the public sector will experience accelerated investment in AI. Hospitals see AI as a means of speeding up diagnosis and testing for COVID-19 and providing automated remote advice to patients in isolation. At the same time, governments will use AI to assess compliance with social distance requirements.

The study notes that after the coronavirus pandemic, AI solutions deployed in the cloud infrastructure will become widespread, showing that companies consider the deployment of smart technologies in the cloud to be a more efficient and flexible process.

IDC interviewed representatives of more than 180 companies in Europe, and 16% of respondents said they see automation using artificial intelligence and new technologies as a means to minimize the consequences of the pandemic. Lack of employees and supply chain failures will lead to the need to automate all production processes.

IDC Customer Insights & Analysis analyst Petr Vojtisek added that artificial intelligence can play an important role for companies in responding to the COVID-19 crisis in the near future, as these technologies automate processes and help reduce the use of human labor.^[20]

2019

The market size of AI solutions for IoT amounted to $2.64 billion

According to the ReportCrux Market Research^[21]global demand for artificial intelligence in the IoT market was valued at about $2.64 billion in 2019, and revenue is expected to be about $15.72 billion by the end of 2027, giving average growth (CAGR) of 25.0% from 2020 to 2027.

IoT is Artificial intelligence a combination of connected to the Internet computers to other connected devices, analyzing big data and providing the required amount of data to the system. Using AI in IoT, it simplifies the collection of data and solving methodological problems, since AI conducts better analysis. In addition, the need to process large amounts of real-time data produced by IoT devices is expected to create new opportunities for artificial intelligence in the IoT industry.

The global artificial intelligence market in IoT is subdivided by technology, end-user categories, components, and region.

By component, the artificial intelligence market in IoT is segmented into software, platforms, and services. The service segment is further divided into professional services and managed services. The software segment, will grow with a high CAGR in the forecast period. Most software systems are designed to solve compatibility problems arising from heterogeneous devices, as well as to process large amounts of data, protect them and privacy. The software helps reduce interoperability issues and creates flexibility to effectively manage various business functions.

By technology, the market is subdivided into machine learning (ML) and deep learning, natural language processing, and others. In 2019, the machine learning and deep learning segment held nearly 65% of the market share. ML has become a reliable analytical method for large amounts of data. The combination of ML and edge computing removes most of the noise generated by IoT devices, leaving the appropriate data for analysis by the edge and cloud analysis mechanisms. By end user, the global artificial intelligence market in IoT is segmented into healthcare, insurance and finance, power and utilities, government and defense, telecommunications, manufacturing and others. The finance and insurance segment held almost 27% of the market in 2014. The health sector is expected to take a leading role by 2027. By application, it is divided into, robotic surgery, virtual medical assistants, process management in hospitals, reduction of dosage errors, identification of clinical trial participants, automatic imaging diagnostics and preliminary diagnostics, these applications should lead to market growth in the forecast period.

Geographically, the market is subdivided into countries in Latin America, the Middle East and Africa, Asia-Pacific, North America and Europe. North America held nearly 39% of the market in 2019. In the Asia-Pacific region, the highest CAGR is predicted. Such growth will be due to large public and private investments aimed at improving AI and IoT technologies, which in turn leads to increased demand. North America is projected to be the leading region in terms of IoT adoption and AI growth, increasing government spending on new technologies is expected to drive market growth during the forecast period.

The main players in the field of artificial intelligence in the IoT market are Google, (USA), Anagog (Israel), Oracle (USA), Salesforce (USA), Hitachi (Japan), Autoplant Systems Pvt. Ltd. (India ), Kairos (USA), SAP (Germany), IBM (USA ) and Microsoft (USA). These suppliers have implemented numerous organic and inorganic growth strategies such as alliances and agreements, new product launches, and mergers and acquisitions to further expand their presence in the global market.

Gartner Hype Cycle for AI

8 new technologies hit the 2019 cycle:

• AI Cloud Services
• AutoML
• Augmented intelligence
• Explainable AI
• Intelligent devices
• Reinforcement Training
• Quantum computers
• AI marketplaces

Investment in AI companies turned out to be a record - CB Insights

At the end of January 2020, CB Insights conducted an annual analysis of global investment trends in artificial intelligence and reported that in 2019, startups specializing in such technologies attracted a record investment of $26.6 billion, concluding more than 2,200 transactions around the world. For comparison, in 2018, about 1900 agreements were concluded for a total amount of $22.1 billion, and in 2017 - about 1,700 pieces for $16.8 billion.

The record recorded by analysts corresponds to the analysis of other organizations monitoring investments in the AI ecosystem. Earlier, the National Venture Capital Association announced that although total venture capital declined in 2019, investors contributed a record $18.4 billion to the development of AI startups in the United States.

Among the areas of investment, healthcare was the leader - investors invested $4 billion in the relevant AI companies at the end of 2019. It is followed by industries such as finance ($2.2 billion), retail ($1.5 billion), sales and cybersecurity. M&A deal activity was also the highest in healthcare, sales and retail. Another 2019 report found that private investors were most often keen to invest in areas such as self-driving, drug research, financial AI and facial recognition technology.

In addition, in 2019, the number of AI companies increased, whose value exceeded $1 billion. Among the new "unicorns" are the developer of autonomous robots for the delivery of goods Nuro and the analytical company DataRobot. All ten new billionaire companies are based in China, the UK or the US. Venture capital investors such as Plug and Play Ventures, Accel and Lightspeed Ventures were among the leaders in investing in the artificial intelligence market in 2019.^[22]

AI rate outpaces Moore's Law

At the end of December 2019, Stanford University released the results of a study according to which the computing power of artificial intelligence has been ahead of Moore's Law for more than seven years.

This law states that the processor speed doubles every 18 months, which means that developers can expect application performance to double in this time frame at the same hardware cost. But a report from a team of researchers at Stanford University, produced in collaboration with McKinsey & Company, Google, PwC, OpenAI, Genpact and AI21Labs, found that AI's processing power is growing faster than that of traditional processors. The turning point when the speed of development of artificial intelligence began to outstrip Moore's law was 2012.

The researchers studied how AI algorithms improved over time, tracking the progress of image identification program ImageNet. Given that image classification methods are largely based on controlled machine learning methods, the report authors considered how long it takes to train an AI model and the associated costs.

In 18 months, the time it took to train the network on cloud infrastructure fell from three hours in October 2017 to 88 seconds in July 2019, the study found. The authors of the report used the ResNet model to estimate how long it takes for image classification algorithms to achieve a high level of accuracy.

In October 2017, 13 days of training was required to achieve accuracy above 93%, which cost developers $2,323. The latest benchmark, available on the Stanford DAWNBench platform in September 2018 with an image classification accuracy of just above 93%, cost just over $12.^[23]

Record investment in AI startups

Artificial intelligence startups received $7.4 billion in funding in the second quarter of 2019, the highest on record. The number of deals was also growing. 488 transactions on artificial intelligence were concluded.

40% of European AI startups are not related to artificial intelligence - MMC Ventures

40% of European startups engaged in developments in the field of artificial intelligence do not actually use it. This conclusion was reached by analysts at the venture capital firm MMC Ventures based on a survey of 2,800 startups in 13 EU countries published in March 2019.

In 40% of cases, we could not find evidence of the use of artificial intelligence, "MMC Research Director David Kelnar told Forbes magazine.

Startups that "people perceive and consider AI companies are probably not," he said.

Some emerging technology developers use the buzzword "artificial intelligence" to draw attention to themselves and their products in order to get more funding. According to MMC estimates, companies that claim to work on AI solutions attract an average of 15-50% of investments.

At the same time, startups themselves do not always declare the use of artificial intelligence, Forbes notes. Some analytical sites attribute the prefix "AI" to companies, and the MMC report does not specify the percentage of companies that purposefully mislead their customers and investors when talking about the use of artificial intelligence.

The partner of intellectual property law firm Gill Jennings and Every, Peter Finney, confirmed the study's findings. The number of AI patent applications has grown as companies seek to make their businesses more attractive to investment, he said.^[24]

The AI sector is a bubble that could soon burst, he said.

MMC studied the most popular AI startups, regions and destinations. Most of these startups in Britain are 479 by March 2019, followed by France (217 startups), Germany (196), Spain (166) and the Netherlands (103). As for the spheres, most startups are engaged in developments related to medicine and health maintenance.

AI market to grow 31% annually -Frost & Sullivan

According to analysts at the international consulting company Frost & Sullivan, by 2022 the total volume of the AI technology market will increase to $52.5 billion, or 4 times compared to 2017 ($13.4 billion). The annual growth rate (CAGR) in the forecast period will remain at 31%. The widespread adoption of AI technologies by 2030 will increase the global market for goods and services by $15.7 trillion, according to TAdviser in Frost & Sullivan on January 15, 2019.

If we analyze the data over the past few years, we can conclude that artificial intelligence will become an integral part of our personal and professional life in the near future, "said Denis Stepanov, director of business development at Frost & Sullivan. - This technology has already been adopted in all major industries. The average annual costs of large companies for applications based on artificial intelligence technology are constantly growing, as companies want to gain a competitive advantage by improving their business processes using machines.

According to Frost & Sullivan, the leader in the amount of investment in the development of AI technologies and the number of deals concluded remains the United States, followed by China and the United Kingdom. However, taking into account the average value of the transaction, the undisputed leader is China, whose average deal size in the last 2 years has been more than $100 million against $15 million and $5.4 million from the United States and Great Britain, respectively.

China is actively developing in the direction of AI technologies. In 2017, the share of world investment that came to Chinese startups for the first time exceeded the share of American companies and amounted to 48%, Stepanov said. - At the same time, 54% of all AI startups are located in the United States, and only 15% in China. Such an increase in the popularity of artificial intelligence in China is due to the fact that the local government is actively promoting technology in many industries, from agriculture to the military-industrial complex. If this AI technology market continues to grow at this rate, the United States risks losing its leading position, and China has every chance of becoming a world leader in the technology development and application market.

The most popular companies among investors in 2018 were Chinese SenseTime ($1,200 million in investments), UBTech Robotics ($820 million), Megvii Technology ($600 million), YITU Technology ($300 million) and American Dataminer ($391 million), CrowdStrike ($200 million) and Pony.ai ($214 million), noted in Frost & Sullivan. Tech AI startups have caught the attention of companies such as Alibaba Group, Fidelity International, General Atlantic, Accel and others.

Research in various technological fields of artificial intelligence is carried out by companies such as IBM, Google, CloudMinds, Affectiva, etc.

In health care, AI is mainly used to identify patterns in medical data and allows for a significant increase in treatment accuracy, minimizing the possibility of making false diagnoses, and, as a result, reducing treatment costs. Thus, the California startup HealthTap has developed an application that analyzes the records of more than 100 thousand doctors to identify the most accurate diagnosis.

AI is also transforming the cybersecurity space, enabling innovation to be applied to protect against threats that traditional network security is not able to detect. Deep Instinct (Israel) has developed its own deep learning algorithms (one of the AI areas) that identify anomalies in network behavior and the presence of targeted cyber attacks ( APT) that cannot be detected using traditional technologies.

The large impact of AI-based applications also extends to industries such as ICT, banking and finance, automotive, video games and entertainment. According to Stepanov, in the future the influence of AI applications will increase, including in the aerospace and defense industries, power, retail.

According to Frost & Sullivan analysts, the steady growth of the AI technology market in the world will also be due to the integration of smart applications in areas such as robotics, the creation and development of cognitive, neuromorphic, quantum and large-scale computing systems.

Among the many startups in this area, one can separately single out the project being implemented MIPT to develop hybrid chips (neural co) processors based on memristor and optogenetic technologies, which is a synthesis of world-class basic research and the applied application of these developments in several areas at once: neuroscience and neuropharmacology, prostheses and artificial body parts integrated with the human nervous system, energy efficient and ultra-fast calculations, "added man-machine interface Denis Stepanov.

2018

AI software market growth by 26.6% to $2.6 billion - IDC

The global market for artificial intelligence software platforms reached $2.6 billion in 2018, an increase of 26.6% compared to 2017. Such data analysts IDC released in June 2019.

David Schubmehl, director of research at Cognitive/Artificial Intelligence Systems at IDC, says that in 2018 companies were moving from testing artificial intelligence to the industrial operation of such applications in their business processes.

The market for software AI platforms is growing, providing the tools and capabilities that organizations use to develop such applications, the expert noted.

IDC expects this trend to continue and is confident that demand for open source solutions will continue to grow.

According to experts, the time has come to act for manufacturers who want to make money on artificial intelligence, deep and machine learning.

Analysts call IBM the largest manufacturer of AI software, which in 2018 earned $240.6 million on such products, which is 26.6% higher than a year ago. IBM's market share reached 9.3%.

In second place is SAS, whose business is growing even faster than that of the leader. In 2018, the company's revenues in the market under consideration increased by 104.6% and amounted to $89 million.

According to David Schubmel, SAS has shown impressive growth in the artificial intelligence technology market thanks to its leading position in analytics.

As organizations move from experimentation to industrial exploitation of AI to meet their business challenges, many are looking for a reliable provider that offers analytical expertise and domain knowledge. Built-in AI SAS capabilities and integration with open source technologies allow the company to take advantage of the technology to automate processes without the heavy lift of AI training models, "said Schumbel.

AI-based solution development is among the key priorities of SAS. This technology is at the peak of demand, and strong positions in the field of in-depth analytics allow us to use its capabilities to the maximum, the company noted. According to IDC, SAS holds 30.8% of the global in-depth analytics market.

The IDC report says that the growth of the market for platform AI solutions and the companies developing them is due to the high demand for artificial intelligence in the corporate sector.

Earlier in 2019, analysts Gartner published the results of a study that showed that the number of companies using artificial intelligence in work processes in four years has become 270% more. 37% of companies use AI systems in one form or another. In 2015, this figure was measured at 10%, in 2018 - 25%.

Experts agree that AI is becoming an integral part of future corporate strategies regardless of the market. For example, 52% of telecommunications companies whose representatives took part in the study already use chatbots to improve the quality of customer service.

At the same time, according to the analytical agency IDC, only 25% of organizations that have already implemented AI systems have developed corporate strategies for using this technology. Half of respondents consider AI a priority, and two-thirds of respondents reported supporting a culture in which artificial intelligence comes first.

According to IDC forecasts, spending on software AI platforms in the world will grow by 36.7% annually and reach $9.5 billion by 2022.^[25]

Chinese AI market growth 52.8% to 33.9 billion yuan

In 2018, sales of software and equipment intended for the creation and operation of artificial intelligence systems in the Chinese market reached 33.9 billion yuan ($4.93 billion at the exchange rate as of July 16, 2019), which is 52.8% more than a year earlier. Such data were published in July 2019 by the non-governmental organization Internet Society of China. Read more here.

Deep learning chip sales reach 164.9m - Tractica

The global market for chips designed to solve deep learning problems reached 164.9 million units in 2018. Such data were released on May 6, 2019 by the analytical company Tractica.

According to the study, the rapid introduction of artificial intelligence to solve practical business problems has led to a number of uncertainties and risks in almost all markets. However, there is a fact that cannot be denied: in the modern AI industry, equipment is a key factor in overcoming the main problems, and chipsets are the "heart" of such devices.

Given the wide availability of AI, it will almost certainly happen that each application in the future will need some acceleration using AI chipsets, whether at the data center or at the edge of the network. Acceleration can take many forms, from a simple AI library running on a CPU to more complex user hardware.

According to experts, the potential of artificial intelligence is best manifested when AI chipsets are optimized to provide the necessary computing power with the right energy consumption for specific AI applications. This trend leads to an expansion of specializations and a variety of AI chips.

It is predicted that global supplies of chips designed to implement deep learning will exceed 2.9 billion units by 2025, and the market volume will reach $72.6 billion in money.

Graphics and central processors are in greatest demand in the market under consideration. However, in the future, special purpose integrated circuits (ASIC) will take the lead. In addition, growing demand is expected for single-chip (SoC) accelerators and programmable solutions (FPGAs).

As for the largest manufacturers of deep learning chips, Nvidia and Intel are called the leaders in Tractica.^[26]

Market size of $8.1 billion - Tractica

At the end of December 2018, the analytical company Tractica presented a report on the development and deployment of artificial intelligence (AI) technologies. According to experts, the volume of the global market for relevant software at the end of 2018 reached $8.1 billion and will only grow in the future. Sales are expected to measure $105.8 billion by 2025.

Software with elements of artificial intelligence is becoming more and more popular due to the growing volume of data, increased computing power of computers and increased performance of algorithms. AI technologies have penetrated almost every industry, allowing machines to speak, listen, move and make decisions, and the many uses for them reveal potential opportunities for business development, attract new investments and contribute to changes in existing business processes.

AI technologies are used in 258 areas at the end of 2018, according to a Tractica report. Chief analyst Keith Kirkpatrick says the following:

The capabilities of AI can be used in a wide range of industries and regions: from advertising and automotive to transport and telecommunications. At the same time, a significant part of the total income is concentrated in highly specialized markets with the need to process large data sets.

Tractica's report, Cases of Artificial Intelligence, details the use of AI in consumer, corporate and state markets. The company identified, contextualized, and offered real-world revenue examples and forecasts for each use case, organizing them by industry sector. The resulting report serves as a guide to AI market forecasting, offering an overview and analysis of each use case included in the model. After conducting an in-depth analysis of the application market, Tractica also identified opportunities in the report to expand the artificial intelligence market.^[27]

In the top ten most popular areas of software use with elements of artificial intelligence in terms of revenue, experts included: video surveillance, monitoring and management systems for IT networks and operations, customer service and marketing, voice and speech recognition, detection and recognition of objects by equipment/vehicles and evading collision with them, patent examination and analysis, photo and video image request, medical image analysis, patient data processing, predictive maintenance.

According to the calculations of the analytical company IDC, the global market for cognitive systems and solutions in the field of artificial intelligence at the end of 2018 will amount to $24 billion, and in 2022 the figure will be measured at $77.6 billion. Researchers estimate the average annual growth rate of this market at 37.3%.

Analysts expect the largest AI investments in 2018 in the retail to trade and banking sectors (more than $4 billion). In areas such as discrete production, medicine and continuous production, significant costs are also expected in AI products.

About 40% of cognitive and AI technology costs come from software, including cognitive platforms that provide tools and technologies for analyzing, organizing, accessing, and providing consulting services based on structured and unstructured information. This segment will be the fastest growing: each year its volume will increase by an average of 43.1%.

How artificial intelligence is improving - the main trends and obstacles

Main article: How artificial intelligence is improving - the main trends and obstacles

2017

Deep Learning Software Market Valued at $3 Billion

On November 26, 2018, Tractica, an analytical company, published some results of a study of the software market designed for deep learning tasks. Sales of such software in 2017 reached $3 billion, and by 2025 the figure could grow to $67.2 billion.

According to experts, deep learning is a fashionable term that has been widely advertised by companies and technology publications over the years. However, until 2016, the introduction of such software brought very meager results that did not meet expectations.

After this period, the true potential of deep learning was realized thanks to advances in the development of equipment and algorithms used to recognize samples in a continuous learning cycle, which allowed programs to self-learn on tasks that did not require explicitly specified program code.

By the end of 2018, practical deep learning projects are focused on areas such as image recognition, text analysis, product recommendations, fraud prevention and data curation. At the same time, according to analysts, the technology can lead to the creation of even more productive and breakthrough applications in the field of self-driving cars, personalized equipment, preventive medicine, etc.

Tractica senior analyst Keith Kirkpatrick says deep learning has become a key area of interest because of its potential to transform entire industries. The market leaders of such ON experts are called companies,, Google,, and Facebook Microsoft. IBM Amazon Baidu

The largest demand for deep learning software falls on North America - in 2017, the share of this region in the market was approximately 50%. The top three were Asia-Pacific and Europe.^[28]

Markets And Markets: Pattern recognition market size to reach $29.9 billion by 2020

According to Markets And Markets, published in 2017, the volume of the pattern recognition market will reach $29.98 billion by 2020 with an average CAGR of 19.1%. Pattern recognition technologies include pattern, optical, code, object, and digital photo recognition. They are either individually or integrally used in areas such as security and surveillance, image scanning and creation, marketing and advertising, augmented reality, and image search.

The key driver of this market is the departure of all processes both in business and in the consumer segment into the clouds, as well as the growing influence of the Internet, smartphones, and social media. The actors of this market are such large corporations as NEC, Google, Honeywell, Hitachi and Qualcomm. There are also many smaller players such as LTU Technologies, Attrasoft, Blippar and SLYCE, and vendors such as Catchoom and Wikitude.

2016

Sales of software with AI algorithms were estimated at $3.2 billion

In 2016, the volume of the global software market using artificial intelligence (AI) algorithms reached $3.2 billion, according to the research company Tractica.

According to experts, sales of this software by 2025 will increase to $89.8 billion, while previously lower rates were called. So, in May 2017, experts predicted an increase to $59.8 billion.

The increase in the forecast in Tractica was associated with the emergence of new areas of application of artificial intelligence in various industries.

Artificial intelligence has already become the key to how Internet companies work today, allowing them to create the most personalized services in accordance with the primary AI strategy, says Tractica research director Aditya Kaul. - The rest of the market, including the corporate and public sectors, is still lagging behind in the introduction of artificial intelligence and still does not fully understand the value of the technology, including the breadth and depth of its use.

However, according to the expert, this lag of companies and state institutions will rapidly decrease. If in 2016 about 26% of sales of software with AI elements fell on the consumer segment, then by 2025 this share will decrease to 8%, although in dollar terms the segment will grow more than 7-fold, the researchers predict. According to them, this trend will be expressed in the fact that the pace of implementation of AI systems in companies will be higher than the growth dynamics of the consumer market for such software.

The report also notes that by 2017, AI solutions are beginning to penetrate into the realms of real life, and the pace of implementation has become the fastest in history, as startups and IT vendors rushed to create platforms and niche products to solve certain tasks of corporate clients.^[29]

Tractica: AI Market Size $643.7 Million, to Reach $38.8 Billion by 2025

Although AI technologies will affect almost every business in the future for 10 years, the main drivers of the market will be the sectors of consumer products, business services, advertising and defense. Tractica predicts revenue growth in the AI market from $643.7 million in 2016 to $38.8 billion by 2025.

Tractica, based on the taxonomy of 191 real cases, divided the AI market into 27 sectors. Experts suggest that usecase such as pattern recognition, algorithmic exchange trading and patient data management in healthcare have tremendous scaling potential, while other cases are so far niche. The dynamics of AI, in their opinion, will be based on six fundamental technologies:

• machine learning,
• in-depth training,
• computer vision,
• natural language processing,
• machine argumentation and
• strong AI.

The development of the AI market in terms of market share will occur mainly in North America, since even now this region is the center of the dynamic development of advanced technologies, production processes, infrastructure, disposable income, etc. The broad adaptation of AI technology in industry, media and advertising, healthcare, BFSI, transportation and the automotive industry is a key factor supporting the growth of the AI market in this region.

Artificial intelligence became a key technology trend in 2016, and the volume of global investment in it exceeds $500 million. According to forecasts of the international research company Markets and Markets, by 2020 the AI market will grow to $5 billion due to the use of technology machine learning and natural language recognition in advertising, retail, to trade finance and healthcare.

Gartner believes that by 2020, about 40% of all interactions with virtual assistants will rely on data processed by neural networks.

AI startups raised $5 billion in investments

According to CB Insights, in 2016, startups in the field of artificial intelligence (AI) were able to attract a total of $5.02 billion in investments, which was the largest amount in the past five years.

The number of rounds of financing projects in the field of AI increased from 160 (2012) to 658 - that is, more than 4 times. The share of non-American startups also increased, gradually increasing from 21% to 38% over the period from 2012 to 2016, although most of the investment still goes to American companies.

Several megarounds of funding were conducted. The largest amount was received by the Israeli application Gett, which uses AI algorithms to deploy autonomous cars. Volkswagen backed Gett with a $300 million investment to expand Gett's predictive algorithms.

Market And Markets: Natural Language Processing Technology Market - $7.63 Billion

The Natural Language Processing (NLP) market is estimated by Market And Markets at $7.63 billion in 2016 with a forecast of growth to $16.07 billion by 2021, with a CAGR of 16 1%. The main drivers of the company consider the increasing demand for a more advanced level of user experience, the growth of the use of smart devices, the growth of investments in healthcare, the growing use of network and cloud business applications and the growth of M2M technologies.

Speech Recognition Technology Market Size - $104 Billion

The global speech recognition technology market is valued by BCC Research at a whopping $90.3 billion in 2015. This market is expected to grow from $104.4 billion in 2016 to $184.9 billion in 2021 with an average pace (CAGR) of 12.1% for the period 2016-2021.

2005-2007: Launch of AI in business and industry

The first commercial AI projects began to be implemented only in the 21st century. The use of AI to solve specific problems in business, finance, healthcare and other areas, including process optimization, market analysis and diagnostics in medicine, has increased.

From 2005-2007, a boom in algorithmic systems began based on predictive analysis in trading robots on the exchange.

Hardware improvements have opened up new avenues for AI-based weather forecasting and in robotics.

Notes