Explore Our Perspectives

Find the latest trends, opinions, and news about technology
  • Apr 27
  • Raghav

The adoption of AI & ML is increasing across companies in numerous industries. More specifically, the spending on AI & ML semi-conductor space is escalating as analysts estimate a 50% revenue growth y-o-y in 2021. Overall, the spending on AI software and hardware will reach approximately $138.2 billion in 2021, growing at a 23.1% CAGR to a $257.8 billion market in 2024. The size of this market can support both large private companies as well as meaningful growth opportunities for large enterprises.

AI research is leading to innovations in fields such as vision and natural language processing (NLP), which potentially answers the questions of enterprises that have struggled to implement AI in their systems. In Q1, 2021, public valuations of AI companies were not too optimistic as listed companies including C3.ai, Palantir, Sumo Logic, and Lemonade suffered declines. These companies’ growth outlooks remain strong, yet market enthusiasm has been tempered for innovative companies. A successive wave of AI special purpose acquisition company (SPAC) mergers also suffered, thus deflating expectations for future growth. In a shorter-term-focused market, AI startups may have to show revenue growth and software-as-a-service-style (SaaS) margins to achieve exits.

AI Core

This included the building blocks of AI & ML deployments such as the DevOps tools needed to build and deploy models to production. Sub-categories include AI as a service (AIaaS), AI & ML developer tools, AI platform as a service (PaaS), automated machine learning (autoML), cognitive computing, data preparation platforms, and quantum AI. The market seems to be heading towards data preparation, the enabling technology best served by startups. Hyperscalers with increasingly advanced cloud-native toolsets are supporting model training and deployment. According to a recent industry survey, two primary reasons for AI & ML project failures are lack of both data quality and production-ready data. While training frameworks are freely available, relevant and clean dataset cases are not—thus creating a need for startups to supply enterprises with data labeling services. To fill this gap, enterprise datasets currently require intensive manual labor—referred to as human-in-the-loop—to be used in AI & ML. 39.5% of data scientists surveyed by Appen spend more than 50% of their working hours managing, cleaning, and/or labelling their data. As a consequence, data preparation is a less crowded space than model training, with high demand.


The combination of marketing and technology to achieve marketing goals and objectives could be increasingly adopted by big firms and organizations aiming to cut costs and increase productivity. Today, recommender systems, digital marketing, conversational AI/chatbots are all prevalent on websites that offer a service for consumption. Wearable devices, IoT, sensor technology, Internet and website tracking cookies, and more help companies to collect vast amounts of data from everyday consumers which can then be used to understand consumer behavior better and to create new products and services. As privacy concerns continue to pick up steam, companies will be looking to find new avenues to pursue their marketing goals so they can continue to track consumer behavior.

AI & Cybersecurity

Using AI, algorithms can learn the ways of its user in order to decipher a pattern of behavior and normality. Once suspicious behavior is detected, it could either alert us or prevent the attacker from going further. This can be applied to a company or an individual user at home. People are now starting to adopt smart homes in which they can control daily tasks in their home using a digital assistant. Training AI algorithms to learn their user’s behavior can help prevent hackers from illegally gaining access to a person’s home. Using home devices is convenient but can also leave a person vulnerable to cyber-attacks, which is where AI can assist in mitigating such risks.

Healthcare AI

With an overwhelmed healthcare system, it means patients with other illnesses and diseases that require emergency services cannot receive the treatment they need. Using AI, hospitals and healthcare systems will be looking to automate certain tasks, such as triage and diagnosing patients, or evaluate medical records of their patients in order to best assess high risk individuals or those who may have something that was missed by previous office visits. This can limit exposure to disease, give priority care to those who need it most, and flag anomalies that can lead to better disease prevention, among other things. Radiologists and other medical professionals have already been using AI to help scan X-rays and MRIs to help find diseases and other problems. 2021 should find them leaning on AI more as accuracy rates continue to rise above what humans can see.


Company Market Map Fig 1.1 – Companies in AI & ML Industry Categorized by Different Sub-Asset Classes Bibliography

Pitchbook. “Artificial Intelligence & Machine Learning Q1 2021 VC Update,” n.d.
“Top 5 Artificial Intelligence (AI) Trends for 2021.” KDnuggets. Accessed July 12, 2021.https://www.kdnuggets.com/2021/01/top-5-artificial-intelligence-trends-2021.html .


© 2021 Verge Advisors. designed by Dibiz Solution