As digital transformation accelerates, we progress from a decade of cloud adoption into an age of artificial intelligence (AI) and machine learning (ML), powered by the cloud.

Years of cloud service development, the emergence of a DevOps methodology, big data, and the adoption of public cloud, has created a foundation for AI to now effectively permeate enterprise development. The cloud has progressed from a less expensive method of storing data and operating applications to becoming an integral part of the deliverance of AI and ML.

Essential for data storage and compute power

AI and ML require massive amounts of data to model their findings, learn outcomes and responses, and improve predictions. They also need colossal amounts of compute power to operate algorithmic processes and model data. A combination of CPUs and GPUs is needed, and cloud providers are responding by offering GPU-backed VMs and containers on a pay-per-use model more effective for AI users. And, cloud providers are beginning to deliver AI as a Service, APIs, and tools for data scientists and developers which take advantage of the infrastructure they provide thus accelerating cloud reliance.

In turn, the easy availability of cloud-based AI provision without huge infrastructure investment is accelerating the adoption of AI. The public cloud model enables businesses to purchase the right amount of cloud storage needed for their ventures and add to it whenever they need to. Gone are the past investments into on-premise infrastructure and massive physical data centres by large corporations. Thus, freeing more capital to invest in developers and their projects, and perhaps culminating in savings for the end-customer.

A maturing cloud services industry will propel the emergence of AI

The public cloud industry already reached a value of $182.4 billion by 2018 and was predicted to hit a value of $214.3 billion by 2019. Indicative of its relative infancy, the market for AI was worth $21.46 billion in 2018 but could grow to a value of $190.61 billion by 2025.

Tractica predicts that AI will contribute to as much as 50% of cloud services revenue by 2025. And, Accenture estimates that 85% of business and IT leaders will invest in one or more AI-related technologies in the next three years.

A Deloitte study found 49% of companies that have already deployed AI in some way, are using cloud-based services. David Schatsky, managing director at Deloitte LLP says:

“Cloud adoption is motivating enterprises to undertake more proofs of concept in their firms with AI because it’s easier than ever before to get started.”

Schatsky believes this route is attractive as cloud providers develop their own AI services for businesses which don’t need these organisations to invest heavily in their own compute power, instead they can purchase it as they need it, in the cloud.

AI expertise combined with flexibility, adaptability and scalability

The shortage of AI skill sets is an obstacle for enterprise adoption but its also answered by cloud providers whose expert teams are developing applications, algorithms, APIs and solutions which can be easily utilised or adapted by their cloud clients. A progression to widely available open-source software means that developers can pull from multiple sources to build their cutting-edge and enterprise projects faster and cheaper than ever before.

To contrast with private cloud and on-premise infrastructure, public cloud services are much more flexible. Platforms and their users can use the public cloud to scale with market demand. Considering the potential growth of AI utilisation and revenue over the next few years the ability to move with this market is invaluable to businesses wishing to maximise the potential of AI. Arguably, public cloud services offer better security for the storage of sensitive big data needed for AI modelling.

AI promises greater automation across many business processes, more powerful devices, infrastructure and even environments, better customer service delivery, improved decision-making capabilities, more capable cybersecurity defences and faster medical diagnosis in healthcare. And, this is just the beginning of cognitive computing.

For AI to deliver its technological commitment to industry and global economies, it requires a scalable foundation of compute power and data storage that is cost effective. Public cloud services offer the underlying, flexible, and adaptable infrastructure required for AI’s complete ratification.

Enterprises and developers can innovate directly from the cloud using a combination of available, adapted, and created technologies, with almost limitless compute power allowing innovation to involve into a complete realisation of the potential of AI.

Ulrike Eder, COO at drie.co