The Digital Revolution

June 11, 2018

‘The cloud’ is a buzzword that has been banded around for over a decade. Unlike many such words which can be seemingly overused and overhyped, ‘the cloud’ has become an integral part of doing business in a technology-driven, modern society. Indeed, when we see a company of Amazon’s size trading at a P/E multiple of 250, we must take note.

Amazon Web Service (AWS) is a significant factor in this valuation. Its cloud service offering has tripled the company revenue in just three years and whilst only accounting for one tenth of Amazon’s revenues, has a net profit margin in the region of 25%. Additionally, it appears that all of Amazon’s operating profit in 2017 was attributable to AWS.

It would therefore seem that a long-standing strategy of investment and capital expenditure is paying off. CEO of Amazon, Jeff Bezos, who is renowned for not spending time entertaining major stockholders, has ingrained this culture into the firm. He still attaches to shareholders annually the original letter when the company went public in 1997:

“We will continue to make investment decisions in light of long-term market leadership considerations rather than short-term profitability considerations or short-term Wall Street reactions.”

The company continues to retain its earnings in favour of expending it for future growth.

Cloud computing, in its essence, is the delivery of on-demand computing services. Initially, it was storage and processing power via a communication platform (typically the internet), which can be provided on a subscription or pay-as-you-go basis. Since, the concept has developed and the likes of AWS is no longer just a storage rental service, but a suit of around 100 services fulfilling a range of needs including deployment, analytics and developer tools.

Although widely used by large corporations, the impact of cloud computing on smaller enterprise is more significant, removing the capital requirement for hardware and software whilst offering security and business-orientated applications with lower up-front licensing costs. This provides an instant, competitive infrastructure. It is estimated that two thirds of US small/medium enterprise are AWS aligned.

Whether Amazon is the biggest provider of cloud services is difficult to determine. The other major corporation in the space is Microsoft, who do not disclose operating profit for its ‘Intelligence Cloud Unit’. IBM has been steadily transferring its expertise and technology to the cloud, and in doing so has become a significant player. As to whether IBM directly compares in size may be a discussion for the hard-core techies; as to whether or not the cloud is being overstated, and as to the exact meaning of business process as a service to which half their cloud revenues are attributed.

The commonly understood industry terms are ‘infrastructure as a service’ (IaaS), ‘platform as a service’ (PaaS) and ‘software as a service’ (SaaS). While Amazon is leading in IaaS, Microsoft has become a powerhouse in all three terms, and that is increasingly important as customers look for depth of offering to save on the potential costs involved in integrating of systems that were not designed to work together. Microsoft now focuses on end-to-end solutions to manage what it calls ‘the entire digital estate’.

Beyond the cloud, most big name tech companies are investing into Artificial Intelligence (AI). Even Intel, the semiconductor giant, has recently made a number of AI acquisitions. AI is a rather generalised term regarding the concept by which machines achieve tasks in a manner beyond being given a straightforward instruction. Most industry efforts today are focused towards machine learning.

Machine learning involves an algorithm that is able to adjust itself to better achieve an outcome. This outcome is typically in the form of classifying information. Essentially, based on the information fed to a system, data is classified and a feedback loop informs the system whether the output is correct so it can adjust metrics to achieve a higher probability of success in the future.

A common example of this is recognising objects in images, which requires huge amounts of data and significant human involvement to tag objects in images as the feedback. Machine learning is most noticeably being used to direct targeted advertising to relevant consumers online, but it is beginning to have an impact in other ways.

In healthcare, diagnosis is largely a system of matching symptoms and probabilities, meaning algorithms could play a major role in processing pattern information to assist their human counterparts. In data security, it is already being used to spot malware files with great accuracy, and can analyse how data is accessed to report anomalies that predict data breaches.

Machine learning is being implemented to suggest media that may be of interest to the consumer, a convenient function when consumers are required to trawl through seemingly endless choices to find their next digest. Then, of course, there are driverless vehicles, which have progressed at an exciting rate.

The list is endless, but those highlight a few of the themes that have been piquing the interest of investors in recent years. So, what is it about AI that might make a tech stock attractive? With such an array of opportunities, cutting through the noise is a difficult task and successful outcomes do not necessarily indicate quality solutions! However, a few key factors warrant our attention:

Data: AI becomes more effective with the greater amount of data it can analyse. Companies with large user bases, interaction with people’s habits and feedback information as part of the service are well placed to capitalise. It may go some way to explain Facebook’s success and in addition, the likes of Amazon will have a wealth of information on consumer spending habits.

The protection of personal data has become a cause for concern this year. Following the news of Facebook’s data breach, more attention is being paid to how our personal information is being used. This, in addition to the new GDPR legislation which has recently been rolled out across the EU, means that the use of big data could become a minefield requiring careful navigation in the years ahead.

Natural Language Processing (NLP): NLP attempts to understand human communication in greater depth with the aim of being able to communicate back in a similar, comprehensible language. Systems with natural language can stand in at the point of contact and direct customers to the information or locations they require. It can also be used to translate or condense large volumes of information, achieving efficiency in this data-heavy generation. Again, access to data here is key; Amazon’s ‘Alexa’ and Apple’s ‘Siri’ seem to have the market lead, but Google’s ‘Home Assistant’ and Microsoft’s ‘Cortana’ are emerging participants.

Deep Learning: Deep learning is considered the cutting edge of AI, and is just one of the many approaches to machine learning. This progressive step involves the intertwining of existing machine learning components in order to create artificial neural networks. This creates a layering effect where each layer has its own purpose and functionality.

This complexity is giving rise to an increasing number of joint ventures in the AI space and these partnerships may become key to success. Traditionally closed companies, such as Apple, may find it difficult to achieve the diversity required for successful neural networks. However, protected intellectual property and the resulting ability to monetise such advances may prove to work in their favour.

The other option to joint ventures would be to acquire, and large tech companies have been involved in a flurry of such activity in recent years. Google has acquired twelve AI start-ups in four years; Apple four in two years; whilst IBM, possibly a sleeping giant after being a leader in the field since the 50’s, has recently purchased three. Apple also poached Google’s AI chief in April.

The digital revolution has so far been defined by an explosion in the amount of stored information and data. The next stage will be the defined by how that data is used and processed.

Continue reading