As the battle for market leadership in Artificial Intelligence heats up, it is yet to be clear what designs will dominate. However, lessons from history may help us back a winner.

Xavier Ferràs

If the winner often takes it all in digital industry market share, who’s going to walk away with the AI prize? 

The competition is fierce, with giants Amazon, Microsoft, Apple, Google, Facebook, NVIDIA and Oracle leading the way. But who will be the VHS victor, and who’s resigned to Betamax history? It’s difficult to assess which company is emerging as producing the dominant design without the benefit of hindsight; Nokia phone fans may testify. 

But research from Xavier Ferràs-Hernández (Esade), Petra A. Nylund (University of Stuttgart) and Alexander Brem (University of Stugggart and University of Southern Denmark), published in the California Management Review, has shed light on the dynamics of AI in business and the potential frontrunners.  

By examining previous managerial frameworks and developing a theoretical synthesis based on a comprehensive literature review, the researchers have identified the designs and business models they predict will dominate the AI market. They also have important recommendations for firms and managers on how to generate competitive advantages. 

Trillions of dollars at stake

Technological change drives economic growth, and AI is the most significant technological development for a generation. An analysis of over 400 application cases across 19 industries suggests AI will generate between $3.5 and $5.8 trillion every year, so it’s no surprise that the heat is on. However, it’s not yet clear how value chains will form or what the final dominant designs will be.  

AI is expected to generate between $3.5 and $5.8 trillion annually

History has shown us that the best technical designs aren’t necessarily those that are adopted by the market. A dominant platform that can be adopted by the majority of the ecosystem requires a balance between open architecture and closed innovations. Closed standards, systems that are too specific or customized, or those which require greater resources, all act as hurdles in the race. Service and business model innovation also play key roles in the emergence of a market leader. 

Disrupting markets

Disruptive technology, as the name suggests, shakes up markets. In its early days there are multiple configurations at play, each targeting different niches. Supply chains lurch between vertical and horizontal modular architectures until a dominant player emerges. 

Vertical industries focus on key competencies and don’t prevent competitors from entering the supply chain. Modular industries seek efficiencies and control through verticalization strategies, with key competencies such as processors, R&D and talent all generating incentives for vertical expansion. 

This modularized configuration is where AI currently stands, with no final dominant architecture. An integration of the leading figures and their processes could see the final victor start to emerge. 

Honing the product

The AI market is currently fluid and pre-mass market infiltration, with no clear winner evident just yet. Experience shows us that, in this phase, start-ups jostle for position with established corporations.  

This period of instability ends with the appearance of a design favored by the market, just as IBM did with the personal computer market and Apple with its still-ubiquitous iPhone. The focus then moves onto honing this design, shifting from product innovation to process innovation.  

Providing AI through the cloud is a clear business area for a mass market

For AI, design competition is still in its infancy. It’s expected to eventually reach the final consumer with new, more sophisticated human-machine interface systems that have increasing social capabilities and strategic thinking skills.  

But although it’s early days, the analysis by Ferràs and co-authors points to one clear business area for a mass market: providing AI through the cloud

Design frontrunners

In the late eighties and early nineties, as the PC market went mainstream, Intel won the battle of the chip manufacturers. Its ‘Intel Inside’ campaign saw it propelled to prominence to become the go-to central processing unit for an emerging market of computer manufacturers. 

In the same way, the researchers predict that AI supercomputing facilities will be provided by market leaders. Retailers, manufacturers and service providers will connect, via terminals, to future iterations of image, facial and speech recognition, writing generation and strategic thinking.  

These cloud supercomputing facilities powered by advanced GPUs (graphics processing units) will enable the provision of AI as a service, according to Ferràs-Hernández and his co-authors. The result will be a scenario where consumer goods are powered by the AI equivalent of ‘Google Inside’, ‘Amazon Inside’ or ‘Microsoft Inside’.  

These household names, along with Facebook and Chinese tech giant Baidu, have already opened the code to develop AI applications for third parties through open-source software. This move has generated a horizontal market of external developers, in much the same way as the Intel CPU opened the door for the mass production of personal computers.  

In the AI revolution, Intel still holds a significant market share - but NVIDIA is currently leading the GPU race and calls itself the world leader in AI computing. 

AI as a cloud service

This software-as-a-service market (SaaS) or, in this case, AI as a service (AIS) — through the cloud, seems to be the emerging business model. Amazon Web Services (AWS), with specific AI libraries, is worth $130 billion and leads the global cloud market with a 32 percent share. Microsoft’s Azure, with more than 20 cloud-based cognitive services, holds 20 percent. Google Cloud, with its libraries of applications for image processing, natural language allocations, speech recognition and intelligent data searching, amongst others, has a nine percent share. 

The global cloud market will grow to $1,026 billion by 2026

Apple currently favors a ‘final user device’ strategy, with its cloud capacities only available on its own devices. NVIDIA also offers cloud solutions, but provides its machine intelligence through partner cloud platforms including Google, IBM, Microsoft and Amazon. 

With a predicted annual growth rate of 18 percent, the global cloud market is set to grow to $1,026 billion by 2026

Skill up to embrace AI data

Where emerging technologies lead, the job market follows. The AI engineering job market grew by 344% between 2015 and 2018 in the US. Amazon has already invested $228 million in talent specializing in AI, followed by Google with $130 million and Microsoft with $75 million.  

And along with the need for suitably qualified and experienced specialist engineers, managers will have to ensure training and development at an organizational level to be able to utilize the competitive advantages AI opens up. Preparing a company and its people for this doesn’t happen overnight.  

AI as a service provides managers with the ability to generate knowledge gained from experience that isn’t tainted by human bias, and the potential applications are limitless. Data strategies will be propelled to the forefront of operations and decision-making, with vastly improved predictions and perspectives. 

All this will be facilitated by the acquisition and storage of big data, and the design of the IT frameworks required to manage it. 

If (or, more likely, when) AI is commoditized, the quality of available data will dictate which companies gain a competitive advantage, and which will be left behind. 

All written content is licensed under a Creative Commons Attribution 4.0 International license.