How Might AI Demand Affect Edge Computing?

Industry analysts and data center operators--also including enterprise customers--seemingly continue to evaluate the costs of adding edge computing capabilities to their existing data center architectures, resulting in arguably-slower market growth than some had predicted and expected. 

And now there is the additional consideration of how support for artificial intelligence operations “as a service” could affect edge computing as well. And it might be fair to say that, at the moment, AI requirements might not boost prospects for edge computing all that much. 

In fact, one might make an argument for a barbell type growth pattern, with AI increasing demand for remote processing on one hand and “on-the-device” on the other hand. Remote processing might still be the best option for heavy processing and storage requirements, while many latency-dependent operations will be increasingly handled right on the device. 

The arguments around edge computing “as a service” have always focused on use cases requiring very-low latency, with additional value from less capacity network demand and possibly some privacy and security advantages. For enterprise computing in general, processing power has been a concern when evaluating private and cloud computing alternatives. 

But cost has been a growing issue as “cloud computing as a service” volume and scale have grown. 

But on-the-device and on-the-premises private processing alternatives always are a possibility. And remote processing always is an alternative for non-real-time use cases and needs. 

So the additional new requirements for AI “as a service” still hinge on latency and processing power requirements. 

Some might attribute the present state of thinking around edge computing as more focused on cost-benefit and payback than hype, with less of the “either/or” rhetoric around edge computing as a “replacement” for remote cloud computing some espoused early on. 

Forecasts of revenue growth have varied significantly, looking at service revenue, capital investment, endpoints supported or processing capacity. 

Firm	Year	Metric	Projected Growth
BCC Research	2022	Global Edge Computing Market (USD Billion)	$60.0 to $110.6 (2024-2029)
Grand View Research	2023	Global Edge Computing Market Size (USD Billion)	$16.45 to $36.9 (2024-2030)
MarketsandMarkets	2023	Edge Computing Market Size (USD Billion)	$60.0 to $110.6 (2024-2029)
Gartner	2022	Edge Computing Endpoints Worldwide (Billions)	1.25 to 25 (2021-2027)
IDC	2022	Global Edge Computing Infrastructure Spending (USD Billion)	$17.9 to $176.1 (2020-2025)
McKinsey & Company	2018	Edge Computing Adoption Rate (%)	30% of enterprises to adopt edge by 2025
STL Partners	2019	Edge Computing Market Trends	Global market growth estimated at a CAGR of 34% during the period 2019-2025.

There's a strong possibility that AI processing will follow a barbell pattern, with the majority of processing happening on devices or at remote data centers, leaving edge processing with a smaller role. Here's a breakdown of the reasoning:

On-device processing will be preferred or necessary for real-time use cases, including facial recognition; augmented reality; image processing and speech interfaces. Privacy issues also are addressed when processing happens on the device and not at a remote data center of some type. 

Remote computing will still make sense when processing power and storage are key requirements, but not real-time response. 

Also, centralized management and control will have advantages for processing-intensive or data-intensive operations, or in instances where collaboration with many other entities is required. 

So one might postulate that edge computing will have value in between the on-device and remote cloud processing ends of the barbell. 

Edge devices (sensors, network gateways) often have limited processing power and battery life, which restricts their ability to handle complex AI tasks, for example, but might also have latency requirements. 

Also, in some instances, connectivity impediments or cost might make local processing attractive. 

So AI processing might follow a barbell pattern, prioritizing on-device processing for applications requiring low latency and prioritizing user privacy. Remote data center processing will be essential for training complex AI models, managing large datasets, and facilitating collaboration.

Edge computing has to provide value someplace in-between. And that is part of the business value issue. As often is the case, products have clear value at the high-end and low-end parts of any market. In the middle is where balances have to be struck.

Video is Becoming a Hybrid

One hears a lot of talk about how different broadcast TV content strategies are from video streaming content strategies. The argument is that broadcast TV and video streaming services catered to different audiences, with distinct revenue models. Some of us cannot see that continuing. 

At a high level, the consistent trend in “video” markets has been in the direction of personalization or customization, on-demand consumption and a shift to viewing of titles rather than “channels” that mirrors and is part of the same trends in nearly all forms of media. 

The key implication is a shift of most of the business towards personalized, niche, custom or more-specialized audiences rather than the broad-appeal audiences the broadcast TV networks have required. 

Where broadcast TV has relied on advertising revenue generated by content pitched to large audiences with content of wide appeal, including sitcoms, dramas, and reality shows, streaming has been characterized as a subscription model, with content often aimed at niche audiences often featuring genre-specific shows, documentaries and foreign language content.

But those generalizations are breaking down. Already, video streaming services are moving to ad support plus live (scheduled) programming, especially sports. And that blurs the lines between “broadcast” and “streaming” content and revenue models. 

The example one always hears is that “eventually, the Super Bowl will be on streaming,” the point being that the National Football League consistently generated the top “live” programming events, with advertising being the revenue driver. 

Were NFL events to shift to streaming distribution, so would audiences and revenue models follow the shift. The point is simply that streaming substantially substitutes for broadcasting in terms of viewership and revenues.

And most predict that the role of broadcast programming will also shift over time as a result, becoming a smaller part of the total content revenue stream and also therefore shifting its role. Many expect a shift to substantially more reality TV as the broadcast staple.

Others also expect the broadcast role to shift more towards the role theatrical exhibition has had for video markets, namely building interest and word of mouth for eventual on-demand or pre-recorded media release. 

In that sense, the content distinctions between broadcast and streaming services might be less important. Perhaps it is the role of broadcast TV that changes, in the direction of the role theatrical release now has for movies: building word of mouth for eventual on-demand exhibition. 

Beyond those changes, one might note a generalized switch in business models in many industries from one-time purchase to subscriptions, enabled by online and internet platforms that enable them. 

The other obvious change is from one-time purchase to on-demand purchase and consumption. 

AI might affect on-demand and one-time-purchase models using personalization. 

But the impact on subscription business models is less clear, except that subscription payment models often are a way to support on-demand consumption. Streaming video services and e-commerce subscriptions such as Amazon Prime provide examples. 

“U.S. consumers may be willing to endure more ads and be corralled into bundles, but they may not be willing to give up the degree of customization they gained from unbundling pay TV, or the personalization they enjoy from social media, say consultants at Deloitte. 

In other words, streaming “subscriptions” basically are favored because they provide on-demand viewing, much as earlier generations of technology (VCR tape rentals, DVD rentals) enabled on-demand, personalized consumption of content.

Linear broadcast bundles have tried to provide choice by providing many genres and content formats, even if broadcast schedules are part of legacy media formats requiring consumers to adapt to provider schedules. 

Streaming subscriptions also are a platform for custom and personalized choice, allowing consumers to watch what they want, when they want it, and not on a fixed schedule. Digital video recorders, in essence, are an attempt by linear video providers to create some amount of on-demand consumption. 

To a large extent, streaming also is favored because it is a “multi-screen” format, allowing native consumption on phones, tablets, gaming devices or televisions. Linear TV providers replicate that feature using their own proprietary streaming services. 

So the argument that content strategies for linear broadcast and streaming are “different,” as much sense as that has made, arguably is changing. “Appointment television” now consists largely of sports or other events, but a slow movement of such content to streaming is happening. 

As linear audiences shrink, so will the necessity of assuming large audiences are the target (“large” is a relative term, of course, looking at streaming content audiences by title instead of TV networks). In fact, we already can see that many linear TV series that did not do so well on linear services can become “hits” on streaming services. 

And the same can hold for movies that did not perform so well theatrically. 

The point is that the same “big audience” formats thought to be essential for linear broadcast also often make sense for supposedly-niche streaming services. So a blurring of lines and strategies is taking place. 

Such a hybrid model is found in other industries as well. Private computing coexists with cloud computing. The rise of cloud computing hasn't eliminated the need for on-premise data centers, nor have entities stopped buying their own computing gear in preference to renting computing cycles and storage. 

Streaming services such as Spotify and Apple Music have become popular. However, vinyl records and physical CDs continue to be produced and collected by enthusiasts.

The rise of e-readers and audiobooks hasn't eradicated traditional printed books. Movie theaters still have a role in the content value chain. 

People use online shopping but also visit stores. They travel but also use videoconferencing. You get the point: a hybrid video model is coming.

AI Will Produce Winners and Losers

Though many executives and analysts are trying very hard to figure out which firms benefit most from generative artificial intelligence and AI in general, the prior experience of firms with the internet suggests there also will be losers.

And those losers could come from industries focused on digital and physical products such as “print” media, as our experience with the internet suggests. 

Study Title	Authors	Year	Focus	Key Findings
How the Internet Changed the Market for Print Media	NBER	2019	Impact on print media	Household adoption of broadband significantly reduced print readership and circulation, leading to revenue decline for newspapers.
The Impact of the Internet on Media Industries: An Economic Perspective	Oxford University Press	2008	Economic impact on media	The internet weakens intellectual property protection, making it easier to distribute content illegally and reducing potential revenue.

Similar losses can be noted in retailing as well, with a shift from place-based and physical retail to online retail. 

Study Title	Authors	Publication Year	Key Findings
The Impact of E-Commerce on Retail Employment	Autor, D., Dorn, D., Hanson, G.	2017	Found that increased e-commerce adoption led to job losses in retail sectors most susceptible to online substitution (e.g., electronics).
Omnichannel Retailing and Customer Engagement: A Review of the Literature	S. Verhoef, M. Kannan, P. Bharadwaj	2009	Highlights the importance of omnichannel strategies for retailers to enhance customer engagement and satisfaction in today's digital age.
The Impact of Online Shopping on Brick-and-Mortar Stores	T. Van den Poel, R. Verhoef	2003	One of the earlier studies exploring the potential negative impacts of e-commerce on traditional brick-and-mortar retailers.

Advertising has seen some of the greatest shifts from the internet, though.

Source: Gemini

Put simply, digital now claims up to 82 percent of all U.S. ad placements and revenue. Print has declined from 42 percent to less than three percent. Linear video dropped from 38 percent to 16 percent. Radio dipped from 10 percent to half a percent. 

Channel	1996 (Billions)	1996 (%)	2023 (Billions)	2023 (%)
Print (Newspapers, Magazines)	80.0	42.1%	10.0	2.7%
Linear Video (TV Broadcast, Cable)	72.0	37.9%	60.0	16.2%
Network Radio	10.0	5.3%	2.0	0.5%
Other (Radio Spots, Out-of-Home)	28.0	14.7%	18.0	4.9%
Digital Ads (Search, Social Media, Display)	-	-	300.0	81.7%

It might be reasonable to expect the content industries, advertising and retailing will again be among the industries to see early AI disruptions. 

Financial services might also be included on the list of industries that saw early internet disruption, and might see further challenges from AI. More recently, various forms of “sharing” (transportation and lodging, for example) also have emerged, and might see further changes from AI. 

But manufacturing and pharmaceuticals seem poised for AI disruption as well. On the other hand, construction might see relatively low amounts of disruption. 

Industry	Potential AI Impact	Drivers
Manufacturing	High	Robots can handle repetitive tasks, improve precision, and optimize production processes. AI can also be used for predictive maintenance and quality control.
Transportation	High	Self-driving vehicles, logistics optimization, and automated traffic management are all powered by AI.
Customer Service	High	Chatbots and virtual assistants can handle routine inquiries, freeing human agents for complex issues.
Finance	High	Algorithmic trading, fraud detection, and risk assessment can be significantly enhanced with AI.
Healthcare	High	AI can assist in medical diagnosis, drug development, and personalized medicine.
Retail	Medium	AI can personalize recommendations, optimize inventory management, and automate tasks like pricing and promotions. However, the human element in customer service and product selection might remain crucial.
Legal	Medium	AI can analyze legal documents, predict case outcomes, and streamline research tasks, but human judgment will likely remain essential for legal proceedings.
Education	Medium	AI-powered tutors can personalize learning experiences, but human teachers will likely remain central for guidance and social interaction.
Media & Entertainment	Medium	AI can personalize content recommendations and automate content creation tasks.
Construction	Low	Manual labor and on-site decision-making are still crucial aspects of construction, making widespread AI adoption less likely. However, AI can be used for design optimization and project management.
Hospitality	Low	The human touch remains essential in hospitality, but AI can automate tasks like booking and guest communication.
Arts and Culture	Low	Human creativity and emotional connection are central to the arts, making AI unlikely to replace artists entirely. However, AI can be used for artistic exploration and content creation tools.

Right now, attention is logically focused on industries and functions that are susceptible to AI automation. But equally big changes could come if AI allows competitors to enter markets in new ways. Think ridesharing and peer-to-peer lodging. 

And there always is the possibility that new industries are born. Think search and social media.

By Definition, Most Enterprise Use Cases for Generative AI Involve Content

Enterprise use cases for generative AI, given that GenAI is a content creation engine, all center on various forms of content, according to the Stanford University Human-Centered AI institute. 

source: Stanford University Human-Centered AI institute 

That noted, AI use cases are arguably far more numerous. 

source: Stanford University Human-Centered AI institute