AI Increases Data Center Energy, Water E-Waste Impact, But Perhaps Only by 10% to 12%

An argument can be made that artificial intelligence operations will consume vast quantities of electricity and water, as well as create lots of new e-waste. But higher volume of cloud computing operations--for conventional or AI purposes--is going to increase in any event.

And higher volume necessarily means more power and water consumption, and use of more servers. 

Some portion of the AI-specific investment would have been made in any case to support the growth of demand for cloud computing. 

So there is a “gross” versus “net” assessment to be made, for data center power, water and e-waste purposes resulting from AI operations. 

By some estimates, AI will increase all those metrics by 10 percent to 12 percent. It matters, but not as much as some might claim. 

By definition, all computing hardware will eventually become “e-waste.” So use of more computing hardware implies more e-waste, no matter whether the use case is “AI” or just “cloud computing.” And we will certainly see more of both. 

Also, “circular economy” measures will certainly be employed to reduce the gross amount of e-waste for all servers. So we face a dynamic problem: more servers, perhaps faster server replacement cycles, more data centers and capacity, offset by circular economy efficiencies and hardware and software improvements. 

Study Name	Date	Publishing Venue	Key Conclusions
The E-waste Challenges of Generative Artificial Intelligence	2023	ResearchGate	Quantifies server requirements and e-waste generation of generative AI. Finds that GAI will grow rapidly, with potential for 16 million tons of cumulative waste by 2030. Calls for early adoption of circular economy measures.
Circular Economy Could Tackle Big Tech Gen-AI E-Waste	2023	EM360	Introduces a computational framework to quantify and explore ways of managing e-waste generated by large language models (LLMs). Estimates annual e-waste production could increase from 2.6 thousand metric tons in 2023 to 2.5 million metric tons per year by 2030. Suggests circular economy strategies could reduce e-waste generation by 16-86%.
AI has a looming e-waste problem	2023	The Echo	Estimates generative AI technology could produce 1.2-5.0 million tonnes of e-waste by 2030 without changes to regulation. Suggests circular economy practices could reduce this waste by 16-86%.
E-waste from generative artificial intelligence"	2024	Nature Computational Science	Predicts AI could generate 1.2-5.0 million metric tons of e-waste by 2030; suggests circular economy strategies could reduce this by up to 86%1 2
"AI and Compute"	2023	OpenAI (blog)	Discusses exponential growth in computing power used for AI training, implying potential e-waste increase, but doesn't quantify net impact
"The carbon footprint of machine learning training will plateau, then shrink"	2024	MIT Technology Review	Focuses on energy use rather than e-waste, but suggests efficiency improvements may offset some hardware demand growth

The point is that the specific impact of AI on energy consumption, water and e-waste is significant. But the total data center operations footprint is not caused solely by AI operations. Computing cycles would have grown in any case. 

So we cannot simply point to higher energy, water and e-waste impact of data centers, and attribute all of that to AI operations. 

Measure	Total Data Center Impact	AI Workload Contribution	AI as % of Total Impact
Energy Consumption	~200-250 TWh/year globally	~20-30 TWh/year	~10-12%
Water Consumption	~600-700 billion liters/year	~60-90 billion liters/year	~10-13%
E-Waste Contribution	~3.5-4 million metric tons/year	~350-500 thousand tons/year	~10-12%

If 90% of Your Business Model is Threatened, You Must Change

It’s fairly easy to explain why cable TV programming networks face an existential crisis. Shrinking audiences attack 85 percent to 90 percent of the revenue base. 

Their revenue is driven by advertising (perhaps 40- to 45 percent of total) and affiliate fees paid by distributors (about 50- to 55 percent). The former is dependent on the volume of viewers, which is declining every year. The latter is dependent on the number of potential viewers a distributor can deliver (subscribers). 

Both are essentially “potential attention” metrics and are declining. Subscriber volume provides potential larger audiences while also making any particular distributor more or less valuable as a platform for such potential audiences. 

If subscription decline is irreversible, then so are advertising and affiliate fee revenues. By some estimates, video streaming revenue already has surpassed the level of linear video subscription revenue. 

But most content business models also have been disrupted by the internet and digital content distribution. So far, most of the angst about artificial intelligence has been on its potential to disrupt content industry jobs by automating the functions. 

Industry	Traditional Business Model	Disrupting Technology/Trend	New Business Model
Cable TV	Subscription-based service for linear channels	Streaming services (Netflix, Hulu, etc.), cord-cutting	Subscription-based streaming services, ad-supported streaming, live TV streaming
Long Distance Calling	Per-minute charges for long-distance calls	Voice over IP (VoIP) technology	Flat-rate long-distance plans, VoIP services
Local Telephone Service	Landline phone service with monthly fees	Mobile phones, VoIP services	Mobile phone plans, VoIP services
Postal Services	Physical delivery of letters and packages	Email, digital messaging, online shopping	Digital delivery of mail, specialized shipping services
Retail Shopping	Physical stores, in-person shopping	E-commerce, online marketplaces (Amazon, eBay)	Online shopping, omnichannel retailing, subscription boxes
Music Distribution	Physical media (CDs, vinyl records)	Digital music distribution (iTunes, Spotify)	Digital music streaming, music subscription services
Newspapers	Print newspapers, subscription-based model	Online news, digital subscriptions	Online news, digital subscriptions, ad-supported news
Magazines	Print magazines, subscription-based model	Digital magazines, online content	Digital magazines, online content, ad-supported content
Movie Theaters	In-theater movie screenings	Streaming services, video-on-demand	Streaming services, video-on-demand, premium video-on-demand
Home Video	Physical media (DVDs, Blu-rays)	Digital video rental, streaming services	Digital video rental, streaming services, video-on-demand

While an apt observation, it is the disruption of existing business models which will have the greater impact. It might be easy to suggest that AI will restructure many business processes in content industries. What is harder to ascertain is the ultimate impact on business models.

The mere fact that AI creates content instead of humans does not necessarily disrupt the revenue model, only changing the cost model. Distribution could be an area of greater change, though, as most content industry disruptions are founded on distribution (display) technology changes. 

Television disrupted movie theater exhibitions. Cable TV disrupted over-the-air broadcasts. The VCR enabled the home video business, the DVD improved it; video streaming ended it. Audio tape, then compact discs largely displaced the records business. But music streaming largely displaced all physical media distribution. 

Industry	Traditional Business Model Characteristics	Potential AI Disruption	New Business Model Characteristics
Content Creation	Human creators, editors, and producers	AI-generated content, automated editing, and production	AI-assisted content creation, personalized content experiences, dynamic content adaptation
Content Distribution	Traditional media channels, physical distribution	AI-powered personalized content delivery, direct-to-consumer distribution	AI-driven content recommendation systems, dynamic pricing, real-time content adaptation
Content Consumption	Passive consumption, linear storytelling	Immersive experiences, interactive storytelling, personalized content	AI-powered interactive content, virtual and augmented reality experiences, personalized content journeys
Copyright and Intellectual Property	Traditional copyright laws, licensing fees	AI-generated content ownership, copyright infringement challenges	AI-powered copyright management systems, dynamic licensing models, new revenue streams from AI-generated content
Advertising	Traditional advertising models (TV, print, digital)	AI-powered targeted advertising, programmatic advertising	AI-driven personalized advertising, dynamic ad placement, brand experiences powered by AI
Education and Training	Traditional classroom-based learning, standardized content	AI-powered personalized learning, adaptive content	AI-driven learning platforms, personalized tutoring, skills-based learning
Entertainment	Traditional entertainment formats (movies, TV shows, games)	AI-generated entertainment, interactive gaming, virtual reality experiences	AI-powered interactive entertainment, personalized gaming experiences, virtual and augmented reality content

The cumulative impact of digital media and internet distribution has enabled a shift to on-demand consumption and away from linear formats across virtually all media. That, in turn, has altered the advertising and affiliate revenue basis for linear video subscription services. 

U.S. subscriptions, for example, peaked around 2010 and have been steadily declining since then. 

source: IBISWorld 

Though there are many reasons for the lower demand for linear video subscription services, the threats to advertising and affilate fee revenue streams are the most obvious.

Challenges for Linear TV & Cable TV	Advantages of Streaming Alternatives
Rigid Scheduling: Viewers must watch programs at set times.	On-Demand Viewing: Content is available anytime, allowing flexibility.
High Subscription Costs: Cable packages are expensive and include unwanted channels.	Lower Cost Options: Streaming services offer affordable subscription plans and bundles.
Ad Interruptions: Traditional TV has frequent and lengthy ad breaks.	Ad-Free or Limited Ads: Streaming platforms offer ad-free tiers or fewer interruptions.
Lack of Personalization: Limited ability to customize programming to individual tastes.	Personalized Recommendations: AI algorithms suggest content based on viewing habits.
Limited Mobility: Watching requires a TV or cable box at a fixed location.	Cross-Device Access: Streaming is available on phones, tablets, smart TVs, and laptops.
Outdated Content Delivery: Programming is tied to fixed schedules, making discovery difficult.	Content Libraries: Vast libraries of current, past, and exclusive content are accessible anytime.
Declining Viewer Engagement: Younger audiences are abandoning traditional TV.	Appeals to Younger Audiences: Streaming platforms cater to younger demographics with interactive and diverse content.
Complicated Bundles: Bundled cable TV packages force customers to pay for unwanted channels.	A La Carte Options: Consumers can subscribe only to the services or shows they want.
Geographical Restrictions: Content availability is often tied to specific regions.	Global Accessibility: Streaming platforms offer consistent access to content worldwide.
Slower Technology Adoption: Traditional TV struggles to incorporate new technologies quickly.	Cutting-Edge Technology: Features like 4K, HDR, offline viewing, and voice control enhance the experience.