Short AI Videos: Do You Really Want to Watch Them?

Artificial intelligence threatens disruption of all sorts of industries and firms. Consider the threats to video sharing sites such as Instagram, which historically has featured user-generated short videos. 

What happens when users are exposed to a flood of AI-generated videos? For many users, a loss of interest is likely. Some users will not want to watch synthetic video resembling short animated videos, videogame style content or other scripted content. 

The “charm” has been real life, captured in short videos, often humorous or otherwise unusual. That interest goes away, for many, if the content is simply short, scripted, imaginary video. A video of a funny mishap is one thing. The same scripted mishap might not be so compelling. 

An amazing human or natural occurrence is one thing. A staged, produced, imagined short story might not have the same compelling nature. User-generated content often feels authentic: raw and genuine, reflecting real-life experiences, emotions, and imperfections. 

AI content will have an authenticity challenge. 

The corollary is that the audience shrinks, which attacks the ad revenue model. 

Threat	How it Works	Implication
Content floods	Mass AI video generation overwhelms feeds	Erodes trust, engagement, and quality drop
Distribution shifts	AI lets creators bypass platform algorithms	Reduces Instagram’s gatekeeping power
Behavior change	Personalized AI feeds displace static social feeds	Users migrate to new experiences
Ad disruption	Ads embedded in AI content or custom feeds	Instagram loses ad market share
AI-powered competitors	New platforms born AI-native	Instagram risks being out-innovated
Deep fake crisis	Synthetic humans and events proliferate	User trust, safety, and brand value erode

AI Might Not Diminish Critical Thinking, But Vested Interests Often Do

One sometimes hear it argued that fewer homes will "get internet" because of changes to Broadband Equity, Access, and Deployment Program rule changes. One also hears arguments that increased use of artificial intelligence will reduce critical thinking skills. 

Sometimes those arguments are highly questionable. There are other reasons why reality, truthfulness or factuality can be challenged, and it has nothing to do with human critical thinking or using AI. Instead, the issue is vested economic interest. 

Advocates for local or state government, for example, have a vested interest in increasing the share of federal resources they can deploy to solve problems. And sometimes they have vested interests in particular ways of solving problems. 

Consider arguments for how to bring better home broadband services to rural areas. For decades, the preference has been for a particular solution, namely optical fiber to the home, with opposition to using other arguably more-affordable and immediately-deployable solutions including satellite service and using mobile networks rather than cabled networks. 

Nobody disagrees that optical fiber to the home is the most “future proof” solution, providing it is economically feasible. The problem is that feasibility often is precisely the issue. 

FTTH Deployment Environment	Typical Homes Passed per Mile	Cost per Mile (All-In)*	Cost per Location (Homes Passed)	Key Cost Drivers
Urban (High Density)	80 – 150+	$50,000 – $100,000	$500 – $1,200	Shorter drops, existing duct/conduit, shared trenching, many users per mile
Suburban (Moderate Density)	30 – 70	$40,000 – $80,000	$1,200 – $2,500	Mix of aerial and buried, moderate trenching cost, fewer homes per mile
Rural (Low Density)	5 – 20	$25,000 – $60,000	$3,000 – $10,000+	Long distances, expensive trenching, new poles/conduit, very few users per mile

Very-rural areas might require investment so high no payback is possible. 

That is the reason a rational argument can be made that FTTH should not be built “everywhere,” and that feasible solutions must include satellite or mobile network access. The argument that “work from home” is not possible unless FTTH is deployed is almost always false. 

I have “worked from home, full time” on connections including symmetrical gigabit per second broadband and on connections offering less than 100 Mbps downstream and single digits upstream. My work has never been adversely affected. 

To be sure, my work does not routinely involve upgrading large files on a sustained basis. But most of us do not require a home-based server role, do not create long-form 4K video content all day and need to upload those files continually. 

So if it is said that changes to BEAD rules mean “fewer households will get high speed internet,” the statement is misleading or false. Fewer households might get internet access using FTTH, but that does not mean they will not get internet. And whether such access is “high speed” or not depends on the definitions we choose to use. 

Beyond that, “high speed” might not actually provide any user-perceivable advantage beyond a few hundred megabits per second in the downstream direction. Whether it makes any difference in the upstream direction might be a more-relevant issue, but even there, actual users might not find their work from home impeded. 

We sometimes forget that society has any number of pressing problems to be solved, and internet access is just one of those problems. Investments we make in any area have opportunity costs: we cannot spend the money to solve additional problems. 

Any engineering problem involves choices. Any allocation of societal resources likewise requires choices. Those choices have consequences. 

It is a perfectly logical and appropriate issue to suggest that serving more people, right now,  is a value as great as serving them with a particular solution or capability. Likewise, being efficient in the use of public resources also is a value we tend to believe makes sense. Virtually nobody ever advocates “waste, fraud and abuse.” 

But as a practical matter, it might well be a waste of scarce resources to insist on one particular solution for all home broadband requirements, when other workable solutions exist. 

For every public purpose there are corresponding private interests. Critical thinking might be said to aid decision making when scarce resources must be committed. And that critical thinking might include weighing claims that certain approaches mean “fewer homes will get internet,” when the truth is that the claim only means “fewer homes will get internet using FTTH:

• in areas where other providers already exist

• where there are locations that might not actually require access (an area might have business users but no home users)

• there are other reasons why subsidized service will still be available

• In areas too expensive to serve using FTTH.

In our justified zeal to ensure that critical thinking skills are not diminished by AI, we should not forget that critical thinking skills often are ignored when vested interests interpret reality in ways that serve those interests.

Not "Just a Job"

The reason many find athletic competitions compelling is that, even if it is a job for many of the participants, there is emotional commitment. 

Coach Dan Lanning of the University of Oregon football team exemplifies that fact. Sure, it's his profession, and he's good at it.

But it's more than that.

Most of us probably don't have deep emotional connections to our work. One cannot miss the emotion here. 

Sure, for viewers there can be a sense of community; tribalism; identity and stories. But it isn't our job; our profession; our work; our roles.

Sometimes it is refreshing and captivating to witness times when that also is true for those for whom it is their work. But it isn't "just a job."

AI Impact Will Come Mostly from Consumer Products and Services, Not Enterprise

It is fair enough to raise questions about whether the coming investment in AI compute infrastructure is matched to new AI revenues that investment is expected to generate. 

“Two trillion dollars in annual revenue is what’s needed to fund computing power needed to meet anticipated AI demand by 2030,” according to researchers at Bain and Company. “However, even with AI-related savings, the world is still $800 billion short to keep pace with demand.”

Bain’s sixth annual Global Technology Report predicts that, by 2030, global incremental AI compute requirements could reach 200 gigawatts, with the United States accounting for half of the capability. 

So here’s the thinking: even if companies in the U.S. market shifted all of their on-premise information technology budgets to cloud and reinvested the savings from applying AI in sales, marketing, customer support, and research and development into capital spending on new data centers, the amount would still fall short of the revenue needed to fund the full investment, as AI’s compute demand grows at more than twice the rate of Moore’s Law, Bain argues. 

The return on investment arguably looks different if we look at AI impact on consumer products, though. 

PwC estimates that up to $9.1 trillion of the total global GDP gain from AI by 2030 will come from consumption-side effects (increased demand due to personalized, higher-quality products and services). 

In other words, productivity improvements are part of the story, but not the whole story. 

AI-Influenced Consumer Spending: A report by Cognizant and Oxford Economics projects that U.S. consumers who embrace AI could drive $4.4 trillion in AI-influenced consumer spending in the US alone by 2030.

The global consumer AI market size is projected to reach approximately $674.49 billion by 2030, growing at a CAGR of 28.3% (NextMSC forecast). 

Feature	Bain Argument (B2B/Enterprise Focus)	Consumer AI (B2C/Consumption Focus)
Primary Metric	Annual revenue needed to fund AI compute capital expenditure ($2T needed, $800B shortfall).	Increased consumer spending and consumption-side GDP boost (e.g., $4.4T influenced spending in the United States, $9.1T global GDP from consumption).
Key Conclusions	Supply-side funding shortfall to build the necessary data centers and computing power.	Demand-side explosion creating massive new market value and consumption.

Study Name	Date	Publisher(s)	Key Conclusion on Consumer Impact	Web Link
Sizing the Prize	Oct 2017	PwC	AI will boost global GDP by $15.7 trillion by 2030. Crucially, $9.1 trillion (58%) of this gain will come from consumption-side effects (increased consumer demand for personalized, higher-quality products and services).	https://www.pwc.com/gx/en/issues/analytics/assets/pwc-ai-analysis-sizing-the-prize-report.pdf
New Minds, New Markets	Jan 2025	Cognizant & Oxford Economics	Consumers who embrace AI could drive $4.4 trillion in AI-influenced consumer spending in the U.S. by 2030, accounting for 46% of total U.S. spending. AI will revolutionize the purchase journey (Learn, Buy, Use).	https://investors.cognizant.com/news-and-events/news/news-details/2025/Cognizant-Study-Shows-Consumers-Who-Embrace-AI-Could-Drive-4.4-Trillion-in-Spending-Over-Five-Years/default.aspx
The economic potential of generative AI	June 2023	McKinsey Global Institute	Generative AI could add an equivalent of $400 billion to $660 billion annually to the retail and consumer packaged goods sectors across the 63 use cases analyzed globally.	McKinsey
The State of Consumer AI	June 2025	Menlo Ventures	The consumer AI market has reached $12 billion in the 2.5 years since generative AI went public. The low conversion rate (3% paying for premium) indicates a massive monetization opportunity, especially for specialized AI tools and Voice AI.	https://menlovc.com/perspective/2025-the-state-of-consumer-ai/
AI's transformation of consumer industries	Apr 2025	World Economic Forum (WEF)	GenAI could yield an extra $1.2 trillion in economic value across seven geographies within consumer industries by 2038. Projected impacts include a 10−20% revenue uplift and a 60% reduction in content production costs.	https://www.weforum.org/stories/2025/04/ai-transformation-consumer-industries-wef-report/

The point is that we do not yet know the size of markets and benefits of AI, to evaluate against the cost of computing infrastructure to support AI use cases. But enterprise impact is likely the lesser of the drivers. Consumer products and services are where most of the returns are likely to happen. 

Overinvestment is Typical for Any Big New Technology, But How Much AI Exposure is There?

Oracle  is lining up $38 billion for data center investments in two states tied to its “Project Stargate joint venture with OpenAI and SoftBank to build as much as $500 billion worth of new facilities supporting artificial intelligence operations over a four year period. 

The immediate $38 billion will support new facilities in Wisconsin and Texas that will be developed by Vantage Data Centers. Critics have argued that Project Stargate, when announced, was unfunded. 

The latest package is a step in that direction. 

Of course, there eventually will be concern about the payback from those investments, and investors already seem to be having an AI allergy at the moment. 

To a large extent, payback concerns center on the huge amount of depreciation the investments will trigger. One projection suggests that AI data centers to be built in 2025 will incur around $40 billion in annual depreciation, while generating only a fraction of that in revenue initially. 

The argument is that this massive depreciation, driven by the rapid obsolescence of specialized AI hardware, will outpace the revenue generation for years. 

Some might recall similar sorts of thinking in the first decade of the 21st century as cloud-based enterprise software began to replace traditional software licenses as a revenue model. Then, as now, there was concern about the relatively-low profit margin of cloud-delivered services, compared to use of traditional per-seat site licenses. 

That argument eventually was settled in favor of cloud delivery, however. So optimists will continue to argue that the financial return from massive AI data center investments will emerge, despite the high capital investment and equally-high depreciation impact on financial performance in the short term. 

But skeptics will likely continue to argue that, for processing-intensive AI operations, there is essentially no “long term,” as foundational chip investments must be continually refreshed with the latest versions, mirroring consumer experience with smartphone generations, for example, or mobile service provider experience with their infrastructure (2G, 3G, 4G, 5G and continuing). 

Among the issues is whether key graphics processing unit generations really need to be replaced every three or five years (possibly up to six years). Longer useful life means lower annual depreciation cost. 

But optimists expect demand will grow to more than match investments. 

Metric	Current Demand (2025 Est.)	Future Demand (2030 Est.)
Global AI Market Value	~$390 - $450 billion	~$1.3 - $1.8 trillion
Global Data Center Power Demand	~55 GW (14% from AI)	~122 GW (27% from AI)
Total AI-Related Capital Expenditure	~$200 billion/year	~$1 trillion/year
Required GPU Compute (Exaflops)	Low single-digit Exaflops	Hundreds of Exaflops
Demand-Supply Balance	Significant shortage	Supply-demand balance may be reached, but with continued investment pressure

Some cite the dot-com over-investment in optical fiber transport facilities as the basis for concern about AI data center investment. But there already seem to be key differences. The speculative investment in optical transport was based in large part on the expected survival and success of the many emerging “internet” firms. 

That did not happen when most of the startups failed. Also, there were some key instances of accounting fraud where firms were booking orders or revenue that did not actually exist. 

Depreciation schedules affect some capital-intensive businesses in a significant way.  

Venture capitalist David Cahn has estimated that for AI investments to be profitable, companies need to generate roughly four dollars in revenue for every dollar spent on capex. 

In the enterprise or business space, subscriptions seem the logical revenue model. In the consumer space, it is more likely that advertising will develop as the revenue model. But the issue is when that will happen. 

But there are other, more prosaic issues, such as the impact of depreciation on reported profitability. 

Historically, hyperscalers depreciated servers over three years.

In 2020 they started to extend server depreciation from three years  to four years. That might be deemed a simple recognition that better software enables existing hardware to provide value over longer time periods. 

As a practical matter, that front loads profits as training and inference revenues are booked before a significant amount of  depreciation expenses are recorded. 

In 2021, the hyperscalers investing in AI servers  further extended useful server life to five years and the useful life of networking gear from to six years, citing internal efficiency improvements that ‘lower stress on the hardware and extend the useful life’​.

Between 2021-2022, Microsoft, Alphabet and Meta followed suit, collectively lifting the useful lives of their server equipment to four years. In the year following, Microsoft and Alphabet further extended the depreciable lives for server equipment to six years, and Meta to five years.

It is too early to know which side of the debate will prove correct.