Infra Suppliers and Mobile Operators Might Need 6G More than Users

We can expect to continue hearing quite a lot from suppliers and service providers about the speed, capacity advantages and compelling new use cases for 6G mobile networks. There are obvious reasons for such claims. 

Simply put, every internet access network has required more capacity over time, and mobile services are not exempt from this need. But there are other reasons for the chatter. 

Suppliers have to convince their customers to replace the existing platform (5G) with the next-generation platform. 

Service providers, on the other hand, have to convince regulators to release new spectrum to support the "revolutionary" new networks. 

And claims about important new use cases are typically part of the argument, aside from the virtually-certain improvements in speed (bandwidth or capacity boosts of 10 times have been normal for each mobile platform since the time of 2G) and lower latency. 

Still, at some point, quantifying the value of a new high-speed network for a single user is a challenge. Beyond a certain point, and that point is nearly always far lower than many expect, faster speeds do not provide any measurable improvement in application performance. 

That noted, It's also impossible to calculate the return on investment of an application that has not yet been invented or is not yet possible because the networks will not support it.

Before 4G, no one could have predicted the rise of on-demand ride-sharing. And while mobile video streaming had been noted as a 3G network innovation, those networks did not generally have the capacity to support such apps at scale. So capacity sometimes does matter.

Still, ride-sharing might have blossomed more because of the use of smart phones, plus GPS, rather than "more bandwidth" as such. 

And while 4G bandwidth improvements did make video streaming usable in nearly all cases, one also has to point to the value of social networks and entertainment as the drivers of value for consumers. And it has generally proven difficult to predict which innovations will emerge at scale for any given next-generation mobile network.

Text messaging emerged for 2G networks almost by accident, for example, as the text capability was developed as a user service after the new Signaling System 7 was adopted for network operations.

No Payback from Huge New AI Data Center Investments?

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.

Google, Like Microsoft before It, Avoids a Breakup

The big takeaway from the ruling by District Court Judge Amit Mehta in the U.S. versus Google antitrust case is that Google will not have to divest its browser. The case is said to be notable because it is the first major technology antitrust case since U.S. versus Microsoft in 1998. 

That earlier case, brought by the  U.S. Department of Justice and 20 state attorneys general, alleged that Microsoft illegally maintained a monopoly in PC operating systems and used anticompetitive practices to suppress competition by bundling Internet Explorer with the Windows operating system. 

As in the U.S. versus Google case, the antitrust action alleged that Google, as did Microsoft, suppressed competition by bundling and distribution tactics. 

Judge Thomas Penfield Jackson actually ordered a breakup of the parts of Microsoft that sold operating systems and the part that sold applications (including the productivity suite bundling word processor, spreadsheet and database). 

But that was overruled on appeal. The 2001 Final Settlement ("consent decree") instead included only API disclosure (Microsoft had to share application programming interfaces so third-party developers could build compatible software) and enabled PC manufacturers to install competing middleware (browsers, media players) and change default settings.

The point is that the antitrust action slowed, but did not eliminate,  Microsoft’s dominance in operating systems. But some argue the decision did at least indirectly create space for new competitors in the browser space.

By way of analogy, some will likely argue that the U.S. v. Google decision, if upheld, likewise will slow, but not deter, Google’s leadership in search. 

As was the case for Microsoft, Google will have to take steps to ameliorate its distribution “monopoly.” 

Google is prohibited from entering or maintaining exclusive agreements related to the distribution of Google Search; Chrome; Google Assistant and the Gemini AI app. This includes arrangements that force the use of one product on the distribution or preloading of another (bundling); revenue-sharing tied to placement or multi-year default deals.

Google also must make portions of its search index and user-interaction data available to qualifying competitors and offer search and text-ad syndication services to rival or potential rival platforms. 

Still, the bottom line is that the decision arguably is good for Google.The DoJ had argued for structural remedies (divesting of the Chrome browser, for example). The behavioral remedies are generally viewed by some antitrust experts as weaker and harder to enforce.

The eventual remedies might take years to put into place, and history suggests they will only slow Google's leadership in search, not end it.

Where is the ROI from Private Wireless and How Can We Measure It?

A study sponsored by Nokia and conducted by GlobalData, the Industrial Digitalization Report, argues that 87 percent of enterprises adopting private wireless and on-premise edge saw a return on investment  in one year.

We might note that such claims are based substantially on improvements in sustainability, safety or security goals, though. Under the best of circumstances, it can be difficult to pinpoint actual productivity gains, such as a reduction in downtime or outages, from the use of any local connectivity or network access platform.

Productivity Element	Key Metric	Explanation
Operational Efficiency	Downtime Reduction (%)	Quantifies the reduction in production line or machine downtime due to more reliable connectivity for predictive maintenance and real-time monitoring. For example, a facility might report a 15% reduction in unplanned outages.
	Asset & Inventory Tracking Accuracy	Measures the improvement in locating and managing assets in real time, leading to reduced search times, fewer lost items, and optimized logistics workflows.
	Throughput Increase (%)	Measures the percentage increase in the amount of work completed in a given time, such as the number of items processed per hour in a warehouse or the number of units produced on a factory floor.
	Cycle Time Reduction (seconds/minutes)	Calculates the decrease in the time it takes to complete a specific task or process, such as a robotic arm performing a repetitive action. This is often enabled by 5G's ultra-low latency.
Workforce Productivity	Worker Output Increase (%)	Measures the percentage increase in tasks completed by employees, often through the use of connected tools, augmented reality (AR) for training, or real-time communication on a secure network.
	Response Time (minutes/hours)	Measures the time it takes for a team or automated system to respond to an issue or request, such as a maintenance crew receiving an alert about faulty equipment.
Cost Savings	Total Cost of Ownership (TCO) Reduction	Compares the long-term costs of a private wireless network to legacy systems (e.g., Wi-Fi, Ethernet cabling). Some claims suggest a significant TCO reduction due to fewer access points, easier installation, and lower maintenance.
	Cable Replacement Cost Savings	Quantifies the amount of money saved by using a wireless network instead of installing expensive and complex physical cabling in large or hard-to-reach areas.
Safety and Security	Reduction in Safety Incidents (%)	Measures the decrease in on-site accidents, which can be enabled by real-time monitoring of hazardous areas, remote-controlled machinery, and collision avoidance systems.

That is not a new problem. It is similarly hard to quantify the impact of just about any connectivity solution.  Measuring changes in throughput is an indirect metric and other measures such as “time saved” are somewhat subjective. 

Beyond all that, most managers and most people are probably going to focus on the expected value of the applications and experiences the connectivity enables, not the actual connectivity itself. Connectivity is a "necessary but not sufficient" prerequisite for deriving value from nearly every appliance and application.

Connectivity Solution	Core Productivity Advantages	How to Quantify (Key Metrics)
Ethernet	Speed, Stability, and Security. Enables high-bandwidth applications, prevents data loss, and ensures consistent performance for mission-critical tasks.	Throughput: Measures data transfer speed in Mbps/Gbps. Latency: Measures network delay in milliseconds (ms). Uptime: Quantifies network availability as a percentage.
Wi-Fi	Mobility, Flexibility, and Collaboration. Frees employees from fixed locations, allowing for dynamic workspaces and real-time on-site communication.	Connection Stability: Measures the frequency of dropped connections. Time Savings from Mobility: Quantifies time saved by not needing to return to a desk (e.g., in logistics).  Number of Connected Devices: Assesses the network's capacity to support a high density of users and IoT devices.
Smartphones	Ubiquity, Remote Access, and Real-time Communication. Extends the enterprise network to field workers, enabling them to complete tasks, access information, and communicate from anywhere.	Time to Resolution (TTR): Measures the average time it takes to complete a task remotely. Task Completion Rate: Tracks the number of tasks completed per employee. Cost Savings: Quantifies financial benefits from reduced reliance on paper or fixed equipment.

In Colorado River Basin, Data Center Water Consumption is Not the Problem

Though we hear lots about data center water use, a little perspective is worthwhile. Actually, industry in general, and data centers in particular, represent a small part of water consumption in the desert-like U.S. southwest, for example.  

In the Colorado River Basin, where I live, agriculture accounts for about 74 percent of direct human water use and 52 percent of overall consumptive use (including reservoir and other evaporation), a study estimates. 

Water consumed for agriculture amounts to three times all other direct uses combined. Cattle feed crops including alfalfa and other grass hays account for 46 percent of all direct water consumption.

Another study by Landsat found that in the Colorado River basin, 52 percent of consumption was for agriculture; natural vegetation representing 19 percent of use; evaporation claiming 11 percent of the water and households plus cities and industry consuming 18 percent of the Colorado river’s water. 

 

We need to be thoughtful about data center water consumption, to be sure. But data centers are not the problem, in the Colorado River basin supporting 40 million people. In Colorado, for example, data center water consumption is a negligible percentage of total water consumption.

In Colorado, agriculture consumes 89 percent of the water used by industrial, municipal and agricultural users, for example. Consumers use no more than 20 percent of total water.

Colorado Annual Water Use by Sector
Sector	Delivery / Withdrawal	Consumptive Use	Approx. Share
Agriculture	~13–13.3 million acre-feet (AF)	~4.7 million AF	~87% of delivery; ~89% of consumptive use waterknowledge.colostate.eduWater Education Colorado
Municipal and Industrial	~1.025 million AF (combined)	~0.371 million AF	~6.7% delivery; ~7% consumptive use waterknowledge.colostate.edu
Self-Supplied Industrial	~0.168 million AF	~0.212 million AF	~1.1% delivery; ~4% consumptive use waterknowledge.colostate.edu
Data Centers (subset of industrial)	Variable; example: QTS Aurora uses ~0.5256 million gallons/year (~0.0016 AF); CoreSite Denver proposed up to 805,000 gallons/day (~2.9 AF/day, ~1,060 AF/year) Business InsiderBusiness InsiderGoverning		negligible