For most of the internet’s history, American households suffered from a genuine shortage of bandwidth. Connections were slow, unreliable and radically unequal. A family’s devices competed for a narrow pipe, video calls froze, movies buffered, and uploading a large file could monopolize the connection for hours. Rural communities sometimes had little more than dial-up in everything but name.
Now, in a growing number of American cities, the situation has flipped. Fiber companies are offering residential connections so fast that consumer demand—and consumer equipment—cannot keep up.
Several American providers now advertise symmetrical residential service reaching eight gigabits per second: 8,000 Mbps downloading and 8,000 Mbps uploading. In some markets, the fastest plan costs surprisingly little more than an already lavish one-gigabit plan.
It is an astonishing amount of capacity. It is also capacity that almost no household can meaningfully use.
That is not an indictment of fiber providers. Quite the contrary: after decades in which cable and telephone companies sold too little bandwidth at too high a price, it is refreshing to see someone build ahead of demand. Fiber placed in the ground today may serve a neighborhood for generations. Offering enormous symmetrical capacity demonstrates what that infrastructure can do and places competitive pressure on cable companies whose upload speeds remain stuck in the tens of megabits.
But we should distinguish between infrastructure worth building and service tiers worth buying.
Eight-gigabit fiber is a magnificent neighborhood resource. As an individual residential subscription, however, it is largely a triumph of marketing over arithmetic.
A household cannot stream its way to eight gigabits
Consider an improbably active family of five. Every person is at home, awake and determined to consume bandwidth simultaneously.
Five televisions are each showing a different 4K movie. At approximately 25 Mbps apiece, that consumes around 125 Mbps. Two family members are gaming online, which uses surprisingly little bandwidth—perhaps another 10 Mbps combined—because latency matters far more than raw speed. Two people are participating in high-definition video conferences, adding perhaps 10 Mbps in each direction. A third is livestreaming to YouTube at 1080p and 60 frames per second, which requires around 12 Mbps. Even 4K livestreaming consumes tens of megabits per second, not thousands.
Add ten cloud-connected security cameras, smart speakers in every room, phones, tablets, thermostats, appliances, automatic photo backups and a teenager downloading something nobody wants to discuss. Ordinary sustained traffic would probably remain under 300 Mbps.
The family could deliberately push consumption higher by starting several enormous game downloads and cloud backups at once. Those bulk transfers could fill whatever bandwidth was available—but that is not the same as needing it.
A 100-gigabyte game takes roughly 13 minutes to download over an ideal one-gigabit connection. At eight gigabits, the theoretical time falls below two minutes. That is convenient, but it does not transform family life. It merely changes “make a sandwich while I wait” into “check my email while I wait.”
Even then, the game company’s server may not deliver data to one customer at eight gigabits. The computer’s network interface may accept only one gigabit. Its storage drive may become the bottleneck. Antivirus scanning, decompression and installation can take longer than the download itself.
A realistic technology-heavy family might occasionally benefit from two gigabits of aggregate service. Sustaining eight gigabits would require something closer to a small production studio, a commercial data operation or a household staging an elaborate benchmark specifically to prove it could consume eight gigabits.
The ordinary five-person family does not have eight gigabits’ worth of simultaneous needs. More importantly, it probably does not own an eight-gigabit network.
The speed stops at the equipment closet
When people buy an extremely fast internet plan, they understandably imagine that speed flowing throughout the house. But the advertised number normally describes the connection arriving at the provider’s optical terminal or router. Everything after that point belongs to the home network, and that network is usually much slower.
Most televisions, streaming devices, game consoles, printers and smart-home hubs have either 100-Mbps or one-gigabit Ethernet. Many desktop computers still have one-gigabit ports. Newer premium computers may include 2.5-gigabit Ethernet. Ten-gigabit interfaces remain uncommon outside workstations, servers and serious enthusiast systems.
Wi-Fi does not solve the problem.
Wi-Fi 7 equipment can advertise spectacular theoretical numbers by combining wide channels, multiple spatial streams and several frequency bands. But a marketing maximum is not the speed that one phone receives from three rooms away.
A wireless client has its own radio limitations. It shares airtime with every other device connected to the same access point. Walls, distance and interference reduce performance. The newer 6-GHz band offers wide, relatively uncluttered channels, but its signal weakens faster through walls than lower-frequency Wi-Fi.
A large home therefore needs several wired access points, not one heroic router shouting from a shelf in the office.
Even a very good Wi-Fi 7 device positioned near an excellent access point might achieve one or two gigabits under favorable real-world conditions. That is wonderfully fast. It is also only a fraction of an 8-Gig subscription.
Thus, the absurdity of the premium plan becomes clear. A company can deliver an eight-lane interstate to the customer’s front door, but inside the house the family is still traveling on one- and two-lane roads.
What it would cost to build a real 8-Gig home
Suppose a technology-heavy family of five lives in an existing 4,000-square-foot house and genuinely wants to use as much of an 8-Gig connection as practical.
The goal is not merely to produce an impressive speed test while standing beside the router. The goal is to give as many devices as possible access to multi-gigabit speeds throughout the property.
The house would need a professional-grade network built around a 10-gigabit backbone.
A plausible retrofit would include:
| Component | Estimated installed cost |
|---|---|
| Router and firewall capable of processing at least 8 Gbps | $700–$2,000 |
| 24-port switch with multiple 10-Gig and 2.5-Gig ports | $1,100–$1,800 |
| Four or five Wi-Fi 7 access points with multi-gigabit uplinks | $1,000–$1,500 |
| 16–24 Cat6A cable runs through finished walls | $4,000–$8,000 |
| Rack, patch panels, battery backup, modules and connectors | $1,000–$2,000 |
| 10-Gig adapters for desktops, workstations and storage servers | $600–$1,500 |
| Secondary switches for offices and entertainment areas | $600–$1,500 |
| Professional configuration, testing and troubleshooting | $1,000–$2,500 |
| Probable total | $10,000–$20,800 |
A reasonable central estimate is approximately $15,000, assuming the home is already built, the walls are finished and the cable routes are moderately difficult.
A hostile retrofit involving fire blocks, vaulted ceilings, extensive drywall repair or no accessible attic could exceed $20,000. A house wired during construction could cost substantially less.
Even after spending $15,000, the family still would not make every device operate at eight gigabits. The purpose of the 10-Gig backbone would be to let several devices collectively share the full internet connection. A high-end desktop or network-attached storage server might approach the maximum over a wired 10-Gig link. Most phones, tablets, televisions and appliances would continue operating at a small fraction of it—and would be perfectly happy doing so.
This creates a peculiar economic comparison. The jump from a one-gigabit subscription to eight gigabits may cost only several hundred dollars more per year. But making the house capable of using the upgrade could require $10,000 to $20,000 in networking equipment and construction.
The inexpensive part is buying the bandwidth. The expensive part is creating somewhere for it to go.
Why companies sell speed nobody needs
Internet providers understand all this. They sell 8-Gig plans anyway because the fastest number performs several valuable marketing functions.
First, speed is easy to compare. Reliability, latency, routing quality, customer service, outage response and neighborhood congestion are difficult to express in an advertisement. “Eight gigs” is immediate. It is larger than five, which is larger than two, which is larger than whatever the cable company offers.
Second, a spectacular flagship plan changes how consumers perceive the entire product line. Beside an 8-Gig plan, one gigabit begins to look restrained and economical. The top tier acts as a price anchor. Few customers need to buy it for the strategy to work; they need only see it.
Third, the fastest plan establishes technological superiority. A provider offering eight symmetrical gigabits can credibly present itself as the company building tomorrow’s network. The customer may purchase only one gigabit, but the existence of eight reassures that customer that the underlying fiber will not soon become obsolete.
Fourth, Americans have been trained to buy the largest available number. More horsepower must mean a better car. More megapixels must mean a better camera. More gigabytes must mean a better computer. More bandwidth must therefore mean better internet.
Yet internet service is not a fuel tank from which every activity draws endlessly. Once the connection is sufficiently fast for the household’s simultaneous work, additional bandwidth sits idle.
It does not make an ordinary webpage load eight times faster. It does not reduce the physical distance to a remote server. It does not fix weak Wi-Fi in the bedroom. It does not improve a movie that is already receiving all the data it requests. It does not make an online game more responsive when the real limitation is latency.
For most homes, the noticeable difference between 100 Mbps and 500 Mbps can be substantial. The difference between 500 Mbps and one gigabit is usually confined to large downloads and multiple heavy users. The difference between one and eight gigabits is mostly visible in carefully arranged speed tests.
The strange psychology of unused capacity
The speed test has become the internet provider’s equivalent of the automobile speedometer.
A car may have a speedometer reading to 160 mph even though its owner never exceeds 80. The unused numbers communicate capability. They imply power even when that power cannot be legally, safely or practically used.
Gigabit service operates similarly. The customer runs a test, sees an enormous number and feels reassured that the service is superior. Whether any application actually requests that much data becomes almost irrelevant. The test itself is the product demonstration.
At eight gigabits, however, even the demonstration becomes difficult. A customer may subscribe to an 8-Gig plan and then complain that a laptop receives “only” 900 Mbps over Wi-Fi. Nothing is broken. The laptop may be operating beautifully. The problem is that the customer purchased a whole-house aggregate capacity and interpreted it as a guaranteed speed for every device.
Providers rarely emphasize that distinction. “Eight gigabits delivered to your optical terminal, to be divided among compatible devices through networking equipment that you probably do not own” would be technically more accurate, but it would make a terrible billboard.
Instead, the giant number is allowed to float free of the physical system required to use it.
We have finally reached abundance—unevenly
There is something encouraging beneath the absurdity. In certain American neighborhoods, connectivity has finally outpaced immediate household need. That represents genuine progress.
The problem is that this abundance remains geographically uneven. One family is invited to choose among one, three, five and eight gigabits while another family, sometimes only a few miles away, struggles with unreliable fixed wireless, aging DSL or a cable upload speed inadequate for modern remote work.
America has not solved broadband scarcity. It has created islands of broadband excess.
There is also a danger that the public conversation will confuse premium speed tiers with universal progress. A country does not have excellent broadband merely because a few neighborhoods can purchase eight gigabits. The more meaningful measures are whether nearly everyone can obtain reliable service, whether upload speeds support modern work and education, whether prices are affordable, and whether providers serve low-density communities rather than only profitable subdivisions.
An additional billion dollars spent advertising and overbuilding the fastest neighborhoods may generate better speed-test headlines. The same investment extending reliable fiber or fixed broadband into poorly served communities could create far greater social value.
That does not mean providers should stop improving their networks. Building fiber with enormous future capacity is precisely what they should do. It means we should stop using the largest residential speed tier as the primary measure of progress.
Buy reliability before bragging rights
A technology-heavy family of five in a 4,000-square-foot home would be exceptionally well served by symmetrical one-gigabit fiber, four properly positioned wired access points and Ethernet connections to stationary high-demand equipment.
Two gigabits might be justified for several content creators, frequent large cloud transfers or a serious home server. Beyond that, the household enters a realm where the objective is no longer meeting demand but manufacturing enough activity to validate the purchase.
Most families would benefit much more from spending money on their internal network than on an 8-Gig subscription. Properly positioned access points, wired backhaul, Ethernet connections for televisions and desktop computers, and a battery backup for the router could improve everyday performance far more than multiplying the incoming bandwidth.
There is nothing morally wrong with buying eight gigabits. Enthusiasts routinely spend money on performance they rarely use. People own sports cars that can triple the speed limit, watches engineered for ocean depths their owners will never visit and professional cameras used mostly for vacation photographs. Pleasure, curiosity and bragging rights are legitimate uses of disposable income.
But they should not be confused with necessity.
The most honest description of an 8-Gig home plan is not “internet for a large family.” It is “a small-business-class data connection available to a residence.” That is technically impressive and perhaps historically remarkable. It is also a connection whose greatest achievement may be sitting mostly unused while assuring its owner that, if all five family members ever decide to download enormous video games, back up several computers, host a production studio and livestream in 4K at precisely the same moment, the internet will not be the problem.
The rest of the house probably will be.
Leave a Reply