Why you shouldn’t buy the miracle broadband network Softbank’s Masayoshi Son is selling


Credit: KAZUHIRO NOGI/AFP/Getty Images

It seems SoftBank CEO and Sprint(s s) chairman Masayoshi Son will say anything to push his merger of Sprint and T-Mobile, including promising to rejuvenate broadband competition in the U.S.

On Wednesday at the Code Conference, Son gave a well-received talk on deplorable state of the internet in the U.S. Though he never mentioned T-Mobile directly, the subtext was that if regulators let him get his hands on T-Mobile, he wouldn’t just make the U.S. mobile landscape the competitive, but the entire realm of internet access.

Give me T-Mobile and I’ll give you a competitor to Comcast-Time Warner, was the message Son delivered, and everybody seemed to eat it up. I think Son is being pretty disingenuous here. He simply can’t deliver the network to meet those promises.

Here’s why.

Source: Shutterstock / Max Krasnov

Source: Shutterstock / Max Krasnov

Mobile and wireline broadband networks are fundamentally different animals, and no matter how much wireless technology improves you’re never going to pump the same amount of capacity through a cellular connection that you would through its wireline equivalent. Son argued that today’s average LTE connection — at 6 Mbps — is just as fast as the home broadband speeds many Americans have today (though as The Verge’s Chris Ziegler pointed out, he seemed to be making up numbers), but Son is conflating speed with the cost of capacity.

The way people use their home broadband connections simply can’t stand up to today’s mobile technology and today’s mobile business models. Speaking at same conference as Son, Netflix(s nflx) CEO and co-founder Reed Hastings said wireless technologies can’t handle the demands of video traffic, and he’s right; especially when you consider the way people use his streaming video service. Netflix and other video streaming services are driving monthly home broadband usage into the hundreds of gigabytes.

Sandvine estimates that average cord-cutter (which Sandvine defines as the top 15 percent of users) gobbles up 212 GBs a month. Let’s assume you consumed that amount of data each month on your LTE connection. Even on their most liberal shared data plans, AT&T and Verizon would charge you $2,805 a month for that kind of usage.

Ironically, Son’s Sprint(s s) would be far more expensive, since it still uses per-megabyte pricing on its mobile broadband plans. Sprint would charge you more than $10,000 if you consumed 212 GBs of data.

Sorry, Masa, but you’re not off to a good start.

What about this future you speak of?

To be fair, Son is deriding the current state of mobile and wireline internet, with an eye to changing it. His plan is to make the networks of a combined Sprint and T-Mobile far more efficient while offering data at prices far more cheaply than we see today.

I definitely buy the argument that mobile carriers encourage the idea of spectrum scarcity and inflate prices higher than they should be. But how cheap could Son go? Even if he were to slash the cost of mobile data to a tenth of its current price, he’s still talking about $200 to $300 monthly broadband plans as opposed to the $60 to $65 we pay for a home broadband connection today.

When looking at the technology involved, Son also stands on tenuous ground. As I wrote in Gigaom’s reinventing the internet report, mobile and wireless technologies will gradually get faster and more powerful, evolving to a point where we may some day be able to consume data over wireless connections as indifferently as we consume it over wireline connections. But that kind of scenario involves much more than the cellular networks Sprint could provide.

Japan's SoftBank Corp. founder and President Masayoshi Son speaks during a press briefing to announce the company's financial results in Tokyo on February 12, 2014. SoftBank said February 12 its nine-month net profit soared 58 percent thanks to strong iPhone sales, but third-quarter earnings suffered because of losses in its newly acquired Sprint Nextel unit. KAZUHIRO NOGI/AFP/Getty Images

Japan’s SoftBank Corp. founder and President Masayoshi Son speaks during a press briefing to announce the company’s financial results in Tokyo on February 12, 2014.  KAZUHIRO NOGI/AFP/Getty Images

We’ll have to tap technologies ranging from Wi-Fi to white spaces supplied by multiple different providers, including the ISPs Son aims to compete against. Our wireless connections won’t just come from Sprint; they will come from Comcast(scmsca) and Google(s goog) and Facebook(s fb), from city governments and even community networks.

Yes, one day we will be able to able to consume 100 or 200 GBs of mobile data for the same cost we pay for a home broadband connection. But it’s not like wireline broadband technologies or our internet consumption habits are standing still. We’re moving away from coaxial cable and copper to fiber links to the home. Meanwhile the resolution of our video programming is increasing and our apps and data storage are moving into the cloud.

Five years from now, when Son is ready to offer 200 GBs for $50 on the cellular network, 200 GBs a month will seem like a paltry amount. Son is chasing a moving target, and the simultaneous advancement of both wireless and wireless access technologies dictate that he’ll never catch up.

Buying T-Mobile won’t change any of this (though according to Reuters, Son got one step closer by getting Deutsche Telekom to sign off on the deal). In fact, a merged Sprint and T-Mobile would most likely be mired in years of dysfunction as they integrated their operations and completely incompatible networks.

After they emerged, Sprint would have more customers and more spectrum to add to its massive stores of unused airwaves, but it would be no closer to offering a broadband service competitive with the wireline ISPs. Subscribers and spectrum are nice things to have, but you can’t use them to break the laws of physics or alter the fundamental economics of mobile.



I posted this in response to another comment further down, but it does directly contradict the numbers in the article as well.

The Sprint broadband mentioned in the article is for mobile broadband devices like embedded laptops, wireless cards, etc. not for tablets and home routers.

You should be looking at the Wireless Router Plans (also on that same page at Sprint.com if you scroll to the right from the mobile broadband plans, those plans can’t even be activated on the tablet or router devices) because those are the plans aimed at the home-use segment that would be the wireline competitive plans, not the road warriors who are using them in airports, on the road, etc.

A 12GB router plan would be $80, with 1.5¢/MB overage cost. This comes out to about $15/MB overage. So instead of a $10,000 overage cost for 212GB it would be $3,000, which is more on-par with the AT&T and Verizon number of $2,805.

To quote the plans: “Monthly usage in excess of plan allowance on network: 1.5¢/MB (tablets & routers), 5¢/MB (mobile broadband).”


Johnny Kragebæk Olesen

Normally I get a better insight in the American telecom marked by reading Kevins articles. But this time he is completely wrong.

I do not know if it is a lack of international view or simply the typical American gut reaction that fewer companies equals less competition, but Kevins arguments are simply plain wrong.

Is it too much to ask Kevin, to do a follow-up article where you could look beyond the US border?

In other parts of the world mobile companies already use mobile broadband to compete against fixed internet connections (xDSL, fiber).

I my home country of Denmark, we have three 4G LTE mobile networks (TDC, Telia/Telenor and 3) and a fierce price competition among the mobile companies.

In Denmark you can today buy a 200 GB 4G LTE mobile broadband subscription for 299 DKK, which is the same price as a unlimited xDSL or fiber broadband connection. The 4G LTE speeds are comparable to the xDSL connections (20-40 down / 10-15 up). And for a reasonable price of 399 DKK, one can subscribe to a mobile broadband package of 500 GB data.

At those prices mobile broadband is a real alternative to fixed connections, but the mobile companies must have access to sufficient spectrum and a network with enough cells to cope with the extra demand from mobile broadband users.

Talking about usage on MY OWN mobile broadband connection; My small family of four is heavy media users – both Netflix, YouTube, Twitch and a bunch of local streaming services – but we seldom use more than 350 GB per month, well within the limit of 500 GB. We also have two Macbooks that use online backup (BackBlaze) and we have several hundreds of GBs backed up – all transmitted using mobile broadband.
So from my point of view mobile broadband is an excellent replacement of a fixed connection.

There is room for three mobile companies in the USA, and everyone knows that the two small operators will be squeezed in the long run due to the way frequencies are licensed in the US.
A marriage between Spring and T-mobile may hinder competition in the near future, but it will benefit the consumer in the long term. Big networks has a lower expense per customer, due to better utilization a the network assets, lower acquisition cost of customers etc.

So see from the other side of the pond, a merger between Sprint and T-mobile might just be the best thing the US telecom marked has seen in decades.


Kevin, have you been to Asia, Japan specifically? I currently live in Tokyo and the mobile speed is super fast. I barely see a difference between the landline wifi and 4G on my phone from Softbank (except for streaming the video which is slightly better on wireline). I have seen Softbank from a also-run company in Japan to beat DoCoMo & AU hands down. My point is that I trust Masayoshi and he can deliver wonders. I have to tell you I dread coming back to the US where the mobile technology is way behind… South Korea is even better in terms of speed.

US has to change if it were to catch up with the rest of the world unless it wants to compete with the lower tier countries. I feel the same as Masayoshi in terms of how horrible the internet speed is in the states. By the way, I pay around $50 bucks for fixed data package in Tokyo using Softbank. You have to come to Asia to know what Masa is saying.

Michael Elling

Kevin, what if both you and Son are wrong.

Wireless 20/20 reveals the 5 yr TCO for a wireless operator is 13% capex and 87% opex. This is upside down from most people’s perception. Also, consider that providers charge you a buck, but deliver only 30 cents of value; given 70% user designated offload. The wireless model is clearly not sustainable, but most in the trade and capital markets don’t understand this, yet.

The reason is that many issues are being conflated and confused. Here are the real issues:
1) how quickly will fiber be pushed to the edge and be available for mobile operators to drive cost down?
2) what is consumer’s propensity for consumption? Mobile vs fixed. Small screen vs large-screen. On-demand vs linear.
3) What if the vertically integrated, silo-ed models of the edge access providers are the real source of high BB pricing and speeds that are below what they should or could be based on the pricing and performance of the , still evolving, Google Fiber model in KC?

It should soon become apparent to the market that artificial distinctions from the past based on technology, application, geographic, or market segments will no longer define service providers. Demand will be defined by 4/8K VoD, 2-way HD collaboration, seamless mobile BB, and IoT. Supply is woefully behind on satisfying capacity (particularly upstream), latency, QoS, security and redundancy issues for all of these rapidly developing trends.

Ayrio Cronin

I’m sorry – that’s business pricing…..I stand corrected

Kevin Fitchard

Hi Ayrio,

Smartphone only. If you use your phone as a modem, there are caps.

Ayrio Cronin

You seem to have left off the fact that Sprint offers unlimited data at a price that is <$60 per month………..

Richard Bennett

You’re assuming that any wireless broadband service would work more or less the same way that mobile broadband works today, but that’s the first idea you need to discard. Verizon is already trialing a fixed-location wireless broadband service using LTE that doesn’t support roaming. With MU-MIMO, we’ll eventually have efficient beam-forming that permits fixed-location wireless links to be treated more like wires.

This will enable the lifting of data caps – Verizon has already raised theirs for their experiements – which makes the service more competitive. Fixed location LTE doesn’t have to be the best network in the whole world to be successful, it simply has to be better than ADSL, which isn’t that hard.

Son talks a lot of nonsense about the relative speed of America’s networks vs those in other countries. Sprint has the slowest network, and that drags down the average, and we’re essentially on par with Japan on LTE and will leap ahead again when we deploy LTE-Advanced. But the idea of wireless ISPs for rural areas has been well proved by the WISPs who do the job with Wi-Fi today, despite the fact that it’s a poor fit for this scenario.

Kevin Fitchard

Hey Richard, thanks for commenting.

I definitely see the argument for fixed LTE as replacement for DSL in areas where the only option is a slow copper or restricted satellite connection. But Son is implying that fixed wireless could be a competitive proposition all over. In my mind being better than ADSL isn’t good enough, which is why I used the example of cord-cutter rather than the overall average (around 40 GBs a month). If Sprint can only provide an alternative that’s as good or slightly better than our worst broadband technology or plan, than it’s not really competitive. At least that’s how I feel about it.

But you’re definitely spot on about how fixed wireless can take advantage of new technologies that traditional mobile networks might not (5G discussions are now eevn centering on massive MIMO arrays at super high frequencies). I glossed over them because I didn’t want to get bogged down in technical details. But I think the logic still applies. As fixed wireless technologies improve, wireline broadband technologies will as well. Wireless will always be behind the curve.

Richard Bennett

Well, yes, wireless will always be behind the curve in terms of raw capacity, but at some point it will have have much capacity that the curve doesn’t matter. Even today, CPU power, SSD bandwidth, transit capacity, and the bandwidth of the human senses are the limiting factors.

Looking at it another way, wired will always be behind the utility curve because it’s confined to a fixed location and people are mobile. What would you rather have today, a 100 Mbps mobile network or a 1 Gbps network that only works on your desktop or TV set? I’d go for the mobile every time. Netflix on Google Fiber only streams at 4 Mbps so most of the capacity of Google Fiber is simply unusable, and will be for at least ten years if not longer.

Lots of people are pushing gigabit networks and FTTH, but in the real world the high capacity of fiber is mainly valuable as a backhaul and aggregation technologies because cable – which is really a spectrum-based system internally – and advanced wireless are advancing so fast.

“Fastest” is rarely a goal in engineering or business, any more than “cheapest” is. What counts is “fast enough” and “cheap enough”. Vectored DSL at 40-80 Mbps – unshared – is perfectly fine for 99% of applications today, and wireless at that range is already here with Wi-Fi and beginning to emerge in markets with lots of spectrum and virgin LTE builds as well.

There certainly are things we could do with all fiber networks at 1 to 10 Gigs that we’re not doing today, but we don’t even have the gear to build the apps anyway.


Sprint MBB charges 1.5Cents/MB in overage which would equate to $15/GB how does that come to 10,000 for 212 GB it looks like a under $700 which would be far under the cost you’ve quoted for the other wireless carriers. Also I’m no expert on wireless but if capacity is the issue here. I’ve read that Sprint has more Capacity than the big 2 combined because of the clearwire 2.5 spectrum. So if you combined that with T-Mobile Spectrum that would give them even more and they would still have A whole lot less customers


The plan you’re looking at is for mobile broadband devices like embedded laptops, wireless cards, etc. not for tablets and home routers.

You should be looking at the Wireless Router Plans (also on that same page if you scroll right from the mobile broadband plans, those plans can’t even be activated on the tablet or router devices) because those are the plans aimed at the home-use segment, not road warriors.

A 12GB router plan would be $80, with 1.5¢/MB overage cost.

Monthly usage in excess of plan allowance on network: 1.5¢/MB (tablets & routers), 5¢/MB (mobile broadband).



I smell a turd in the works. Sprint already had an opportunity to challenge home DSL with Clear – a solution that provided UNLIMITED 4G at an affordable price. They actually offered home 4G routers exclusively for the purpose. What did they do once they got majority control? KILL IT.

The dirty truth is wireless broadband in the US sucks, I’ve said it for years:


Sadly, this situation isn’t changing any time soon.

Craig Campbell

I’m very skeptical of his promises. I used to spend a lot of time in Japan, and Softbank was pretty widely derided, the general opinion being that most of their customers were because they were the only network at the time carrying the iPhone. Now that both KDDI and NTT Docomo offer the iPhone, I’d be interested to know how/if the landscape has shifted.

Kevin Fitchard

Thanks for commenting Craig,

Yeah, it’s amazing the fever that overtakes people when they want to pull off a mega-merger. Facts, perspective, context — they all go out the window. We saw it with AT&T-Mo and now we’re seeing it again. Son takes a lot of credit for bringing the iPhone to Japan, but wouldn’t you think it originally had something to do with SoftBank’s networks initially matching up the iPhone’s specs?

Craig Campbell

Yeah, at the time I think Softbank was the main UMTS/HSPA network. Docomo certainly is now, but at the time of Softbank’s iPhone launch, they may still have been migrating from their PDC (Mova) to their UMTS/HSPA (Foma) network. Or they might have been mostly UMTS by that point, but not willing to ink a deal under Apple’s terms (not being as desperate for customers as Softbank was, just coming out of it’s J-Phone and disastrous Vodafone Japan days). KDDI, the number 2 network, would have been all CDMA/1X/EVDO at the time (before the days of CDMA iPhones on Sprint and Verizon) and only launched LTE in the last couple of years.

Comments are closed.