Upgrading America


Achieving a Strategic Bandwidth Advantage

And a Psychology of Bandwidth Abundance


Drive High-Performance Knowledge Exchange


Fujitsu Conference on

Paving the Road to Unlimited Bandwidth:

Technologies and Applications for a Connected Age

San Jose, California

June 13, 2012


The National Broadband Plan addressed the traditional four questions of telecom policy: how do we get networks everywhere, how do we get everyone on, how do we have a competition framework that drives consumer welfare and how do we use existing networks better?

The Congressional mandate to write the plan, however, required us to examine a 5th question: what policies would drive innovation over broadband? This inquiry led to another—a question at the heart of this conference, a question that both excites and troubles me:


What happens if we remove bandwidth as a constraint to innovation?

The question excites me because, as I learned from a brilliant doctor, genetic sequencing—which uses so much bandwidth it is cheaper and faster to send the results by Fed Ex—can revolutionize medicine. Eliminating bandwidth constraints can make this new tool more effective for treating, among other maladies, cancer.

The question excites me because, as I learned from numerous educators, immersive gaming technology—which requires massive bandwidth-- can improve the effectiveness of education and job training, particularly for those for whom traditional methods fail.

The question excites me because, as I learned from a broad spectrum of people, from those doing scientific research to those providing business services, from those focused on security to those who wish to invigorate civic engagement, the coming age of Big Data has the potential to accelerate advancements in addressing our most vexing problems. But to achieve its full potential Big Data will need Big Bandwidth.

The question troubles me for the same reasons.

That doctor told me, with evident pain, he is treating people today who will soon die but who, if he had the tools of tomorrow today, he could save. So too, delays in removing bandwidth as a constraint mean that soon a child will grow up illiterate who shouldn’t. A man seeking to retrain for a job of the future will not receive the help he needs. A critical discovery to address climate change could be delayed because we are not collaborating as effectively as we could.

There is a cost to delay, even if we can’t see it, even if economists can’t measure it.

In this light, we should all want to accelerate that moment where only untapped human ingenuity, not bandwidth, constrains our progress.

Since leaving the FCC, I have devoted most of my time to that mission by organizing Gig.U—three-dozen university communities dedicated to accelerating the deployment of next generation networks and services. It has already led to projects in Maine and Florida and a $200 million private sector commitment for the first multi-community gigabit deployment in the United States. We will soon make other announcements that broaden the opportunity to discover what happens when we can innovate unconstrained by bandwidth.

While I have been working on Gig.U, policy makers have been using the foundation of the Plan to address the traditional four questions. That agenda remains vital. Recent actions have made progress. I hope it continues.

What I would like to do today, however, is to argue that over the next few years, the prime mission of communications policy ought to be to eliminate bandwidth as a constraint on innovation and productivity.

In doing so, I will offer three assertions that lay the foundation for a communications policy that would accelerate next generation network deployments and services.

They are about communications policy in the United States, though I believe that the first and third are applicable around the world. They represent a policy shift, one that can be adopted no matter who wins the election; a shift that would lead to an acceleration of economic growth and an improvement in how we deliver essential government services.

They are:

As the principal activity of our economy and society is knowledge exchange, and as the Internet is its commons, the prime purpose of communications policy should be to improve knowledge exchange by delivering a strategic bandwidth advantage and a psychology of bandwidth abundance.

The current market dynamics and regulatory structure in the United States will not deliver the wireline upgrade necessary to deliver a strategic bandwidth advantage nor a psychology of bandwidth abundance.

Network upgrades are not only the best way to achieve a strategic bandwidth advantage and psychology of abundance; it is also the best way to address the traditional four questions of telecom policy.

So let’s go through these.

As the principal activity of our economy and society is knowledge exchange, and as the Internet is its commons, the prime purpose of communications policy should be to improve knowledge exchange by delivering a strategic bandwidth advantage and a psychology of bandwidth abundance.

What do most of us do?

Knowledge exchange.

At the heart of our daily efforts, we gather data, store it, analyze it, and through a feedback loop adjust behavior and start the process all over again. The better we do it that, the better the outcome.

This is not a novel thought. Decades ago, the visionary Peter Drucker suggested that: “(t)he most important contribution management needs to make in the 21st century is similarly to increase the productivity of knowledge work and knowledge workers. …It is on (the productivity of knowledge work), above all, the future prosperity—and indeed the future survival of developed economies—will increasingly depend.”

Knowledge work is not just the province of traditional white-collar jobs. Knowledge work is now part of every sector.

Whether in manufacturing-where the Boeing Dreamliner, the largest manufacturing effort ever, was done in multiple countries thanks to advanced tools of knowledge exchange—or retail--where smart use of those tools enabled a small company from Bentonville, Arkansas to become the world’s largest retailer--or agriculture—where farmers in Africa exchange texts to multiply their income, all increasingly depend on knowledge exchange.

And this is before the era of Big Data—now dawning--multiplies the importance of knowledge exchange.

History teaches us many lessons about how to benefit from technology-driven change.

One is the importance of combinatorial innovation.

For thousands of years, population and food production rose in tandem, which, as Malthus famously wrote, doomed mankind to a state of mere subsistence and misery.

But something got us out of that trap—the Industrial Revolution.

What characterized the advance of science and technology in the Industrial Revolution, in contrast with discoveries made in earlier eras, was the way advances combined with others; the steam engine was not the product of a single innovation but literally of thousands.

It was this exchange and combination of ideas, not isolated discoveries, that created the escape hatch from the Malthusian trap.

The historian of innovation, Steven Johnson, credits the coffee houses of London, with facilitating the exchanges that led to unprecedented growth.

Perhaps future historians will credit free Wi-Fi in coffee houses around the world with magnifying that exchange. But there is no doubt that the broadband infrastructure is the primary location where combinatorial innovation occurs today.

A second lesson about technological change is the importance of leadership in the exchange of inputs.

A precursor of great advances in an economy is when we exchange one set of inputs with another.

We saw it at the beginning of the agricultural age when we replaced human labor with animal power.

We repeated it with the industrial revolution when we replaced human and animal based inputs with coal, iron and trains.

We repeated it over a century ago by exchanging those inputs with steel, electricity, cars and planes.

Over the last 20 years, we have seen another historic change as we exchanged inputs based on atoms with those based on microprocessors, fiber optics, and digitization.

That input exchange, has, among many things, transformed knowledge exchange.

Countries that lead in driving, and adapting to, input change receive a greater share of the benefits from that change.

Thus, every country should want their residents to replace their use of atoms with the use of bits, chips and bandwidth wherever possible, as soon as possible.

The broadband ecosystem brings these two lessons together.

This transformation from the tangible to the intangible, from the material to the immaterial, from bricks to bits, has created a world in which the principal activity becomes knowledge exchange and the principle method is over the Internet.

One way in which this transformation differs from previous ones is how it scales. Building on the insight of economist Paul Romer as to how bandwidth is a non-rivalrous public good, former FCC Chairman Reed Hundt has written in a brilliant, soon to be published article, “it is technically and economically feasible, for example, to distribute robust public goods to everyone on the Internet – about three billion people and almost everyone in the United States – at a cost that approximates zero.”

This opens up all kinds of opportunities as “the low marginal costs will combine with human creativity to produce the kind of abundance that replaces a short, cross country phone call priced at dollars with a video conference uniting persons on five continents priced at free.”

This reflects a different kind of wealth creation than we have ever seen. Building a fantastic Ferrari benefits the seller and buyer but its price reflects its scarcity and its impact is limited.

Building a fantastic teaching tool, such as the Khan Academy, can teach millions; the wealth it creates, benefits all.

So what are the building blocks to increasing the probability of being both the innovators and the first beneficiaries of innovation?

First, a strategic bandwidth advantage.

Think about it this way. 25 years ago, the farmer and doctor in rural Iowa needed the same thing—dial tone. Today, their needs are different. The farmer needs great mobility service to be able to connect to critical data while outside and moving while the doctor needs great fixed connectivity to review MRIs with experts in distant cities.

We need a strategy that gets them both the kind of bandwidth they need, even though it is now different. Throughout the economy we should have the necessary bandwidth going to the right places to drive economic growth and the most effective delivery of essential public services.

Spectrum plays a big role. Recent actions, such as Congress passing incentive auction authority, Cable increasing its Wi-Fi network, and the Presidential Commission recommendation on dynamic access represent progress. Still, as AT&T CEO Randall Stephenson wrote the other day, we still need to make more spectrum available for broadband use.

In addition to getting the bandwidth we need today to optimize productivity for all, we also need hubs of huge bandwidth to assure we build and retain the talent experienced in designing, building, operating, and innovating on top of the best networks in the world. These networks are always going to be ahead of where the mass-market applications will be but are critical to the innovation process.

I knew from the data I looked at with the Plan that our problem was not that we didn’t have gigabit networks everywhere, but that we had them nowhere. But the problem was crystallized when a network engineer told me in the Gig.U process, that he hoped his company would build a gigabit test bed because ‘no engineer ever went to GM to work on a used Chevy. But that is what I am doing.’ We need to provide our engineers the challenge of working toward the best, the network to come.

It is on those world-leading networks, that we can combine new functionalities with world-class area expertise to invent what will become big bandwidth based medicine, education, business services and improvements in so many other areas.

It is on those networks we can create, in effect, a Bell Labs 3.0, a community of collaboration brought to us not by the benevolence of a monopoly of the industrial age but, as the recent book on innovation by Brynjolfosson and McAfee put it, by the “parallel experimentation by millions of entrepreneurs (which) is the best and fastest way (to drive innovation).”

That insight, on the need for world-leading networks to drive improvements for all, lead to the Plan’s recommendation that we need to always have a critical mass of communities with world leading networks; networks that don’t just follow demand but networks that lead demand.

In short, we should have the goal of having a strategic bandwidth advantage that delivers to all Americans the bandwidth we need to be our most productive today and to lead in the creation of a better tomorrow.

The second thing we need is a psychology of bandwidth abundance.

A psychology of abundance is critical to the development of the entrepreneurial mind. The biographies of the founding entrepreneurs, from Gates and Allen, to Andreessen, to Zuckerberg, all involve decisive moments in which seizing an abundance of computing or network power enabled them to see new opportunities.

This psychology is not just about entrepreneurs.

Unlimited, or big bundles, drive use that in turns drive innovation, investment and jobs. Consumers considered wireless a luxury until AT&T introduced the One Rate in 1998; compelled to do so by new competitors and enabled to do so by a policy that dropped the rate wireless providers paid wired phone companies to connect calls. Average minutes of use then soared fivefold, with subscribership and revenues growing as well.

The same story is true for the Internet. AOL initially resisted unlimited packages, but when the FCC rejected the phone companies’ efforts to apply per-minute charges to Internet use, and a new entrant started selling unlimited packages, AOL did the same and Internet usage in the United States soared. Europe kept the per-minute model and their use lagged.

In both cases, we saw market forces at work but the dynamic also involved policies designed to, again citing to Chairman Hundt’s article, “assure low-cost access to key inputs that enable long-term investments in transformative infrastructure…(and) remove barriers to replacing old methods with the new.” In these two cases, a combination of removal of high terminating access charges and access to bandwidth, unleashed market forces that provided consumers that psychology of abundance.

The same story reappears with the smart phone. The Chairman of AT&T recently said that unlimited data was a mistake for his company. Methinks he dost apologize too much. Our country benefitted enormously. The AT&T/Apple partnership launched the apps economy, placing the United States at its epicenter. It’s already created almost a half million jobs. By 2015, it is projected to generate nearly $40 billion in sales.

Further, it was not a ‘mistake’ so much as a necessary response to the competitive landscape in the mid-2000’s. AT&T needed a competitive differentiator and the iPhone, without an unlimited plan, would have failed. Further, AT&T knew that more spectrum was soon on its way, thanks to a decision a decade before to utilize the DTV transition to obtain a “digital dividend” by moving spectrum from broadcasting to mobile.

In all three cases, the United States economy benefitted from consumers enjoying a psychology of abundance. As Amar Bhide notes in the book “The Venturesome Economy”, entrepreneurial activity is more likely to succeed here than elsewhere because of consumers willingness here to try unproven products. In the case of the Internet, mobility and 3G apps, the willingness of the United States consumers to try new things make them the best market to build for, a big plus for investment here. But consumers’ willingness to try new things depends the cost of the trial being their own time, not an incremental payment.

In short, knowledge exchange, our principal economic and social activity, benefits from a strategic bandwidth advantage and a psychology of bandwidth abundance. Therefore the prime policy mission should be to produce that advantage and psychology.

Second, the current market dynamics and regulatory structure in the United States will not deliver the wireline upgrade necessary to produce a strategic bandwidth advantage or a psychology of bandwidth abundance.

As a starting point, we should aspire to the creation of bandwidth proceeding at the pace of Moore’s law, so that we see increases in abundance similar to those we see in technology.

As to the bandwidth of Internet transit, we do.

Thanks to the progress of optical fiber communications, over a period of 25 years, the system capacity has increased by almost 5 orders of magnitude.

As a result, the price per mbps has dropped from nearly $1,200 per to less than a buck.

We don’t see that kind of advancement in the access portion of the network, but there are positive elements of the United States access network. We lead in building out and using 4G wireless networks and thanks to DOCSIS 3.0, a greater percentage of our population has access to a 100 Mpbs networks than all but a few countries.

From a perspective of use of wire line bandwidth, however, we are mid-tier and numbers suggest a pattern of decline. And as discussed above, if we want to lead, we want to lead in use, not theoretical access.

There are other numbers that are more positive for America but none show leadership in use, none show positive trends and, none show that any individual American cities are hubs of abundant bandwidth leadership where new big bandwidth sectors will likely be birthed.

But let’s turn our attention to, as Wayne Gretzky said, not where the puck is but where it is going.

Here’s a big fact.

For the first time since the beginning of the commercial internet, the United States does not have a national wire line provider with plans to build a better network than the currently best available network.

That means that for most Americans, five years from now, the best network they have is the network they have today.

That is not going to be true in a lot of other countries. Our wireless leadership is great for the mobile transformation occurring now, but looking down the road only wireline can provide the excessive bandwidth that provides the platform for the creating the next generation of big bandwidth services.

We enjoyed five wire line upgrade cycles, with cable and the Telcos seeing and raising the network functionality of the other.

But that poker game is over. Verizon is finished with its fiber rollout. Both it and AT&T are focusing their capital expenditure on building out wireless networks. One could point to Verizon’s recent 300 Mbps announcement—which discussed bigger bandwidth as the foundation for new services, sounding like a Gig.U press release—or AT&T’s recent statement that maybe they will upgrade rural lines, as counterpoints. Still, I can’t find anyone on Wall St. who is betting that any Telco will try to invest new capital expenditures to build a better network than Cable.

Cable could upgrade, without much difficulty. Ironically, that is part of the problem. Wall Street knows that an upgrade for the telcos is probably a losing, and certainly a risky, proposition. Telcos would have to spend orders of magnitude more than cable to produce a world-leading network, and would be unlikely to ever make the money back, as cable could under price them for the same functionality.

But if the telcos don’t press cable, cable cannot justify any incremental spending to upgrade its network. No company has overbuilt its own network without external pressure.

As noted Wall Street analyst Craig Moffit recently reported, “the numbers suggest that (cable) broadband is an almost comically profitable service, with direct gross margins of about 97%.”

No doubt Wall Street has a different view of comedy than Hollywood but to be clear, I am not criticizing Cable for achieving that margin. They invested at great risk. Not long ago, AOL and the telcos tried to get regulators to kill cable’s ability to invest in broadband and they almost succeeded. Fortunately, they didn’t and cable has delivered a platform that for today’s purposes is great.

We shouldn’t blame cable for succeeding or the Telcos for a rational capital allocation strategy, but we should also not confuse ourselves about the future.

When it comes to the wireline access network, instead of talking about upgrades, we are talking about caps and tiers. Instead of talking about investment for growth, we are talking about harvesting for dividends.

The Government should not manage the capital allocation decision of private parties but it should recognize that our progress demands an investment environment that creates the conditions that allows us to invent the future, not just harvest from the past.

The Government should not attempt to micromanage the packaging or pricing of a service. But those in policy positions should understand this: a country that is talking about an upgrade, and not caps, will be better off in a few years; a country that is talking about caps and not upgrades, will not be better off in a few years, and likely will be worse off.

It is both wrong and counter-productive to blame any private sector entity for the situation. Rather, we should try to understand we don’t see that abundance seen in the transport portion of the network reflected in wireline access upgrade plans in the United States.

In this regard, two equations provide insight.

First, what are the total benefits of a new or upgraded network?

Figure 1: The basic equations for addressing the barrier

They are the sum of the benefits to the private investor, the third party content and apps creators, the local community, the region and the country.

But the only benefits that matter to the investment decision are those that matter to the investor.

And as to those, it is a simple case of costs v. benefits.

Figure 2: The returns do not justify the investment


Or, broken down, the relationship between the new or incremental capital expenditures and operating expenditures v. the risk adjusted revenues, plus the system benefits, plus the threat of competitive losses.

Today, the cost side is greater, and therefore no upgrade or new network is in the offing.

It is actually worse in the United States because, as we spread the cost of network across two platforms, the risk-adjusted revenue is cut in half. That would be balanced if today there was a threat of a competitive loss without an upgrade, but that is not the case today.

How do we address that problematic math?

Changing the math is conceptually not that difficult.

To reverse the relationship between costs and benefits, we need to have an environment that lowers capex, opex and risk, and increases revenues, system benefits and competitive threat.

Figure 3: Changing the math

Easier said than done but we have done it before.

And that leads to my third point.

Third, the next upgrade is not only the best way to achieve a strategic bandwidth advantage and psychology of abundance; it is also the best way to address the traditional four questions of telecom policy.

Here is a chart, listing all the major new networks or upgrades in the United States.

Figure 4: Network build-outs, upgrades and ecosystem changes

While these networks primarily were built with private money, every upgrade or new network was preceded by a change in public policy that changed the math in one or more of the six ways discussed above.

Further, when Gig.U engaged broadband vendors and service providers in an RFI process, the answers reflected numerous ways other beneficiaries, particularly local communities, could act to make investments in next generation networks possible now.

These are like the actions the state and local government took in Kansas City, creating the environment in which Google could provide the community a world-leading network. It is what state and local officials did with Gig.U projects in Maine and Florida, and is also what is at the core of proposed $200 million commitment by a private company, Gigabit Squared, to build six gigabit test beds in university communities. This announcement creates a private version of a race to the top for communities seeking to bring world-leading networks to their people.

Two lessons from that RFI process stand out as relevant.

First, any community that wants its residents to have access to a Gig can do so. The barrier is not technology or economics. The barrier is organization; specifically, organizing demand and the better use of underutilized assets.

Second, university communities have the greatest motive and easiest organizing task. Big data communities have the greatest existing demand and as high bandwidth-based enterprises are born, they will want to locate there. University communities have the assets and organization tools best suited to the task, which is why our effort to accelerate next generation networks begins in those communities.

There are other projects—such as Google Fiber and U.S. Ignite—and a myriad of policies at the federal, state and local level that could also assist in the mission.

But the purpose of my talk is not to catalog the policies; rather it is to sell the primacy of the mission.

Part of the reason I don’t think now is the time for a laundry list of policies is that the Gig.U is in the midst of discovering the most effective levers. I can say with some confidence, we can accomplish this mission without big, new spending or a new regulatory regime. After all, in less than a year, a small budget and no regulatory power, we have already stimulated gigabit developments for 8 communities. We should continue to let the data guide us toward what works; though Chairman Hundt’s admonition to policy makers to “assure low-cost access to key inputs that enable long-term investments in transformative infrastructure…(and) remove barriers to replacing old methods with the new” appears to be the right recipe.

Moreover, while one can’t predict how policy options arise, if in the next few years, policymakers ask themselves as they review their choices, which option will drive a strategic bandwidth advantage and which option will drive a psychology of bandwidth advantage, then we will accelerate the day when we reach our goal.

If we ask the right question we have a good chance to get it right; if we ask the wrong question, our chances of getting it right are slim. Without a Northstar, we are like the sailors in the race of whom it is said, they were lost but they made great time.

There is another reason for this focus on upgrades—one generally misunderstood in policy discussions. Simply put, if we are smart about that mission, then that new bandwidth and that abundance will get networks everywhere, get everyone on, generate a competitive dynamic and drive better uses, far more effectively than any other means.

If you think about the history represented by that chart, where we most succeeded in making progress on addressing the four questions was when we created an environment in which private sector actors either saw the opportunity, or as with AT&T and the iPhone, saw the danger in not acting, to provide Americans with an upgraded level of functionality and a new psychology of abundance.

So today, we can spend billions of dollars connecting rural America to baseline broadband by building on top of old technology, or we can figure out an upgrade strategy using new technology to bring far bigger broadband at far lower prices.

We can spend billions trying to get everyone one a network, or we can create upgrade options for low-income individuals through the utilization of untapped resources in the existing network can bring a compelling value such that market forces do most of the trick.

We can, like Korea, mandate spending billions to upgrade everywhere to drive more effective use of the network, or we can upgrade in those places we know have, and are likely to do so in the future, create the kinds of improvements that scale everywhere and create new market forces that incent the private sector to invest in a broader upgrade.

Competition is trickier. Many upgrades created a new asymmetry in the market that triggers a competitive dynamic but not all do. Still we should understand that it is And we should understand that after that that the answer to creating a competitive dynamic that increases consumer welfare lies not in a rule that slows everyone down but in an environment that speeds somebody up.

That is why, I hope next year, the President of the United States tells the Chair of the FCC that his or her mission is to deliver a strategic bandwidth advantage for the country and a psychology of bandwidth abundance for consumers, and that the nation will holds the Chair accountable.

On a personal note, I hope it is the same President and the same FCC Chair. But whoever it is, this mission is, and should always be, bipartisan.

Let me close with my favorite answer to the question of what happens if we eliminate bandwidth as a constraint on innovation.

It would be enough for Big Bandwidth to improve knowledge exchange to accelerate curing that doctor’s patient of cancer, helping the illiterate child to read, retraining the unemployed, or providing the partner in collaboration Big Data will need.

But my favorite answer came at the event announcing a gigabit project in Maine. Many testified eloquently of project’s importance but none more than Nate Wildes, a student, who said that while having the world’s fastest network in Maine would enable the business he started to work better and bring “immediate benefits” to the state, the most exciting thing about unlimited bandwidth “is what we don’t know yet.”

In this regard Mr. Wildes is following in the great American tradition of those who travel with hope, energy, and urgency but no map. In that tradition, whether it be Columbus, Lewis and Clark, the astronauts or the fathers of broadband, we don’t follow a map; we make the map.

I hope today, and in conferences like this to come, we make the map to discover the undiscovered country.


- Blair Levin, Executive Director


June 13, 2012