Energy and Utilities Forum
IDC's Energy Insights
It is impossible to talk about energy without starting with oil. But let’s start in 1999, when prices had collapsed to $10 per barrel. Imagine if, in that year, then Vice President Gore proposed a $60 tax on a barrel. As many of you may recall, the Clinton/Gore Administration did in fact propose an energy tax a few years earlier, one equivalent to just $3 a barrel. That proposal earned a prompt bipartisan slap-down in Congress. Everyone seemed to agree that such a tax would reduce oil demand, but everyone worried the tax would hammer the economy.
To note the obvious: today oil trades at 20 times the decade-old Gore tax benchmark, Wall Street, while jittery, remains in high territory. And, the demand for oil has actually risen modestly. So what happened?
In a word, technology happened. By this I do not mean oil drilling technology. Though that is important too. I am referring to a deep technology transition underway in our economy. In simple terms — the dot-com-era techno-dweebs were right, even if a little early and hyperbolic. It is a new economy. Our GDP is increasingly tech- and information-dominated. We are, in short, being assimilated by silicon. The effect is to put the U.S. economy ever more economically distant from the cost of oil, and all raw fuels.
But first, one cannot explore this trend without genuflecting to the obvious; high-priced oil price does matter. The price of everything matters at some level. High-priced oil takes money out of peoples’ pockets, which is of course especially painful for lower incomes. And high commodity prices do ripple through an economy – but it is a ripple, not a tsunami. This was not always so. Consider first a much earlier, analogous trend.
In bygone eras, and today in undeveloped countries, the price of food and the health of the agriculture sector utterly dominated national economies. In an industrial economy, while self-evidently food still matters – no one stopped eating after the industrial revolution – fluctuating food prices and agriculture viscidities can’t drown an industrial economy. And now in a post-industrial increasingly tech-based economy, rising oil prices are similarly eclipsed by larger forces. We still eat, and we still drive lots of oil-dependent vehicles. But we have passed, largely unnoticed, another pivotal economic milestone.
In order to see this transformation, the assimilation of our economy by silicon and digital technology, we have to count electrons, not barrels. All the tools, machines and appliances of the digital economy exclusively depend on electricity. 1984, when Apple introduced the Macintosh is an appropriate benchmark year to measure a core trend. Since 1984, U.S. demand for electricity has risen over 65 percent. This growth was not a result of more lights, air conditioners, motors and pool pumps, as it was during the post WW-II industrial boom. There are more of all such conventional electric devices, but engineers have made this class of mid-20th century appliances astoundingly more efficient. So much so that national electric demand should have – to steel a phrase from a perennial flat-electron-growth forecaster – demand should have at best, stayed flat as the Kansas horizon – unless there were something else new consuming electricity.
As every school kid knows, there is today a vast new constellation of electricity-consuming technologies that didn’t exist two decades ago. There are now hundreds of millions of digital devices from desktop to palmtop, used in homes and offices, on factory floors and shipping docks, and there are thousands of million watt industrial-scale digital technology enterprises from chip fabs to server farms. All plugged in to the electric gird. All part of the silicon revolution.
The reductionists way to track this digital electrification trend is just to count semiconductor silicon itself. Every square inch of semiconductor silicon shipped costs about one-kilowatt-hour <CK> just to create. And then that same square inch of silicon, once put in to use, ends up consuming, on average, another kilowatt-hour a year. Annually, we ship semiconductor-grade silicon measured in square miles.
Now, in order to measure the traditional, non-digital part of our energy, we count barrels. Since the same 1984 Rubicon, when Apple introduced the Macintosh, U.S. demand for oil has also risen, but only about 30 percent over that two-plus decade period. A growth well under half the increase in demand for electrons. Considering that over this period, the U.S. economy doubled, it requires no special insight to observe that our economic growth has been fueled dominantly by electrons not barrels, or more precisely the digital technologies consuming those electrons.
When the first oil-shock hit in 1973, despite three-quarters of a century’s electrification — 60 percent of the U.S. GDP still depended directly on barrels of oil (or the equivalent). Ironically, and not entirely coincidentally, it was in 1984, the year of the MAC’s introduction, that the barrel-share had slipped to 50 percent – and the electric share of GDP rose to 50 percent of the GDP. Today 60 percent of the GDP depends on electricity. The barrel-electron roles have completely reversed since the first oil shock. This single fact explains the unanticipated resilience of the economy to oil at $70 per barrel. The oil-based, engine, and boiler parts of our energy economy are receding in the economic distance.
This is not a new trend. It started with three guys named Edison, Westinghouse and Tesla. And was powerfully accelerated a half-century ago by another three guys: Shockley, Kilby and Noyce – fathers of the transistor and integrated circuit. More and more of our economy depends on electric devices – and thus electricity — while less and less on the direct use of oil or similar combustibles.
You would not guess such a transformation has occurred based on our current media and public policy preoccupation with oil. We are giving short shrift to the electric sector. Legislators should focus on, and business owners should assume they will not focus on, our economy’s high and rising dependence on electrons, and move up the list of energy concerns the availability, and the reliability, of electricity supply.
On the supply side, electricity shares one common denominator with the oil patch – lead times are long for major additions to wholesale production and delivery. On the demand side, reliability and resiliency (not to mention security) are increasingly vital and generally more challenging than in the oil part of the economy. This is in, large measure because of the uniquely difficult physics problem in storing any significant amounts of electricity.
To be sure, in the wake of 9/11, the great Northeast black-out of August 2003, and hurricane Katrina, there has been a spate of activity surrounding power reliability and continuity. But, it has been largely cast in the context of the transmission lines, and the utilities that own them. Lost in the shuffle is a simple truism: utilities that supply and deliver wholesale electrons can, fundamentally, only reduce the frequency and duration of outages. They can improve resiliency, and recovery speed – but only somewhat. Outages, a priori, can never be eliminated. Given the rising importance of electric power to a wide and increasing array of businesses and critical services, there is an entirely new industry focused on enterprise-level power reliability, security and continuity. This enterprise-level requirement will grow faster than, and be amplified by, the inherent growth in overall electric demand.
Running counter to this reality is an entire sub¬-industry, one is tempted to call it a cult, devoted to the proposition that growth in electric demand can, should and will soon stop.
Electrification will now abate – the argument runs – because of two principals; efficiency, and saturation. We are at the end of a century-long cycle of electrification, with saturation now being reached even in the rapid growing digital domains. We are near the end of the creation of the Internet and the PC on every desktop. And, engineers are now turning technology to make all those legacy and new digital devices ever more efficient.
There is one remarkable feature about the notion that improving efficiency will decrease demand; it is firmly enshrined as an enduring politically bipartisan principal. It is and has been in every single energy proposal, budget and roadmap. Efficiency is firmly enshrined as the key tool to cut future energy demand. But it is, unfortunately, a fallacy. Increasing efficiency is synonymous with decreasing the cost of operation – and absent punitive or sumptuary laws, lower costs do not decrease, they increase demand. The U.S. economy is 2 times more energy efficient today than 50 years ago, and we use 3 times more energy.
What about the saturation argument? The very long-run and inexorable rise in the demand for kilowatt-hours is not the consequence of some mindless use of electrons to over heat water, leave lights on and set air conditioners too cold. Rising electric demand, it hardly seems necessary to state, reflects the rising use of new or existing hardware, or different uses for technologies
Every B-school student sees the curves showing the growth-to-saturation for everything from air conditioners and TVs, to PCs and cell phones. This barely classifies as informative. The question is not whether existing technologies are at, or are approaching saturation, but whether innovation has ceased in creating new devices, or finding new uses for existing ones. The saturation of the radio didn’t predict TV, nor the refrigerator the air conditioner, nor the mainframe the PC. New devices just keep emerging, as do new uses for existing devices. Innovation keeps happening.
The hollow architecture of the saturation argument is nakedly visible in the example of the Internet and our data economy. It was only a decade ago – inconveniently, at the dawn of the Internet — that prominent no-growth experts confidently forecast demand for PCs and computing were nearly saturated. It was only a few years ago that we were told the Internet was overbuilt, both in terms of the server farms and networks to deliver digital content. The digital punditocracy hailed the advent of super-efficient blade servers that would stem the electron appetite of huge data centers, with their tens of thousands of servers inhaling vast quantities of electrons. It didn’t happen.
The technorati in the blogosphere are abuzz over a recent article by a Google engineer that observed that the electric cost to run servers may soon exceed the cost of the servers themselves. A few are speculating, hyperbolically, that since Google is now the world’s largest owner of computers, they have the world’s biggest electric bill. This is almost certainly not true – – yet.
For those who know mantra #6 in Google’s 10 guiding Philosophical statements – “You can make money without doing evil” – you might appreciate the various reactions, indeed angst, over the ‘discovery’ by some of the greenest of the technorati that megawatt-class server farms are electric hogs. From a wholesale electric perspective, computing is just another load. Microprocessors look like a lot of little heaters, that both consume electricity, and require yet more electricity for ancillary cooling – and yet more electrons for all the conditioning equipment needed to convert wholesale electricity in to the uniquely perfect form that computing requires.
The traffic growth on information highway itself is starting to heat up again too. The trade press has covered the fact that the over-built arrays of fiber optic networks, filled with so-called dark fiber, are starting to light up. Like asphalt highways, the roads and fiber pipes don’t themselves consume electricity. The traffic does. The typical wholesale user is buying connections that can handle 4-times as much traffic as just two years ago.
Looking forward on the retail side; BellSouth's Chief Architect told an audience in March that the advent of high-definition TV on-demand on the network will increase a typical users data traffic 500-fold. To serve all this traffic, it takes hardware – electric-consuming hardware. And, not to put too fine a point on it – – but there is no conceivable increase in efficiency emerging that will completely offset a 500-fold increase in data traffic.
Wall Street analysts measure this capital spending as an economic bellwether. Overall tech-firm capital spending is on a tear. And the spending is not for more chairs and parking lots, but for electron-gobbling digital hardware. Energy analysts can count these same capital trends as a bellwether for the demand for both wholesale electrons, and the ancillary reliability hardware and business services.
So we might not be surprised at the forecasts from the Energy Information Administration (EIA). Electric demand will keep growing. The EIA forecast can be put in barrel-of-oil terms as well – fuel resources equivalent to 80 billion barrels of oil will be required to fuel the lifetime needs of just the additional power plants built to meet rising demand over the next couple of decades,
The EIA forecasts the most rapid growth in the commercial sector, where over half of all the additional electric demand will come from a catch-all category called “other.” The “other” category was initially created years to ago to sweep up formerly minor electric uses, after accounting for HVAC, lighting and the like. The “other” category includes medical imaging equipment and “new telecommunications” equipment. Since one doesn’t expect an MRI on every desktop, it isn’t a stretch to speculate where most of this demand is coming from.
The demand for digital devices is ubiquitous and far from sated in every niche of today’s $12 trillion U.S. economy, and tomorrow’s $20 trillion economy. I’ve only touched on some obvious examples of new electric demand. There are many others, many we won’t easily guess.
In time, electrons will cross another energy Rubicon, directly displacing barrels in our transportation sector. While barely noted outside of techno-circles five years ago, we now hear increasingly not just about hybrid cars, but plug-in hybrids. The latter provides the option to choose the electric grid, when convenient, over the fuel in the tank, for urban driving. A pint of oil is displaced with every kilowatt-hour taken from the grid. This is no longer a fantasy given emerging battery technology — derived as well from the revolution of the silicon economy. The rap on hybrids is that they’re more hype than hope. But most of what makes today’s hybrids expensive, strongly resembles what initially made computers expensive too; silicon technologies.
And then there are the lasers. I am not referring to the ubiquitous milliwatt class of lasers in every CD and DVD player, or more powerful watt-class lasers used for optical communications. Rather, the emergence of kilowatt, even megawatt classes of lasers. Solid state, diode lasers are now sufficiently efficient, cheap and robust that industrial applications – some already in play – are now emerging as a real, and major industry. Ultra high power lasers don’t just replace thermal and mechanical processes in cutting, heating, welding and the like, but enable remarkable new manufacturing capabilities. I dare say today’s electric forecasts no more accommodate this reality than did those of the early 1980s anticipate demand from computing and the Internet.
The laser, maybe more so than the microprocessor, epitomizes why electricity is the preferred energy form of the 21st century. To briefly wax philosophical – electrification reflects a millennia-long trend towards more highly ordered energy. Energy where the entropy has been stripped out. Entropy is a very strange thing. It is easiest understood by example.
It is entropy that defines the essential difference between photons from a fireplace and a laser. There is a lot of entropy – disorder – in the fireplace’s photons. The laser’s photons dance in exquisite synchronicity with nearly all the entropy stripped away. Strangely enough we measure these two forms of energy with the same units – BTUs or their equivalent. There is nothing in the BTU numbers that can tell you the difference between these two – a gallon of each is exactly the same in BTU equivalence. We only see the difference in the price.
We pay the equivalent of a $2 a barrel for the BTUs from the fireplace. And we pay the equivalent of a $200,000 for a barrel of BTUs of a laser’s photons. The market, needless to say, recognizes and willingly pays for the higher quality – lower entropy – in the barrel of the laser’s photons because those BTUs can do so much more. And we will never get enough of those BTUs – all of which, to the point, are only made possible via a sophisticated assembly of electricity-consuming semiconductor devices.
Fortunately, the long-run supply of electricity is, for all practical purposes, infinite. From the perspective of geopolitics, energy and economic security — over 90% of our electricity is domestic, and furthermore, not produced using oil. We have centuries of supply of coal and uranium – the source of 70% of today’s electricity. And there are ample technologies to use these resources cleanly and reliably – if we so chose. Still, there’s the oil question. Not just the price, but the supply too.
While the price of oil matters a lot less now than it used to for most of our economy, it remains no less important for airplanes, ships and cars. Fortunately, prices still matter for those who find and produce oil. High prices create obvious incentives for more exploration, production and implementation of new oil technologies. High prices also make viable many more oil-efficient technologies, and fuel switching whether to corn-alcohol, coal-liquids – – or eventually electricity.
At a fraction of today’s oil prices, long-term availability of liquid fuels is hardly bleak. One need look no further than 3,000 billion barrels of oil in North American tar sands, coal and oil-shales – easily 10 times the Middle East’s resources. As we tap in to these gargantuan resources, supplies will continue to expand, prices will inevitably, eventually, soften.
This may be the last, relevant, secular cycle in oil, occurring as it does on the cusp of our transition to a new energy economy. Perhaps the days of $10 oil will never be seen again. Regardless, the days are numbered when our silicon-dominated economy will whipsawed by oil prices. We are, instead, standing at the cusp of another great tech-driven economic expansion.