It’s time for a break from the sturm und drang of the political season. Herein an exploration of a macro trend that bodes well for America – and for investors. Of course governments are capable of either facilitating or stifling tech revolutions. But progress is (nearly) inevitable when tech cycles are foundational.
“The revolution we’ve witnessed since the 2007 introduction of the iPhone has emerged mainly from the fact of the network that links the local to the remote, and less from the fact that the phone became a computer.”
And as a bellwether of the next Big Thing consider the ‘overnight’ success of tiny tech company Impinj [NASDAQ:PI]: it ‘leapt’ from its founding 15 years ago to successful IPO last month. It is a company that epitomizes the class of technologies making possible the Internet of Things (IoT), a transformation about which there is already much hyperbole. But this time the hyperbole actually understates the prospects. The misnamed IoT is in fact the next wave in communication revolutions. To understand why and to go beyond superficial hype—common and excusable in the popular press—we need to explore some of the underlying concepts (which I’m previewing here from my next book).
Impinj wasn’t co-founded by a hoody-wearing Millennial college dropout but by Carver Mead, a retired CalTech prof who is one of the great living physicists from the 20th century’s unique confluence of many luminaries who have since passed on, not the least of which was Richard Feynman, a Mead colleague.
NB for students of innovation: On average, more companies are founded by people over 55 than under 35 years old. And the latest data show that Millennials are creating companies at half the rate that Boomers did at the same age. We should all hope more Carvers emerge yet, along with more like his co-founder, hoodie-free physicist and CEO, Chris Diorio, a seasoned Gen X’r who is ancient by Silicon Valley standards.
Gauging whether the IoT, or Impinj, is a big deal or tech hype and figuring out where investors should look for related opportunities requires understanding what it is about technologies that enables networks to expand. The clue is revealed in Prof Mead’s comment—which he says with an impish glint—that Impinj’s radio frequency (RF) computer chips use zero energy. The computer and communications revolution has always been about power. Electrical power that is.
A little background: The relentless and near magical reduction in energy needed to perform a logic operation as well as the energy required to transmit data on networks has been the central defining feature that enabled – and continues to enable – the modern computer and communications revolutions. Doubt that?
No other engineered system in the history of humanity has seen such astronomical progress in energy efficiency. It takes over 10 billion times less energy today to execute a single logic operation, and to wirelessly transport one byte, compared to what was required at the dawn of the computing and communications revolutions respectively. For perspective, if you could do that in physical systems it would mean that a loaded Boeing 777 could fly to Asia by burning just 10 milligrams of aviation fuel. (A housefly weighs about 10 milligrams.) In our universe, the physics of objects and atoms doesn’t permit billion-fold gains in efficiencies. Only in the physics of information is that possible, and that’s what has generated the commensurate revolutions.
It is precisely because of the collapse in watts-per-byte that far less material and space are needed to process and, critically, transmit data. The astounding expansion in portability, ubiquity and the scale of networks is the direct consequence of this phenomenon. One billion people today use networked smartphones. That each smartphone has ten thousand times more compute power than a room-sized 1970s IBM mainframe—each of which required the power of five homes to run—is less relevant than the fact that the wireless connection to each smartphone can carry a thousand times more data than did the entire U.S. communications system in 1970s.
The revolution we’ve witnessed since the 2007 introduction of the iPhone has emerged mainly from the fact of the network that links the local to the remote, and less from the fact that the phone became a computer. While a smartphone is useful even when not connected to a network, it is the network that enabled the emergence of services and companies associated with the likes of Google, Uber, and AirBnB. In fact of the approximately 200 Unicorns—private start-ups with valuations greater than $1 billion—90% of them offer not new hardware but new kinds of financial, health, and social services possible only because of now ubiquitous broadband wireless networks.
Networks are the sine qua non defining our age and the next. In communications theory and economic practice it is the number of connections on a network that is the core determinant of impact. And this is the key: As networks expand their impact expands far faster. Nearly four decades ago Bob Metcalfe, who invented Ethernet and is now a prof at U. Texas, Austin, posited that the value of a network is proportional to the square of the number of users on it. (Credit George Gilder, in his epoch-setting book, Telecosm, with labeling that phenomenon Metcalfe’s Law.) So doing the math, a network with 100-fold more users has 10,000 times more value; a thousand-fold more users yields a million times more value.
Metcalfe originally intended the concept to capture a qualitative rather than a precise quantitative effect. Nonetheless, there has been some academic dispute over the accuracy of the ‘law’ as an economic predictor. Experiential evidence aside—witness the historical impact of the telephone vs the telegraph, the mobile Internet vs the desktop Internet—Metcalfe’s published a 2013 retrospective showing that Facebook’s market value has in fact tracked to a close approximation the square of the growth in its users. Of course in the real world Facebook and other similar companies create more overall economic and social value than is measured by a stock market valuation. Perhaps the only thing Metcalf got wrong is the network multiplier might be even greater.
Each of history’s four communications revolutions is thus best viewed through the prism of the scale of connections: telegraph (thousands), telephone (millions), cell phones (hundreds of millions), and smartphones (billions). There are clearly billions more people yet to connect to ‘mobile’ and while that’s important it’s not the next quantum leap. It will require a phenomenological change in underlying technologies to enable another epoch-setting order-of-magnitude increase in a number of connections.
Clearly the emergence of what has thus far been inartfully called the IoT tautologically entails enough “things” on the planet to count in the trillions. A critical Rubicon was crossed on or about 2008, the year that more things than people were connected to the Internet. In some ways that was not particularly remarkable because big things (like trucks, big motors, power plants and commercial HVAC systems) have long been networked and long outnumbered people, and those kinds of networks expanded as connection costs dropped. Every Uber car is connected as are most new cars and fleet trucks and GE aircraft engines, for example.
However, IoT enthusiasts have been disappointed with the current scale. Early forecasters a mere half-dozen years ago expected a trillion connected things by now. That hasn’t happened. Best estimates put the IoT world today at a ‘mere’ 10 to 20 billion connections. So we have a race amongst pundits to tamp down IoT expectations.
It may seem odd to call the number 10+ billion small, but it’s all relative. And networks start small in part because they begin when the associated technologies are good enough to launch but not yet good enough to fully scale. In 1853, North America had only 100 telegraph stations; people had to go to a telegraph station the same way they would go visit a computer mainframe one century later. The scaling pattern has been the same each time. Only 2 million phones were in homes by 1905—a technology that was widely recognized as revolutionary at the time—even 25 years after launch. The number of cell phones was still only 30 million a decade after launch.
When it comes to the state of today’s IoT, obviously there are many more things that could yet be connected beyond what early adopters have done with pipelines, pumps, cars, tractors, and medical hardware. The number of potential connections logically extends deep toward the bottom of infrastructures, things and people. Useful information is associated not only with the steps or heartbeats we measure, but also with the trillions of switches and nodes in machines and infrastructures.
As networks expand we can expect the trajectory to be the same as in each previous expansion. First comes the enabling new class of electronic components, then the transformation and expansion of infrastructures, and then collaterally we see existing and entirely new businesses capitalize on the associated opportunities, services and new modalities of operating.
We have, as they say, seen “this movie” before. Most recently, the components (CPUs, RF chips, sensors, batteries) got good enough to enable an iPhone; contemporaneously, engineers conquered the spectrum’s seeming limits, unleashing broadband speeds wirelessly (first 3G then 4G, soon 5G). Then came entirely new companies, the likes of Uber, Paypal, AirBnB, along with many others in the pipeline, taking advantage of a billion wireless ‘nodes’. The velocity of the post-2007 disruption from what is now simply termed “mobile” came from the ‘overnight’ success of the underlying technologies stretching back over the previous decade or two.
Enabling real-time communications connections with or within nearly any thing will require even greater reductions in power than have already occurred. And this time the keys to the kingdom will go not so much to companies that further cut the energy cost of logic (though that remains important), but mainly to those able to further cut the energy costs of the connections themselves – i.e., the sensors that connect to and measure the physical world’s attributes, and the RF chips that wirelessly connect that information to the outside world.
A full exploration of all the players and enablers in the emergent IoT ecosystem is beyond this column’s scope. Instead, herein some key metrics and names of some of the early iconic leaders in the four domains found in every network ecosystem: enabling components, network infrastructure, software, and new business models.
IoT Components
Every tech revolution is propelled by the emergence of a new class of enabling components. For the telegraph it was the magic of the new electric dynamo, copper wires and glass insulators. The latter was Ezra Cornell’s key invention, leading to his role in building the 19th century’s ‘Victorian Internet’ that made him a billionaire (measured in today’s dollars), equivalent in relevance and scale to any Silicon Valley titan today. For the computer, of course, it was the vacuum tube and tape drive. The PC came from the large-scale integrated circuit (IC) and the floppy drive. The cell phone was made possible by low-power ICs and a new class of low-power semiconductor radio (RF) chips. Each wave in the evolution of communications networks came from components that radically reduced the energy needed to acquire, manage and transport information.
The progress in compute-communicate energy efficiency over the past half-century has been nothing less than astonishing. Yet we need even more fantastical gains in power efficiency in order for logic and (necessarily) wireless connections to shrink down to the scales where the edges of the networks can be attached to anything from socks on retail shelves to laminations on your skin, and from individual components inside any machine to the very components of your body. Thus at the core of the IoT we need the near-zero-power IC and near-zero-power RF chip.
In order to appreciate the scope of this challenge, consider that the energy distance between the ARM chip in your smartphone and a nanowatt-class chip to unleash the IoT is comparable to the power difference between a Saturn 5 moon-rocket and a backyard drone.
Which brings us back to Impinj where, as noted, Carver Mead likes to claim they’ve conquered the nanowatts goal by going down to zero. In physics (about which Carver needs no education) nothing of course can in fact operate with zero power. However, engineers are always good at the work-around. One can design a chip that requires zero on-board power by wirelessly sending power to it from elsewhere. This idea originated more than a half-century ago with the invention of the Radio Frequency Identification (RFID) chip. The power comes from an external source (a so-called “reader”) that beams RF energy at an unpowered chip which then animates the chip so it can send a radio signal back to the “reader.” An RFID chip needs no battery, wires or energy harvester on board. The concept should be familiar since it’s in every one of the millions of EasyPasses in cars.
RFID chip tech is well-suited for knowing any thing’s identity, location and authenticity. And there are hundreds of billions of things for which just that much information can be invaluable; not just vehicles but also, for example, the parts from which they are made during the manufacturing process. RFID tech is useful as well for tracking and monitoring things like clothing, food (re spoilage and safety), luggage, medical supplies, or boxes in warehouses or trucks, or parts or objects of any kind in the world’s massive labyrinthine supply chains. RFID is also used to track Disney guests, farm animals, and athletes in sporting events. Notably, RFID works in a part of the radio spectrum less susceptible to interference than other wireless devices.
Carver Mead’s physics insights—about which he has had manifold in many domains: “Google him,” as they say—from two decades ago at CalTech, led to the founding of Impinj. One of Carver’s students from that time, Chris Diorio, is today Impinj’s CEO. At the simplest reduction, Impinj makes a next-generation ultra-efficient RFID chip. The company also makes the associated “readers” (a business unit built on the earlier acquisition of Intel’s “reader” technology) as was well as Cloud-based software and analytics to make seamless the utility of the associated data flood. Taking an important lesson from history, Impinj has clearly decided to engage more of the ecosystem than merely the enabling chips. Smart.
Evidence of its vision was visible two years ago when Impinj successfully forged a global alliance, first partnering with Intel, Google,Smartrac and now with over 100 global members. The alliance, called RAIN—a new acronym for radio frequency identification intended to incorporate the analytics—was built to facilitate next-generation RFID deployment via standards, in particular on how data about things are stored and managed in the Cloud. The idea, never mind the reality, of the Cloud didn’t exist when first generation RFID was developed. As the RAIN alliance is happy to note, and we note that Diorio is RAIN's chairman, the spectrum is free in 78 countries.
Impinj is not alone even though it is in dominant in the RFID space. Other RFID companies include NXP, Alien Technology, AMS, Phychips and Zebra Technologies. But for a full picture of Impinj, it’s bragging rights to 200 patents and history with 13 billion tags sold to date, read the good bits in the company’s recent S1 SEC filing for their July 2016 IPO, where one often finds more useful information than at corporate Web sites. Impinj happily notes that Macy’s saw a nearly 10% jump (in retail land, 10% gains are “jumps”) in same-store sales and a $1 billion inventory reduction using RAIN across 850 stores. Impinj points to similar kinds of inventory savings at hospitals and a 30% jump in “capture” of billable medical information. In both these domains RAIN-class tags are still used on less than 10% of “taggable” clothing or healthcare items.
The universe of things that are ultimately taggable with RAIN chips, which today cost pennies (and soon far less), is far greater than the necessarily cautious claims made in SEC filings. And there is much more to the IoT world than RFID chips.
Sensors & Semiconductors
The next IoT frontier is to go beyond identify-locate-authenticate where RFID plays, and to actively measure something about or within a thing or a person; e.g., temperature, strain, vibration, velocity, chemical identification, etc. To do that requires exquisitely sensitive sensors, i.e., those that are tiny enough and able to detect and measure using trivial amounts of energy. (I should note that while Impinj makes no claims about pursuing these capabilities, one would be unsurprised if such were in their skunk-works.) The world of sensors is fortuitously and portentously on the cusp of radical advances.
More than a billion people are already unknowingly familiar with accelerometers, one of the most useful sensors. The accelerometer in your phone detects when you tilt your phone, are walking, even breathing. Compared to just a decade ago these microscopic MEMS-built silicon devices are now half the size, use 70-fold less power, and cost a-tenth as much. Cameras of course are sensors and are now so tiny and efficient that they are not only on every smartphone but can be embedded in swallowable pills for internal diagnostics. Printed inks and conductive yarns are enabling clothing, objects and skin to adopt sensing functions. Electronics assembly giant Flex [NASDAQ:FLEX] formed a partnership with a fabric maker, MAS Holdings in one sign of convergence between the digital world and the clothing world.
And when we consider the biological world of plants, animals and humans, the potential number of ‘things’ that can be usefully measured rises into the stratosphere. The potential for biocompatible and conformal sensors that can be tattooed on or implanted in living tissue is rapidly advancing. The opportunities for real-time monitoring and diagnostics (and therapeutics) seem like science fiction, but are rapidly emerging realworld industries. For a particularly lucid and detailed exploration of is now emerging in the world of bio-compatible electronics, read a New Yorker piece from two year ago about that domain and the brilliant work of professor John Rogers (now at Northwestern University). Or read anything written about or by Scripps digital-medicine pioneer Eric Topol. At least one forecast sees the worldwide health-care IoT market exceeding $400 billion well before a decade passes.
Pushing to the limits of tiny and wirelessly connected though will still require power, and unlike RFID, will require active local on-board power. There are numerous ways for embedded chips to harvest power from the surrounding environment: vibration, noise, light, heat; even ambient RF fields create a ubiquitous sea of energy in our environment. But the reality of physics in the nanoworld means that chip-scale power harvesters produce energy measured in nanowatts. Just how tiny these scales are is hard to fathom: a single flea hop generates 100-fold the energy of a nano-energy harvester each second. And it takes a 100-flea-hops of energy to transmit just a few hundred information bits per second, never mind kilobits or megabits. But as incredible as these metrics sound, they are now being sequentially conquered in the engineering of the nano world.
In the meantime, the sensors and semiconductor logic are already here and deployed to make “connected” cities, homes, and vehicles a reality and big business today where the power challenges have already been consquered. Semiconductor sales into IoT applications are running at $18 billion a year and growing 20% annually. It’s a domain with a variety of MEMS sensor companies like InvenSense [NYSE:INVN], and communications companies like Cypress Semiconductor [NASDAQ:CY] that recently acquired Broadcom’s IoT business unit. And some wild cards on the horizon to think about: watch for an IPO from next-generation semiconductor companies like Efficient Power Conversion, GaN Systems, or Transphorm, all pioneers in the use of gallium nitride instead of silicon which promises to enable small, efficient systems for wireless recharging.
Networks
The architecture of the IoT ecosystem a priori requires networks to transport information to remote users and especially the Cloud. Here the technical differences between smartphones, which have driven network growth thus far, and smart things leads to different protocols and even underlying designs within the conventional cellular networks. First, the superfast video-centric bandwidth of conventional cellular networks is not only overkill for the vast majority of data transport needs in the IoT, but more critically, is far too energy hungry and expensive.
Thus we have recently seen entirely new IoT standards emerge for transporting data on cellular networks that are based on using just one-tenth and one-hundredth the available and standard cell bandwidth. (Remember, speed and bandwidth directly correlate to energy use.) These new IoT-centric wireless standards are intended to allow IoT-specific things to work on the existing cellular networks, thus piggy-backing the deployed physical infrastructure. But using the new low-power standards will require new classes of lower speed and lower power chips. We can expect Qualcomm, Intel and Altair (owned by Sony) to develop IoT class RF chipsets useful on the LTE cell networks. So far though, the first to market is France’s nimble ($100 million market cap) and innovative Sequans Communications [NYSE:SQNS].
It is also likely that entirely new networks will prosper that are optimized specifically for the different characteristics of IoT connections of low power and low-bandwidth, in particular those using frequencies that have more immunity to interference and greater ability to penetrate walls and objects. The latter is of particular importance for gathering information from and especially controlling real-time activities in the physical world (e.g., autonomous vehicles) where there can be no tolerance for the kind of ‘noise’ or dropped connections still too common on the voice and video wireless networks. Several players and standards are already rolling on the path to build out such entirely new low-power wide area (LPWA) radio networks dedicated to IoT; companies like Ingenu, LoRa and Sigfox. While it will be a daunting challenge for new networks to compete with the established cellular infrastructure, if there’s one thing we’ve learned from past communications tech revolutions, it is dangerous for incumbents to ignore disruptions from the edges and at the ‘bottom’.
Datacenters & Software
Network infrastructure expansion of course emerges from and is driven by data traffic. Cisco has long been tracking and forecasting Internet traffic in general and for “things” as well. In its June 2016 forecast Cisco sees machine-to-machine connections as the fastest growing contributor to total overall Internet traffic over the coming five years. And between now and 2020 machine-related data will grow 600%. By then machine connections will comprise nearly half of all Internet connections of all kinds, including TVs, mobile, PCs, etc. Cisco thinks medical and automotive applications will dominate growth in bandwidth consumption.
Consequently, at least one industry insider sees IoT’s growth requiring as many as 4,000 new enterprise-class datacenters. For those that follow datacenter power issues, that much new infrastructure at the head end represents about 200 GW of power demand, a scale equal to the national grid of India. This may explain, in part, why some investors are so bullish on datacenter REITs, including Iconiq Capital, an investment firm that reportedly manages funds for Mark Zuckerberg.
In one bellwether as to whether we should be bullish or bearish about the IoT’s future consider that Google is rumored to be developing a new OS for the IoT, just as they created Chrome for PCs and Android for smartphones. Google only thinks big. It’s a logical move if the IoT is big enough to support its own OS. And it is logical in a technical sense because an OS optimized for crop and heartbeat monitoring or the movement of packages in warehouses and on drones has technical (and thus software) features at least as different as those between PCs and phones.
The IoT will also require and is already generating entirely new classes of software for security, systems management, and analytics. Many of the new companies will doubtless in coming months and years become acquisition targets by the existing Internet giants. CBInsights, which tracks venture investments and especially the unicorns (private companies with $1 billion valuations) as icons of enthusiasm or hype, there are already two unicorns in the IoT cybersecurity domain: ForeScout Technologies and Zscaler.
As with the human Internet, the machine Internet will require a wide array of security solutions as well as security embedded in the physical silicon of the chips at the edges of networks and inside the things. Just one telegraphic example in this domain: at the top of the food chain we find $380 billion market-cap Amazon partnered with $13-billion market cap Microchip [NASDAQ:MCHP] to develop “end-to-end” encryption that begins in the silicon inside the end-use IoT chip synchronized up into the Cloud.
There’s much yet to evolve to fully build out the basic Internet architecture to deal with a different character of data traffic; i.e., the shift from network devices that are dominantly “consumption-based” – i.e., information like movies and maps that mainly flows from the top (at datacenters) down into user devices – to devices that, on average, operate in the inverse way, i.e., where information is produced at the edges and flows up into storage and analytics engines. Some of this traffic shift is already happening; e.g., when consumers use the ‘things’ feature of their smartphones, notably GPS location information.
There is much more to know about many objects other than location in real time. In fact an entire constellation of physical attributes of things is where we expect the knowledge of things to become useful and even actionable. The IoT accelerates the opportunities to manage physical events in real time, the most obvious example of which is any autonomous vehicle, but this is also true in manufacturing and supply chain systems. Data about things that are collected and analyzed requires information and instructions that flows back to where the things reside in the physical world. This traffic feature of the IoT is quite different from the Internet as we mainly use it now.
Shift the direction of primary traffic flow has implications for any infrastructure, whether oil pipelines or highways or the Internet. And when it comes to the IoT there is also an unavoidable feature of physics to consider: the speed of light. The speed of light determines how long the roundtrip in goring from acquiring information at the edge of the network, to the Cloud and then back to the edge with the result or instruction. Even light is not fast enough for real-time control of many things using big centralized and remote datacenters (where the ‘thinking’ part of ‘smart’ things generally takes place). Thus many datacenters will necessarily move close to the action near the edges of the network and not remotely sited where power and land are cheap and easy to access. This reality doesn’t mean the end of the big remote datacenters (many other requirements are driving a boom there), but it does mean we can expect a further proliferation of so-called edge datacenters and ‘micro-datacenters’ that are already important for delivering jitter-free video to mobile devices. This is where the telecos may yet emerge as important players or partners as datacenter functions spread out even further towards the edges.
We see yet another bellwether of the emergence of new classes of IoT engineered systems with the creation of an IoT-dedicated business unit within VM Ware [NYSE:VMW], a Dell owned company that remains public. But the macro indicator of the rise of IoT is not so much what the traditional mega-Internet companies are doing but in the convergence of goals and solutions with the traditional mega-Industrial companies where “things” are made or reside. From GE and GM, ABB and Siemens, to Parker Hannifin and Boeing—wherein Boeing recently moved to Microsoft’s Azure Cloud for IoT class analytics from the 300 airlines using its aircraft—we see programs, projects and pronunciations of IoT-centric futures that are not rooted in Silicon Valley hoody-wearing wishful thinking, but in old-school industrial practicalities seeking competitive advantages in the increasingly competitive global markets. Spend a few minutes at Teradata’s [NYSE:TDC] Website and the focus is obvious regarding industrial opportunities to turn petabytes into profits using IoT analytics. Or visit the Website of analytics firm Splunk [NASDAQ:SPLK] which was founded within the machine/thing world.
The future for analytics of things is so vast that it won’t be owned just by large or incumbent players (at least not for a long time). There is far too much specificity for applications when going from analytics of say human blood flow versus trucks. Many seemingly prosaic applications will have substantial impact when multiplied by the scale of what society does. The potential is epitomized by niche players like Fathym that brought a 10x reduction in sensor cost to snow-clearing trucks in Fairbanks, Alaska to monitor things like air and road temperature, barometric pressure, humidity, windshield wipers in order to perform analytics on road clearing and de-icing. Or there are examples at greater scale like ARI Fleet Management which manages more than 1 million things with wheels (trucks and corporate vehicles) and now collects more data in two weeks with IoT systems than they have accumulated over the past 20 years.
Forecasts, Dreams & Realities
But to circle back to where we started: The data also make clear it is still early days for the IoT. Research firm IOT Analytics recently published the results of a comprehensive study to figure out what kinds of real-world IoT systems are actually operational rather than aspirational. They did so by “mining” corporate Web sites as well as published reports on actual projects. (IOT Analytics excluded consumer wearables in this particular analysis.) They found 640 deployments where those in industrial and urban (“smart city”) applications accounted for nearly half the entire count. Notably, they found oil & gas one of the biggest industrial players. The retail and “smart supply chain” together accounted for only 8% of the projects found and smart agriculture just 6%. But in a world where we can expect tens of thousands and then millions of customer or application-specific IoT deployments this is only the start.
Both business sentiments and venture bets point to the beginning of the tip up in the proverbial “hockey stick” curve of growth once new technologies take hold. Vodafone started four years ago publishing an annual IoT Barometer based on a global survey of business leaders. In this past summer’s release, the survey found 89% report increased IoT budgets; 76% think IoT will be critical for future competitiveness; and 63% claim “significant” returns on IoT investments. And check CB Insights visually elegant map of $7 billion invested by “old guard” industrial companies in 56 start-ups targeting factory floor transformation.
Another window on the future? The lens of trends in venture capital spending. Stipulating that venture investing often reflects not just what engineers are doing on the leading edge but, truth be told, also pack mentality and fashionable tropes. Still, consider that IoT venture investing has more than doubled in the past several years, hitting $3.2 billion last year. (Although the data suggest a modest slow-down to a skosh below $3 billion invested in 2016.)
Still, the failure to reach a trillion-node IoT by now has some putting the knock on IoT as mere hype. They are sounding a lot like the reaction of pundits and economists that put the knock on the promise of the Internet during the late 1990s. For an iconic and much circulated example of that myopia read Nobel-winning Paul Krugman’s column from 1998, when he wrote:
““The growth of the Internet will slow drastically, as the flaw in ‘Metcalfe’s law’ …. [b]ecomes apparent: most people have nothing to say to each other! By 2005 for so, it will become clear that the Internet’s impact on the economy has been no greater than the fax machine’s.””
To be fair, Krugman recently tried—in my view unconvincingly—to ‘context’ his 1998 column, and he wasn’t alone in skepticism about the Internet’s impact. Only two years earlier in another well-known and oft-referenced example, Federal Reserve Chairman Alan Greenspan referred to Internet enthusiasm in 1996 as “irrational exuberance.” They weren’t alone in their skepticism. Even earlier Nobel economist Solow famously said: "You can see the computer age everywhere but in the productivity statistics.”
The truth is that the error was in expectations about the velocity of changes to fundamental infrastructures. In 1998, Amazon was only a few years old (and selling only books) and Google was just founded when Krugman made his embarrassing pronouncement. It took another decade before it was obvious what the impacts would be from the Googles and Amazons and the broader Internet disruption and creation of entirely new businesses that subsequently emerged. This pattern will be repeated. We are at the end of the beginning of the next great network wave. It’s 1998 for the IoT.
And history has convincingly shown that as networks expand, as the edges move out, the economic and social impacts spread out not as small ripples but as a series of tsunamis through society. The prospect of networks expanding by yet another quantum leap is—as we have outlined herein—rooted in further and magic-like reductions in the power for logic and communications in order to enable the IoT to chase to the bottom of everything. Keep in mind here the iconic admonition of that great physicist Richard Feynman from 1959: “There’s plenty of room at the bottom.”
Finally, back to the politicians, and not the election. Give Congress credit with its Resolution 847 passed on September 7, 2016 “Expressing the sense of the House of Representatives” which states: “Whereas the evolution of the Internet of Things is a nascent market, the future direction of which holds much promise.” Amen to that.
This piece originally appeared on Forbes
______________________
Mark P. Mills is a senior fellow at the Manhattan Institute. Follow him on Twitter here.
This piece originally appeared in Forbes