Software Is Eating the World, But Steel Is The Hardware It Runs On: A Reply to Marc Andreessen’s “It’s Time To Build”
Software Is Eating the World, But Steel Is The Hardware It Runs On: A Reply to Marc Andreessen’s “It’s Time To Build”
By Sheldon Kimber
As the full extent of the pandemic and the United States’ complete and utter lack of preparation began to sink in, Marc Andreessen wrote a blog post entitled “It’s Time To Build.” As one of the world’s most powerful venture capitalists, cofounder of Netscape, and early investor in Facebook, when Andreessen has something to say, people pay attention.
His post is part indictment and part call to arms. He states that the pandemic is just one of many crises that we as a country have made worse by negligence, lack of preparedness, and an overall failure to invest in *building* the infrastructure that supports a resilient, equitable, and thriving society. We have run out of time and must immediately determine why we have stopped making such investments and reckon with the implications of the past few decades of inaction. Nowhere is this inaction more clear than in our current climate policies. “We chose not to *build*,” he notes, and it is this complacency that is the greatest threat to our country.
Andreessen admits at the end of his essay that he expects it to become the target of criticism and encourages critics to conceive of their own ways to build, instead of critiquing his call to arms. While I respect and admire his point of view and leadership efforts, and agree with most of what he wrote, I’m ready to offer another perspective. It’s not that Americans don’t want to build – we very much do – it’s that a lack of societal consensus and political will have created a market structure that encourages the “construction” of rapidly scalable software, tech, and consumer products, but makes investment in lasting, physical infrastructure difficult.
But before I detail what I think should be built and how I am a part of building it, it’s important to first discuss some of Andreessen’s observations to uncover the larger issues, ironies, and truths of his words.
There’s No Lack of Desire to Build
My concerns with Andreessen’s argument center around his contention that the reason we have stopped *building* as a nation is a lack of desire. To that I say, nonsense. Based on my 20 years of experience working to build some of the physical infrastructure that he is calling for, his argument is both inaccurate and ironic.
Plenty of people in our country want to build great schools, new power plants, and clean infrastructure and are eager to solve climate change and re-establish the leadership of U.S. manufacturing. Many millions would work hard to *build* these things. It’s not that Americans lack the desire to build, we’ve simply been told that it’s not worthwhile or that it’s not as valuable as making social media apps and online ad platforms. Everyone who can make these revered and financially lucrative things is making them, while everyone who can’t is stuck on the other side of a yawning divide being told “pull yourselves up by your bootstraps, embrace the market, and learn how to code.”
Our political policies, capital markets, and brightest minds are increasingly drawn to building only a narrow slice of the societal infrastructure necessary for us to thrive. Isn’t it ironic that a call to build largely physical and industrial infrastructure lives on the website of a venture capital firm with the tagline “Software Is Eating The World”?
Don’t get me wrong. My essay is not meant to be a bitter tirade against capitalism or Silicon Valley and the tech industry. I will always be a great champion of market solutions and remain an absolute zealot for the use of technology in all parts of business, education, and society. I went to business school in the Bay Area and lived there for decades, surrounded by and immersed in the tech bubble itself. It’s a wonderful place full of endless optimism, opportunity, and imagination. It gave me the knowledge and opportunity to become who I am today.
However, I observed an overarching problem during my stint in Silicon Valley: Many of its inhabitants seem unable to grasp that different parts of our country and sectors of our economy work in a fundamentally different manner than the Bay Area and the tech sector. This disconnect is an important part of why we no longer *build.*
Infrastructure Is Persistent
Unlike many of the technologies made ubiquitous by Silicon Valley, infrastructure does not rapidly scale. Infrastructure exhibits tremendous inertia due to its long-lived nature, and the time and difficulty it takes to replace it. Infrastructure isn’t transitory, it’s persistent.
The companies with technologies that can scale cheaply and rapidly through direct-to-consumer markets like software and electronics thrive. They become powerful because they can build, deploy, and scale their businesses without much in the way of regulatory or societal support or consensus. They appeal to one of the few remnants of functioning democracy left in our country — the democracy of unregulated consumer markets, the one that votes yearly on iOS versus Android in the context of the new phones they buy, or the apps du jour that add convenience or fun to their connected lives.
With software, a new version comes out every few months or a patch is issued every few weeks to remove errors in the code or immediately address customer feedback. Feedback is instantaneous and can be gathered passively through user interaction data without soliciting the direct participation of customers. Even consumer electronics hardware is increasingly disposable and encourages the upgrade of obsolete equipment within a two-year product cycle.
Compare that market scenario with something like large-scale renewable energy, where technologies and companies sell extremely capital intensive, long-lived products into highly intermediated and regulated wholesale markets. There is no direct appeal to consumer preference. Nothing happens without broad societal consensus, which is required to change preferences and make purchasing decisions. Change is glacial.
An even sharper contrast can be seen with an infrastructure product like coal power plants, which first became operational in the U.S. in 1882. We now know that coal power plants had a very large “bug” on day one. They slowly killed those who lived near them and even more slowly contributed to our global climate catastrophe. Coal power companies didn’t, couldn’t, or wouldn’t patch this vulnerability quickly because infrastructure, unlike consumer technology, exhibits a very high degree of inertia, its persistence is intertwined with differences in financing, and regulation and consumer preferences must be filtered and translated through slow-moving political and financial systems to create any kind of market change.
Software is not nearly as capital intensive as power plants and has far more distributed purchasing decisions. The hundreds of millions of dollars required for a new power plant (or plants) to replace just one flawed coal generator takes years to raise. It requires the consensus of many investors in the capital markets that they trust the technology and are confident they will see a return on investment over the 30-50 year life of the asset. Most new purchasers of Windows, Office or a new phone app have a very short-term payback period for their investment, and the purchase price itself can be financed directly without need for outside capital.
On the other hand, U.S. electricity markets are highly regulated, and suppliers are intermediated by utilities and rarely have direct interface with end-customers. Direct-to-consumer renewable energy options are limited to things like rooftop solar or battery storage at residential or commercial locations. This is only allowed because it is relatively small and does not rely on the utility transmission or distribution system.
As a developer and provider of large-scale solar power plants, I only have a handful of potential customers. These large, quasi-governmental entities speak for my millions of potential end-users and have enormous bargaining power. We have no sales force at my company. We competitively bid some small mark-up to our cost and are evaluated against our competitors under rules that cost us millions in legal fees to negotiate with regulators and hope that we are awarded a contract for our product.
There are no real incentives to create new features and few opportunities to innovate my business model. Consumer preferences filtered through regulated procurements slowly reflect societal changes or preferences at best and do not respond to sales or advertising. The potential for viral growth does not exist. Scalability discussions might as well be taking place on a different planet than those kinds of conversations in the tech industry.
Infrastructure Needs Societal Consensus
These types of markets are the ones in which schools, hospitals, power plants, transmission lines, and factories are built. These infrastructure products require societal consensus among regulators, elected officials, and capital markets that aren’t often required for scalability and growth in consumer-facing technology markets. It’s on this point that Andreessen falls short, and this is the critical context that I think those reading his essay should all have.
As a builder of the very infrastructure that Andreessen calls for, just once I want to talk directly to my customer, to convince them of my product’s reliability, the critical need for this cleaner alternative, and the competitive value proposition. Just once I’d like to find capital providers that can achieve the scale and cost of capital required for these assets but have the imagination and forward-thinking vision of a Marc Andreesen.
What I want to do is build. I love taking empty GIS maps and turning them into some of the largest solar plants in the country. I love steel in the ground. And I am not alone. Hundreds of thousands of other women and men in the clean infrastructure business also want to build. Union laborers want to build gigawatts of solar and wind plants. Steelworkers want to build clean steel foundries that use competitively priced green hydrogen and renewable electricity to make the most cost-competitive steel in the world. U.S. manufacturers want to produce clean infrastructure equipment for global markets.
Instead, we’re stuck.
Not out of complacency or a lack of desire. We’re stuck without a functional government that can compromise and define a common vision. We’re stuck focusing only on those things that scale rapidly without the challenges of political or social consensus. We’re stuck because politicians and capital markets only seem capable of dreaming big dreams about faster-to-market, quicker-returning, consumer technologies.
We’re stuck because no one dreams in steel anymore.
A Tech Sized Vision For Clean Infrastructure
What we need is a new integrated vision for what we build, one that embraces the scale of change and consensus required to break free of this inertia. We need that vision to be proactive and not reactive, to look to the future and not fight over the past. Instead of putting tariffs on commoditized product markets like solar modules, which China already dominates, we need to get out in front of owning the hydrogen electrolyzer market and other emerging decarbonization technologies.
We need to use those cheap PV modules to make the U.S. the world’s largest producer of clean energy for the next 50 years. With that inexpensive clean power, we can become the cleanest and most competitive producer of energy and (formerly) carbon-intensive products. I’m talking about products we’ve made for generations — steel, cement, and glass — along with new zero-carbon products such as green hydrogen. The long-term labor intensity and job-creation potential of these industries is far greater than that of solar modules and other commoditized components.
We need a federal clean energy standard and a mandate on electric vehicles to bring our transportation and electric sector emissions to near-zero levels by 2050 and kick off the largest infrastructure buildout in human history. We need to speed the permitting and deployment of new clean infrastructure. We need to scale technologies that are “good enough” and favor rapid deployment, not wait for people to invent solar paint or invite next-generation nuclear plants into their backyards. We need a price on carbon with a dividend paid back to all Americans.
These are not complicated policies. They are straightforward and simple but will utterly transform the American infrastructure market and revitalize American industry, all the while preserving our planet for future generations. What is needed is the political will and societal consensus to make these changes. It’s time to dream in steel again.