My Grandfather Worked in a Steel Mill. Now, the Site Will Host a Data Center.
A poignant anecdote raises a range of complex questions about the future of the economy and inspires me to launch this newsletter.
After serving in World War II, my grandfather didn’t want to work in the northeastern Pennsylvania coal mines like his father, a Polish immigrant. He supported his young family by driving a cab and, briefly, a milk truck.
Then, a life-changing opportunity came along.
In 1953, my grandfather took a union job at the new U.S. Steel Fairless Works mill near Morrisville in southeastern Pennsylvania. He started in the warehouses and soon got a position operating a crane that moved finished rolls of steel onto trucks bound for Detroit and other domestic manufacturing hubs.
That job, my father says, is what made his life – and by extension, my life – possible. It provided a steady income, benefits, and a pension. And most importantly, it took my grandparents out of Dickson City, the declining coal town where they grew up, and allowed them to raise their children in a growing community.
Stories like my grandfather’s loom large in the imagination of many leaders today. With just a high school diploma, my grandfather went from living in poverty in the Great Depression to having a long career helping to make real, tangible things.
Trying to reclaim the economic moment he lived through has been a driver of policy for Democrats and Republicans (though their preferred strategies and tactics certainly differ). I understand the appeal.
Rebuilding the country’s industrial base can advance competitiveness, support national security, and bring benefits for workers too. While at the White House and now as a strategic advisor to regional innovation coalitions, I have focused on building pathways into jobs manufacturing everything from medicines to drones. Many roles in advanced manufacturing require a two-year degree or less and continue to deliver a wage premium compared to other industries.
So, I found it poignant and symbolic when on June 9, Governor Shapiro announced that the former site of the Fairless Works mill will host (drumroll, please)...an Amazon data center.
As part of the tech giant’s $20 billion investment in Pennsylvania, it will become one of two initial “innovation hubs” to power the advancement of AI. The site, located on the Delaware River, has plenty of a resource important to both steel mills and data centers: water.
Right after I heard the news, I tried to write a hot take on what this development means for workers, particularly those like my grandfather who never went to college. But none of my initial arguments felt quite right to me. That’s probably because neither this story nor this current moment in the economy lends itself to simple takes (if any do).
Instead, what it puts in sharp relief is the complexity and uncertainty facing leaders across sectors who are seeking to drive innovation, spur broad-based economic growth, and create good jobs.
First of all, state and local leaders can’t make policy and cut deals based on nostalgia. They might want to bring about a manufacturing renaissance, and many are helping to do that (including in Pennsylvania). But they also have to stay grounded in the reality of our economy as it is today.
Given that, leaders like Governor Shapiro understandably want their state on the forefront of the AI revolution and benefiting from the many billions being spent to develop it. But what benefit these investments will bring for the state’s workers and communities is far less certain.
The announcement promises that the deal will bring “1,250 high-paying tech jobs” and “thousands” more building and maintaining the data centers. That is certainly better than a vacant site, but it’s likely to be well below the more than 8,000 workers the Fairless Works mill supported at its peak in 1974.
The projections accompanying announcements like this are often overly optimistic, and Amazon’s specific track record is mixed, to say the least. The company’s union busting tactics are well known.
Whether the deal will directly create jobs is one important issue, but its indirect impact on jobs is a much bigger question. Direct jobs are those directly related to the project (for example, HVAC technicians who maintain the complex climate control systems in data centers). Indirect jobs are those created through the supply chain or increased economic activity driven by the project (for example, cooks at restaurants where the HVAC technicians spend their wages).
Typically, one would assume that both the direct and indirect job creation numbers will be positive. But how valid are those assumptions if the investment is advancing a technology that many believe could wipe out wide swaths of human labor? It might add 1,250 high-paid tech workers, but some experts think AI may already be wiping out millions of entry-level jobs.
Recently, a prominent AI researcher told me that he’s betting on 2040 as the year when all of our jobs become automated. (After that date, he stated that we’d all be living off universal basic income – a response that reflects how bizarrely sanguine many AI leaders are about the potential societal impacts of the technology.)
With these types of investments, state and local leaders will often seek to get the company to contribute to skilling up workers for the jobs created in the relevant industry and align public investments toward the same goal. The announcement states that the state is investing $10 million in workforce development efforts at schools, community colleges and union halls, and Amazon is committed to “partnering with Pennsylvania educational institutions and workforce organizations.”
I’m glad to see these commitments, but they’re also a drop in the bucket relative to the massive rethinking of our country’s education and workforce development systems that AI may require.
That’s a lot for state and local leaders to balance.
But there’s also the issue of dealing with the fact that data centers aren’t great neighbors (though to be fair, neither are steel mills). They bring noise and diesel fumes and require incredible amounts of electricity. Local communities have already raised concerns about the deal’s implications for their power supply.
I also haven’t touched on the ongoing disruption and uncertainty caused by the Trump Administration’s destructive funding cuts or the potential impacts of the massive bill moving through Congress.
After reflecting on this story, I felt that it was as good a time as any to launch this newsletter. My intent is to use it as an outlet for exploring a wide range of issues and policies – and as in this first edition, offering up as many questions as answers. The reference in the title will be clear to many who have worked in and around government. It’s part of a header added to documents to signal that they’re not final (and, therefore, should be protected from Freedom of Information Act (FOIA) requests).
I’ll end this newsletter where I started it – with my grandfather. According to my father (who also worked in the steel mill when he was younger), my grandfather worked hard, almost never passing up an opportunity to do a double shift, and appreciated the economic value of the job that he had. But he didn’t love it. He also imagined other potential futures for himself, including studying to repair radios and other appliances. My grandmother, a gregarious woman who chain-smoked and wore sunglasses at night, suggested that she could manage the front of the house for a repair shop. It probably would have succeeded.
When it comes to the economy and to AI and our society at large, the future isn’t set. What often seems inevitable is a choice that we’re collectively making. Steel mills might become data centers, but they don’t have to. Jobs can disappear, but they don’t have to.
The communities and the country that we live in are still in draft. We’re still deliberating. We can still decide on a different future.