Punch Cards and Power Plays: From IBM to Big Tech, or How We Learned to Stop Worrying and Love the Data State

There is a distinct class of modern astonishment that deserves to be bottled and sold as a novelty item. It tends to appear whenever wealthy executives, draped in the language of innovation and disruption, are seen sidling up to political power in ways that feel, how shall we put it, slightly illiberal. “Surely not,” comes the cry. “These are the people who gave us social media filters and same-day delivery. Why would they have any interest in power, control, or influence?”

History, as ever, sits in the corner nursing a drink, waiting patiently for its cue to say: “Oh, you sweet summer child.”

Let us begin, not with Silicon Valley, but with something far less glamorous: punch cards. In the early twentieth century, before “the cloud” meant anything other than impending drizzle, data processing was a mechanical affair. Information was encoded into stiff paper cards with holes punched into them, fed into tabulating machines that could sort, count, and analyse at speeds that would have seemed miraculous at the time. It was, in short, the cutting edge of information technology.

These machines were sold, leased, and maintained by a certain American company that would later become synonymous with corporate computing. Through its German subsidiary, it provided the infrastructure for large-scale data processing in a country that, in the 1930s, was becoming increasingly interested in categorising its population in rather sinister ways.

Now, it is worth being precise here, because this is where discussions tend to veer off into either denial or melodrama. The company did not invent the ideology of the regime it served. It did not write its laws, nor did it draft its racial policies. What it did do, however, was provide tools that made the administrative side of those policies vastly more efficient. Census data could be processed faster. Populations could be sorted and classified with greater ease. Bureaucratic machinery, already formidable, was given a technological exoskeleton.

If this sounds abstract, that is because bureaucracy always does. Evil, in its modern form, rarely announces itself with theatrical flair. It arrives instead as paperwork, logistics, and systems optimisation. The horror lies not in the machine itself, but in what the machine enables.

Of course, the question that inevitably follows is whether the executives at the top of this American firm knew precisely how their technology was being used. This is where historians begin to argue in earnest, and where certainty becomes an endangered species. Some suggest that the company maintained a degree of oversight and therefore bears a share of responsibility for what its machines facilitated. Others argue that once war broke out, communication became patchy, control attenuated, and the German subsidiary effectively operated within the constraints of the regime it found itself in.

What is not in dispute is that the technology was used, and that it mattered. The regime in question was obsessed with classification, with knowing who was who and where they were. Information was not merely administrative; it was instrumental. The ability to process data at scale did not create the regime’s goals, but it did help it pursue them more effectively.

Now, before anyone reaches for the nearest fainting couch, it is important to resist the temptation to flatten history into a moral fable with clear villains and heroes. The uncomfortable truth is that many companies operating in that period, both domestic and foreign, adapted to the political environment in which they found themselves. Some did so enthusiastically, others pragmatically, and a few reluctantly. The spectrum of behaviour is wide, and the motivations are rarely pure.

Profit, as it turns out, is a remarkably persuasive argument. Stability, even of the authoritarian variety, can be attractive to those whose primary concern is the smooth running of business operations. The suppression of labour unrest, the promise of large state contracts, and the general tidiness of a system where dissent is not so much debated as eliminated can all be framed, if one squints hard enough, as “favourable conditions”.

And so we arrive, with a slightly uneasy sense of déjà vu, at the present day. The machines have changed, the cards have vanished, and the hum of tabulators has been replaced by the silent churn of data centres the size of small towns. Yet the underlying dynamic, the relationship between private control of information systems and public power, remains stubbornly familiar.

Today’s technological landscape is dominated by companies that do not merely process data but inhale it. Social interactions, purchasing habits, location histories, personal preferences, professional networks, these are all collected, analysed, and monetised on a scale that would have made earlier generations of industrialists weep with envy. Data, we are told, is the new oil, which is a charming metaphor until one remembers that oil has a long and rather unsavoury history of entanglement with geopolitics.

The individuals at the helm of these enterprises are often presented as visionaries, and in some respects they are. They have overseen the creation of systems that connect billions of people, that enable instantaneous communication across continents, and that have fundamentally reshaped the way information flows through society. They have also, somewhat less poetically, built infrastructures that are extraordinarily useful to anyone interested in understanding, influencing, or controlling populations.

Governments, being neither foolish nor immune to temptation, have noticed.

The modern state, even in its most benign democratic form, has an enduring interest in information. It wishes to know who its citizens are, what they are doing, and, occasionally, what they might do next. In less benign forms, this interest can become rather more intrusive. The crucial point is that much of the relevant data is no longer held by the state itself, but by private companies.

This creates a relationship that is, to put it mildly, complicated. On the one hand, governments rely on these companies for infrastructure. Cloud services host public sector data. Social media platforms act as de facto public squares. Analytical tools developed by private firms are used in everything from policing to public health. On the other hand, these same governments have the power to regulate, tax, and, if sufficiently irritated, dismantle the very companies they depend upon.

It is a delicate dance, and like all such dances, it involves a certain amount of mutual flattery. Executives speak warmly of collaboration, innovation, and shared goals. Politicians praise the dynamism of the private sector while quietly drafting legislation that may, or may not, clip its wings. Behind the scenes, there are meetings, negotiations, and the occasional raised eyebrow.

Into this already intricate arrangement, one must add the variable of political alignment. In recent years, it has become increasingly apparent that some figures in the technology sector are not content to remain apolitical. They express opinions, fund campaigns, and, in some cases, cultivate relationships with particular political movements or leaders.

This is not, in itself, shocking. Wealth and influence have always had a tendency to gravitate towards power. What is perhaps more interesting is how this alignment intersects with the control of data.

Consider, for a moment, the incentives at play. A government that is friendly towards a particular company may be less inclined to regulate it aggressively. It may be more willing to award contracts, to overlook certain practices, or to shape policy in ways that are beneficial. In return, the company may provide services, expertise, or simply a lack of resistance.
None of this requires a grand conspiracy. It is simply the logic of mutual advantage.
The more disquieting question is what happens when the boundaries of what is acceptable begin to shift. If a political environment becomes more tolerant of surveillance, more relaxed about data sharing, or less concerned with civil liberties, the tools that have been built for entirely legitimate purposes can acquire a rather different flavour.

This is where the historical parallel, cautiously handled, becomes instructive. The lesson of the earlier period is not that companies inevitably support authoritarian regimes, nor that technology is inherently malign. It is that systems designed for efficiency and profit can be repurposed in ways that their creators did not fully anticipate, and that companies, faced with new realities, often adapt rather than resist.

Adaptation, after all, is what successful businesses do. They respond to market conditions, regulatory environments, and political climates. They pivot, they recalibrate, they find ways to continue operating. There is, in this, a kind of moral neutrality that can be deeply unsettling. The same flexibility that allows a company to thrive in a competitive market can also allow it to function within a system that is, to put it politely, less than ideal.

One might object that today’s world is fundamentally different, and in many ways it is. Democratic institutions, however imperfect, provide checks and balances that did not exist in more overtly authoritarian systems. The media landscape is more pluralistic, public scrutiny more immediate, and the capacity for dissent, though occasionally strained, remains intact.

Yet it would be naïve to assume that these safeguards are immutable. They depend on norms, on laws, and on the willingness of both institutions and individuals to uphold them. Technology, for all its marvels, does not inherently strengthen these safeguards. It can just as easily erode them if used carelessly or cynically.

There is also the small matter of scale. The punch-card systems of the early twentieth century were impressive for their time, but they pale in comparison to the capabilities of modern data analytics. Today’s systems can track behaviour in real time, identify patterns across vast datasets, and make predictions with a degree of accuracy that borders on the uncanny. The potential for both benefit and abuse has grown accordingly.

And so we find ourselves in a situation that is, if not identical to the past, at least rhyming with it in a slightly disconcerting way. Private entities control the infrastructure through which information flows. States have a vested interest in accessing and utilising that information. The incentives for cooperation are significant, and the consequences of that cooperation are not always fully understood at the outset.

The temptation, at this point, is to assign motives with a broad brush. To declare that today’s tech elites are consciously plotting some grand alignment with political power in pursuit of control. This makes for satisfying rhetoric, but it risks obscuring the more banal and therefore more dangerous reality.

It is not necessary for anyone to desire an authoritarian outcome for one to emerge. All that is required is a series of decisions, each individually justifiable, that collectively move in that direction. A data-sharing agreement here, a regulatory concession there, a quiet acceptance of practices that would once have been controversial. Over time, the boundaries shift, and what was once unthinkable becomes merely unfortunate.

If there is a dark humour to be found in all of this, it lies in the consistency of human behaviour. We flatter ourselves that we are more enlightened than our predecessors, that we have learned from history and would not repeat its mistakes. And yet, when presented with the same underlying incentives, profit, power, and stability, we often respond in remarkably similar ways.

The machines may have changed, but the logic has not.

None of this is to suggest that we are on the brink of some inevitable descent into dystopia. History is not a script, and parallels are not destinies. It is, however, a reminder that the relationship between technology, business, and power is fraught with complexities that cannot be wished away with optimistic slogans.

If there is a lesson to be drawn, it is a modest one. Systems matter. Incentives matter. And the quiet, unglamorous world of data processing, whether conducted with punch cards or algorithms, has a habit of shaping outcomes in ways that are only fully appreciated in retrospect.

One might even say that the real danger is not that anyone is deliberately building a machinery of control, but that, in the pursuit of efficiency and profit, we are all contributing to its construction piece by piece, congratulating ourselves on our ingenuity while ignoring the blueprint taking shape beneath our feet.

History, still nursing its drink, would probably raise an eyebrow at that.
And then, with impeccable timing, it would order another round.