São Paulo Pride's 36.8K count erases millions: the politics of quantification

On Sunday, June 7th, 2026, the 30th São Paulo LGBTQIAPN+ Pride Parade filled Avenida Paulista with three decades of accumulated struggle, visibility, and resistance. By Monday morning, the major news outlets had settled on a single number to tell the story: 36,800 people. It was a figure that landed like a stone in still water—not because it was wrong in any technical sense, but because it erased something vast.

For the past twenty years, São Paulo's Pride had drawn between two and four and a half million participants, making it the largest such gathering on Earth. The 36,800 count came from the Political Debate Monitor, a research partnership between Cebrap-USP and the nonprofit More in Common, using drone imagery and automated head-detection software to measure crowd density. The methodology was sound by its own standards. The problem was what it chose to see and what it left invisible.

The researchers had selected 27 aerial photographs taken at precisely 14:37—a single frozen moment in time. The number they produced was not a measure of total attendance but a snapshot of peak concentration at one specific hour. The technology itself carried built-in blindness: the researchers acknowledged their system achieved 72.9 percent precision and 69.5 percent accuracy in identifying individuals. That margin of error swallowed entire categories of people—those who arrived early and left before the peak, those who stayed late, those who occupied the sidewalks and adjacent streets like Augusta and Frei Caneca, which traditionally overflow with crowds. A parade is a flow across hours, not a photograph. No drone image could capture that.

Three days earlier, on June 4th, the same research team had counted the Marcha para Jesus, a Christian gathering, and arrived at 33,800 attendees. The organizers had expected two million. Here was the machinery of quantification operating identically on two events with entirely different logics, scales, and contexts—yet no major outlet questioned whether these numbers were actually comparable. The researchers themselves had noted they could only partially capture the Jesus March because it ended near Campo de Marte, where drone use was unsafe. They never even recorded the event's true peak. Still, the numbers circulated as equivalents, and a narrative began to form: the Parade was shrinking. The Parade had nearly matched a Christian march. The movement was losing force.

But the numbers told only part of a larger story. The 30th Parada had unfolded amid severe financial and political pressure. According to APOLGBT-SP, the organization's revenue had collapsed by sixty percent between 2025 and 2026 after major corporate sponsors withdrew. In 2025, twelve brands had backed the event. In 2026, only three companies participated officially. Fewer sponsors meant fewer floats, less infrastructure, and more vulnerability. Meanwhile, the São Paulo City Council had approved a municipal bill on May 20th that would ban children from the Parade and prohibit it from public streets entirely, forcing it into private venues. To remove the Parade from the streets was to remove it from public consciousness—a strategy of erasure dressed as regulation.

What the activist response to the 36,800 figure represented was something scholars call "statactivism"—a militant use of quantification that refuses to surrender the authority to count to powerful institutions. The Brazilian LGBTQIAPN+ movement had long operated this way, producing its own statistics to challenge official state numbers and demand public policy in the language the state itself required. The movement's insistence that millions had occupied Avenida Paulista was not a rejection of numbers but a claim on the right to produce them, to define what counted as real.

The São Paulo Parade had begun in 1997 with two thousand people. By 2013, it had grown to roughly five million—the result of three decades of political organizing, of bodies that had to occupy public space simply to exist. To say that 36,800 people were present on June 7th was a political choice about what constituted mobilization, what counted as movement, what counted as resistance. In an era of rising conservatism, advancing far-right movements globally, and an election year, that choice mattered. The people were there. The reckoning of how many, and by whom, was itself a battle over the future.