Cloudflare's chief financial officer has projected that machine-generated internet traffic will exceed human activity by a factor of 1,000 within five years, rendering people "a rounding error on the internet." Thomas Seifert made the forecast during the company's Q2 earnings call on Thursday, marking a dramatic acceleration from Cloudflare's earlier predictions. The internet infrastructure firm had previously estimated that machine traffic would overtake human use in 2027, but measurements showed machines actually crossed that threshold in May 2026.
Cloudflare reported $696 million in quarterly revenue, representing a 36 percent year-over-year increase, though the company's losses more than tripled to $205.7 million. The firm projects total annual revenue between $2.865 billion and $2.87 billion, with planned capital expenditure at 14 to 15 percent of that figure—approximately $430 million. Seifert acknowledged his track record on predictions, noting he has "called it wrong at every point along the way," but maintained that current trends point toward exponential growth in non-human traffic driven by artificial intelligence. Investors responded positively despite the losses, pushing shares up 16 percent in after-hours trading to a new high representing 68 percent growth year-to-date, buoyed by record sign-ups among large enterprises.
Seifert emphasized that human traffic won't decline in absolute terms—machine-generated packets are simply "just how fast we're seeing non-human traffic grow." CEO Matthew Prince distinguished Cloudflare's approach from hyperscaler competitors, stating the company focuses on "selling actually work getting done" rather than leasing commodity servers. Prince characterized rival cloud providers as having "super-low" GPU utilization rates because "what they're selling is just a box, and it's up to the customers in order to actually maximize the utilization."
Cloudflare executives argued the surge in machine traffic will require greater efficiency and heightened security measures, with AI applications fueling the demand. Prince contrasted Cloudflare's low capital spending—a fraction of the trillions competitors are pouring into AI infrastructure—with what he described as hyperscalers using their balance sheets to purchase identical servers for AI companies. The company intends to maintain its serverless offerings and extract maximum value from existing hardware through improved scheduling and efficiency, rather than expanding into commodity compute rental. Prince declared that approach unattractive because "over time, it's a commodity business." The firm's strategic bet hinges on squeezing performance from every dollar spent while competitors race to build massive server farms, positioning Cloudflare as the efficiency layer atop an internet increasingly dominated by automated systems. The divergence between Cloudflare's capital-light model and the infrastructure arms race suggests competing visions for who captures value as machines become the internet's primary users, though both paths assume AI's traffic footprint will reshape the economics of connectivity itself.