The 32‑bit Ceiling

In the early 1980s the architects of the Internet imagined a world with only a few thousand networked computers. A 32‑bit address space—2³², or 4,294,967,296 addresses—seemed more than enough. The reality that emerged was starkly different: by 2011 the last blocks of the global IPv4 pool had been allocated, and the Internet had to survive on a finite number of addresses.

How the Exhaustion Happened

The exhaustion timeline is a simple list of dates when each regional Internet registry ran out:

Region Registry Exhaustion Date
Asia‑Pacific APNIC Apr 2011
Europe RIPE NCC Sep 2012
Latin America LACNIC Jun 2014
North America ARIN Sep 2015
Africa AFRINIC Apr 2017

By 2019 every registry had hit its ceiling. The Internet, however, continued to grow.

NAT: The Unseen Savior

Network Address Translation (NAT) fundamentally altered the meaning of an IP address. Before NAT, every device required a unique public address. After NAT, a single public address on a router can represent thousands of internal devices. The router keeps a translation table that maps internal private addresses to the public address, allowing responses to be routed back correctly.

“NAT didn’t extend the life of IPv4; it redefined what an address means.” – Anonymous Network Engineer

While NAT introduced complications—peer‑to‑peer connections, application breakage, and the requirement that all connections be initiated from inside—it also made IPv4 viable for the modern era of inexpensive, ubiquitous devices.

The Market for IPv4 Addresses

With the supply exhausted, the remaining blocks became a commodity. In 2011 Microsoft paid $7.5 million for 666,624 addresses—about $11 each. By 2024 prices had climbed above $50 per address, and in 2025 large blocks traded around $24, while smaller blocks hovered between $31 and $35. IPv4 addresses evolved into real‑estate: finite, transferable, and appreciating.

IPv6: The Permanent Fix

IPv6, with its 128‑bit address space, offers 340 undecillion addresses—far beyond the needs of any foreseeable population. Yet IPv6 and IPv4 are not interoperable; a dual‑stack approach is required during migration, effectively doubling network complexity.

"IPv6 is the only long‑term solution, but the transition is a marathon, not a sprint." – IPv6 Adoption Study

Why the Designers Didn’t See It

The original designers of IPv4 focused on a research‑lab environment, not a world of cheap, mass‑produced devices. Their assumption that a 32‑bit space would suffice was reasonable at the time, but the exponential growth of networked objects—thermostats, cars, wearables—exposed the hard limit.

The Ongoing Transition

Today most networks run dual‑stack, supporting both IPv4 and IPv6. IPv4 continues on borrowed time, sustained by NAT and market forces, while IPv6 grows slowly but inexorably. The story of the Internet’s address exhaustion is a testament to engineering ingenuity and the power of market dynamics to overcome seemingly immutable limits.

Source: Connected App – "Why We Ran Out of IPv4 Addresses"