Summary

  • In 1993, an NSF access point was reportedly ready, but participants said the expensive communications channel needed to reach it still required financing; BITNET and RUNCOL were earlier, different services.
  • Uniandes histories place functional access in late May or mark 4 June as the opening, while InterRed’s statutes are catalogued to July and CONPES 2739 followed in November.
  • Those records establish different stages—network operation, public recognition, an operator’s remit and national policy—not immediate nationwide access or a single source of authority.

By 1993, one of the people organizing the Colombian network later recalled, the National Science Foundation’s Latin American access point was already available. Yet Colombia did not have an Internet service: the costly communications channel still needed a financing decision, and universities could not pay for it alone. This is a participant’s retrospective account, not a circuit log, but it changes the question. The problem was no longer simply whether a route existed at the far end; someone had to fund a path that Colombian institutions could actually use.

A connection attempt around 31 May 1994 failed, while later accounts place a public opening on 4 June and a functional connection in late May. The record does not identify the first successful packet.

The link had a long prehistory. In 1990, the Red Universitaria Colombiana, or RUNCOL, connected academic institutions to BITNET. The University of the Andes research history describes the route through Columbia University in New York to an IBM computer in Bogotá, from which other Colombian institutions could participate. BITNET carried email, discussion lists and documents. It gave universities experience with remote communication, but it was not the same network as the TCP/IP Internet link that followed. Treating it as an early Internet connection would erase a useful technical and institutional distinction.

RUNCOL’s own records show what “participation” looked like before a national link existed. In 1991, ICFES and the University of Quindío signed an agreement for the university to join RUNCOL and use BITNET. The agreement describes membership in a defined academic network; it does not make the university a representative of every future Internet user. The 1994 transition required a different combination of technical work and institutions.

The archive’s research introduction says that, from mid-1993, the necessary relationships were assembled. It names the University of the Andes, EAFIT in Medellín and the University of Valle in Cali, with Colciencias, ICFES, Telecom and the Bank of the Republic also involved in different ways. That list is not a complete roster of who controlled each connection. It does show that “the country connected” was shorthand for coordinated work across organizations with different roles, resources and responsibilities.

The archive’s oral histories make the financing problem more concrete. Participants recalled that the NSF access point had been secured, while the communications channel to reach it was costly. The later arrangement involved public and academic support and a contracted channel, with IMPSAT remembered as the initial provider. The recollections do not provide the signed circuit contract or a complete contribution ledger, so they establish the bottleneck and remembered arrangement—not each institution’s exact share.

An archive record dated 25 July makes the funding arrangement more concrete. Uniandes rector Ernesto Guhl told InterRed’s assembly that the university intended to join as a titular member and set out its contribution in three installments. The letter records an institutional commitment and payment schedule, not a complete cost ledger; it does not show what share of the international circuit Uniandes ultimately paid.

One later account describes the path from Homestead, Florida, through IMPSAT’s operations on Cerro de Suba, then Torre Colpatria, and finally the University of the Andes. It is a retrospective institutional description, not a surviving carrier diagram or a measurement of where traffic could go. The detail makes the connection tangible, but it does not establish redundant routes, stable uptime, broad public access or reach to every Colombian campus. A path that delivered packets to a university node was an important infrastructure achievement; it was not yet a national consumer service.

The operating body also has its own chronology. The archive catalogues InterRed’s statutes as 1 July 1994. They describe a private, nonprofit corporation with mixed participation, and Article 4 gives it an expansive remit: operate networks and related services, direct the Science, Education and Technology Network (CETCOL), connect Colombian networks to the Internet and other international networks, and encourage access across the country. These are powers and purposes written into a charter. They do not establish that InterRed had performed every listed function before it was incorporated, or that access had become universal.

The charter did not instantly transfer service. In a 2023 interview, Hugo Sin recalled that Uniandes continued providing Internet service to universities and other organizations after InterRed was created; the corporation may have taken over the broader service role one or two years later. That later memory distinguishes legal purpose from operational handover, though it does not pin down an exact transfer date.

A letter dated 29 August shows that geography still shaped reach. The archive summarizes Hugo Sin’s message to the Regional Autonomous Corporation of Cauca: although Uniandes was the node managing Internet in Colombia, the distance meant the connection had to run through the University of Valle. It documents one regional arrangement, not a complete national topology or universal service.

The sequence matters. The oral account places a connection attempt around the founding assembly on 31 May; the institutional “opening” date is 4 June; the charter record is dated a month later. In November, Colombia’s approved CONPES 2739 set broader science-and-technology policy and discussed public-private collaboration and international research networks. That later policy did not cause the May connection retroactively. Nor does a corporate charter alone explain who paid for an individual circuit or who could authorize each operational change. Each document answers a different question.

That order is easy to lose when a country’s Internet history is compressed into a birthday. “The first connection” can mean an experimental TCP/IP exchange, a connection that worked long enough to be reported, a formal public opening, or the start of an organization’s service. The Colombian sources do not settle every boundary. They do let us see a transition: academic networking created practical experience; a coalition built and tested an international path; InterRed acquired a defined operating purpose; and government policy later recorded a national research-network agenda.

The distinction also narrows claims about representation. Researchers, universities, government agencies and a carrier all took part, but participation is evidence of expertise and contribution, not proof that those present could bind every institution or future user. InterRed’s statutes specify who could join, pay contributions, vote and serve in governing bodies. The existence of those rules is more concrete than calling the arrangement simply “multi-stakeholder,” though it still says nothing by itself about how broadly the membership spoke for the public.

Sources

Editorial lens: Heng Lu, Note 73, “The Multi-Stakeholder Mirage” (4 July 2026). It informs the distinction between participation and authorization; it is not historical evidence.