More than 2.5 kilometres of previously unrecorded pre-Hispanic road, including a zigzagging main road of at least 1.2 kilometres, has been picked out by a drone-mounted laser scanner in the cloud forest across the river from Machu Picchu, on the side of the valley where the Inca had seemed not to concentrate their urban planning

Machu Picchu has never been a single ruin. Within the national park around it sit over 60 archaeological sites, Ziółkowski says, linked by a network of roads and serving various satellite functions in relation to the main complex, and they sit mainly on one side of the Urubamba.

That distribution has carried an inference with it. “They are situated mainly on the left bank of the river, and it seemed that this was where the Incas had concentrated their urban planning efforts,” says Mariusz Ziółkowski, director of the Centre for Andean Studies at the University of Warsaw.

A drone has now put more than 2.5 kilometres of pre-Hispanic road on the other bank.

The Dirección Desconcentrada de Cultura de Cusco, the Cusco arm of Peru’s Ministry of Culture, announced the result in a statement dated 10 August 2026. A laser scanner flown over the forest around the Mandor wall detected the roads along with what the ministry calls possible platforms and terraces, and one main road stands out among them, at least 1.2 kilometres long, laid out in a zigzag. The ministry lists it among the evidence making up the 2.5 kilometres rather than as an addition to them.

The ministry reports no excavation and no date, and says the chronology and function still require specific archaeological investigation. The Warsaw summary line says otherwise, and that divergence is set out below.

The University of Warsaw, whose Centre for Andean Studies ran the fieldwork alongside the Cusco directorate, placed the survey in late July and early August 2026 in its own announcement on 19 August. The Cusco statement dates the reconnaissance to July.

The bare-earth model under the canopy

The instrument fires rapid laser pulses at the ground and times how long each takes to come back. Many pulses hit leaves. Enough slip through gaps in the canopy to reach the soil that the ground returns can be separated out afterwards, leaving a model of the bare earth with the forest stripped away.

That is the whole point in the ceja de selva, the steep cloud-forest fringe of the eastern Andes where Machu Picchu sits. The Cusco directorate describes the method as a way to detect and record possible archaeological evidence hidden beneath the park’s dense vegetation and broken terrain.

The scanner here was carried on a drone rather than a crewed survey aircraft, and Bartłomiej Ćmielewski of Wrocław University of Science and Technology supervised the scanning. A drone flies low and slow over a small block of ground, which buys density of coverage at the cost of breadth.

The team was small and Warsaw paid for it. Jan Kłaput of the Warsaw centre led the survey and documentation, Paweł Dąbek came from the Wrocław University of Environmental and Life Sciences, and the Peruvian archaeologist Jeff Rocky Contreras Soto, of Apu Space, worked alongside them. Nino del Solar Velarde, who heads the Machu Picchu archaeological research coordination office at the Cusco directorate, supervised the fieldwork.

Roads on the far bank

The roads appear to lead to what the Warsaw announcement calls platforms and terraces, and the Cusco directorate calls possible platforms and terraces, near the Mandor wall on the right-hand side of the Urubamba valley, directly opposite the site itself.

“The roads we have discovered alter this picture – it turns out that the other side of the river also underwent significant cultural transformations during the Inca period,” Ziółkowski said.

The correction is narrower than it first sounds, and worth stating precisely. Neither statement says the right bank was thought to be empty. What Ziółkowski says is narrower: it seemed to be the side where urban planning was not concentrated, and 2.5 kilometres of road running toward possible terraces is not what an unconcentrated effort usually leaves behind.

Which is where the qualifiers start to matter, because the ministry hedges the two things it cannot yet stand behind: the platforms and terraces are only possible, and the chronology is open.

The chronology is still open

The main road’s characteristics are compatible with Inca construction techniques, the Cusco statement says, but specific archaeological investigations will be necessary to determine its chronology and function.

That gap is real, not procedural. The zigzag is the trait the ministry offers as an example of that compatibility, and a switchback is also the obvious way for anyone to take a steep slope on foot in any century.

The label the ministry actually uses for the roads is pre-Hispanic, which is wider than Inca. The platforms and terraces keep their hedge into the ministry’s own standfirst, which calls them possible platforms and terraces to be investigated. And the directorate has ordered focused archaeological research, including field survey and recording, to establish the characteristics, chronology, function and state of conservation of what the scan identified.

The two announcements also diverge, inside the sources rather than downstream of them. The University of Warsaw’s summary line describes roads “dating from the Inca period” found during “archaeological excavations”, while the body of that same announcement describes LiDAR remote sensing and the Peruvian statement leaves the dating open. The ministry’s wording is the one to hold to, since the body of the Warsaw announcement describes remote sensing and contradicts its own summary line.

Neither statement gives coordinates for the finds, either. Both teams are exercising caution about disclosing the exact location of the finds, Ziółkowski said, out of fear, in his words, of “the unbridled and, unfortunately, often harmful curiosity of various amateur treasure hunters.” A bare-earth model is a map of undefended things.

Next year’s trench

The first goal of the next campaign is to finish the remote-sensing survey, Ziółkowski said. On that basis the team plans fieldwork involving at least the partial excavation of the roads and the structures associated with them, which is the step that can produce datable material.

The institutional groundwork is in place. Peru’s Minister of Culture, Alberto Beingolea Delgado, appointed at the end of July 2026, announced ministry support for further research during an inspection visit to Machu Picchu on 11 August, and discussions about a framework for collaboration over the next few years are under way between the Warsaw centre and the Cusco directorate.

The collaboration itself is not new. It began in 2008, between the Centre for Andean Studies, the Regional Directorate of Culture in Cusco and the Machu Picchu National Archaeological Sanctuary, on the UNESCO World Heritage list since 1983.

What the directorate has ordered is prospection and field recording, aimed at four specific things: characteristics, chronology, function, condition. Three of those four, chronology, function and condition, are the ones the drone could not supply, and they are the ones that decide whether the picture of Machu Picchu’s territory really changes or merely gains a footnote.