---
title: "These $2 tiles do not double your Wi-Fi. They redraw mmWave paths"
locale: "en"
url: "https://irz.fr/en/articles/flowform-mmwave-tiles-en"
markdown_url: "https://irz.fr/en/articles/flowform-mmwave-tiles-en.md"
category: "tech"
tags: ["mmWave", "metasurface", "networking", "additive manufacturing", "6G"]
published_at: "2026-08-21T13:55:00.000Z"
author: "Léa Perrin"
translation: "https://irz.fr/fr/articles/flowform-tuiles-mmwave-fr.md"
---

# These $2 tiles do not double your Wi-Fi. They redraw mmWave paths

FlowForm prints passive reflectors for roughly $2 each and uses them to create new mmWave paths through a room. The measured gain is real, but much narrower than “Wi-Fi twice as fast.”

A **$2 printed tile that doubles wireless speed** is excellent headline material and terrible shorthand for what FlowForm actually does. The number survives scrutiny; the implied promise does not.

The UC San Diego work was accepted at SIGCOMM 2026,[2](https://flowform.wqzhao.org/)[3](https://conferences.sigcomm.org/sigcomm/2026/accepted/)[4](https://dl.acm.org/doi/10.1145/3789240.3829102) and the researchers put each passive reflector at roughly two dollars. In five indoor test environments, average mmWave link rates nearly doubled.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications) What they measured was a planned millimeter-wave deployment inside real rooms, not a coupon for twice the speed from a home Wi-Fi router.

One tile also tells us very little about the system. FlowForm begins with a map of a particular environment, designs **many passive surfaces for that site**, and treats the reflected paths as a network rather than as isolated lucky bounces.

## Not Wi-Fi

FlowForm targets **millimeter-wave**, or mmWave, communications, where enormous radio bandwidth comes with an awkward physical weakness: put a wall or a person in front of the receiver and a useful path can disappear.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications) At these frequencies, adding more access points is one answer, while active relays and reconfigurable intelligent surfaces can provide alternative routes at the price of power, control links and extra coordination.

The UC San Diego team makes a different trade. Instead of continuously reprogramming a reflector after installation, it puts the intended radio behaviour into the surface geometry before the tile reaches the wall.

The complexity has not vanished; it has moved from runtime electronics into survey, geometry and placement. For the deployment to work, the room has to be understood before the tiles are made.

## Six inches

Each FlowForm tile is roughly six inches square, about fifteen centimetres, and starts as a 3D-printed plastic part coated with conductive paint. Many thousands of structures smaller than the relevant wavelength cover its face, passively shaping the reflection of an incoming signal.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications)

> Illustration: Close-up of a FlowForm tile next to a US penny showing the scale of its surface pattern. A 15 cm tile contains many thousands of sub-wavelength elements. The pattern is 3D-printed in plastic and covered with conductive paint. Credit: [David Baillot / UC San Diego Jacobs School of Engineering](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications).

There is no transmitter behind the pattern and no battery to charge. The tile only redirects energy already travelling through the room. That makes **$2** a plausible parts-and-fabrication figure for the reflector itself, but not a budget for the rest of the job: somebody still has to survey the site, compute a layout, install the pieces and revisit them if the building changes.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications)

Site specificity is the price of making the hardware so simple. Researchers start from the geometry of the room and the radio locations, then select surfaces and placements for the dead or fragile paths they want to repair.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications)

## Two flows

FlowForm divides those paths into **major flows** and **minor flows**, names that make the architecture easier to picture.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications) Major flows are focused relay chains: they move the signal farther through the building or steer it around an obstacle, acting as the backbone. Minor flows branch away with broader beams and feed the areas where people actually need a usable link.

> FlowForm
> **A backbone, then branches**
> - Map the room and the areas that need coverage.: 1
> - Build focused major flows across distance and around obstacles.: 2
> - Branch wider minor flows into user areas.: 3
> - Let standard mmWave radios choose among the available paths.: 4
> The surfaces remain fixed. Diversity comes from the network of paths built into the room.

Fixed reflectors sound poorly suited to moving users, but FlowForm relies on behaviour mmWave radios already have. Access points scan steerable beams repeatedly, while the planned network gives a user location several reflective paths at different angles; as someone moves or turns, the radio can find whichever path currently performs best without receiving instructions from the tile network.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications)

The team therefore reports no required firmware change, no new standard protocol and no runtime control channel for the metasurfaces.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications) The intelligence was spent earlier, in the geometry and placement.

## What x2 means

The evaluation covered **five real indoor environments** with a mmWave radio testbed, rather than one carefully framed tabletop demonstration.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications) UC San Diego reports that FlowForm nearly doubled average link rates and more than doubled coverage area in challenging environments.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications)

Neither figure says that every download now takes half the time. The link-rate result is an average across the researchers’ test sites, with all the dependence on layout, user position and baseline that an average carries. Coverage area answers another question altogether: how much of the room can sustain a useful connection.

“Double wireless speeds” erases those distinctions, and it also undersells the more unusual result. FlowForm recovered useful capacity in places where a wall, corner or person had made the direct mmWave path poor or unavailable, using pieces of printed plastic rather than another powered radio.

The researchers also compare FlowForm with an idealised active RIS and report comparable performance at much lower cost and complexity.[1](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications) That benchmark is evidence about this experiment, rather than a claim that every passive FlowForm tile can replace every active RIS design.

## Fixed by design

FlowForm saves electronics by accepting immobility. An active RIS can alter its response electronically as conditions change; a fabricated FlowForm surface keeps the geometry it was given.

When a room stays close to the map used during design, the bargain is attractive because the surface has nothing to power, address, update or synchronise. The cost of that simplicity is paid when the room changes. A large change to the building exposes the opposite side: move a partition, reorganise user zones or substantially change the radio installation, and part of the old path design may become less useful.

Reconfiguration then means physical work. Tiles can be moved, replaced or remade, but that effort belongs to deployment cost and should not be confused with the two-dollar fabrication price.

## The wall joins

Wireless infrastructure is usually drawn as active boxes placed inside an environment, with walls shown mainly as losses between those boxes. FlowForm changes the drawing by making part of the environment itself participate in the radio system.

A pattern on a fifteen-centimetre square can therefore contain something close to a routing decision, decided during design and expressed physically before the first packet crosses the room. The surface does not know about packets; its geometry simply makes certain radio paths more useful than they were before.

We do not get a universal recipe from five indoor sites, and site-specific fabrication remains a practical constraint. But the experiment changes the menu of responses to a blocked radio path. We can add another powered box, or in some rooms we may be able to change the surfaces the existing waves already strike.

## References

1. [UC San Diego Today, “Inexpensive Reflective Tiles Pave the Way for Cost Effective Millimeter Wave Wireless Communications”, 18 August 2026](https://today.ucsd.edu/story/inexpensive-reflective-tiles-pave-the-way-for-cost-effective-millimeter-wave-wireless-communications)
2. [FlowForm, ACM SIGCOMM 2026 project page](https://flowform.wqzhao.org/)
3. [ACM SIGCOMM 2026, Accepted Papers](https://conferences.sigcomm.org/sigcomm/2026/accepted/)
4. [ACM Digital Library, FlowForm, DOI 10.1145/3789240.3829102](https://dl.acm.org/doi/10.1145/3789240.3829102)