Sometimes I wonder how my parents managed to go for road trips—paper maps, random directions from strangers, and all the hassle. Things improved slightly in the late ’90s, when Yahoo launched one of the first online map services. You’d type in a start and end point, wait for the page to reload (patience was a virtue in the ’90s), and come up with a basic route plan. By the early 2000s, it was handling up to 35% of all map queries in the U.S.

The real turning point came when Google entered the scene in 2005-2007 investing millions of dollars into advanced mapping infrastructure and significantly upgrading the user experience (allowing smooth dragging/zooming) and streamlining it all: just type in your destination and you’re set with a route, traffic updates, and an ETA.

Yet, despite its sophistication and dominance with over 67% market share globally, Google Maps isn’t without flaws. Particularly in developing countries like India, we have to spend hours in traffic congestion due to the ongoing construction of metros, uneven roads or some major/minor accidents on the way. 

Infrastructure alone is not the problem here. There have been tragic accidents directly linked to faulty navigation—from a driver crashing off an abandoned bridge or a bus crash in Washington due to ignored height restrictions. Extrapolating from these anecdotes, it’s plausible that hundreds of accidents worldwide each year could be linked to navigation errors, with Google Maps implicated in a significant portion due to its dominance. 

The underlying problem here is why Google Maps has not been able to provide us with accurate or real time data about most roads today? 

The major reason is attributed to how it collects its data?

Exhibit: Google maps using vehicles for Street View data collection 

Exhibit: Apple Maps personnel collecting street view data in AP(India)

Google Maps relies on a combination of satellite imagery, street view vehicles, and user contributions to build and maintain its maps. Satellite imagery provides a bird's-eye view of the Earth's surface, sourced from various satellite constellations, including NASA's Landsat. Street View, on the other hand, utilizes a fleet of cars equipped with 360-degree cameras to capture images of roads worldwide as shown in the picture above and by trekkers for inaccessible areas. As Google Maps’ user base grew, every user’s movement contributed data. The app passively collected location and speed data from hundreds of millions of phones, enabling real-time traffic and precise travel times. This network effect allowed Google Maps to emerge as a leader globally. 

Network effect of Google Maps 

However, due to such capital intensive techniques of data collection the frequency of updating real time data is low leading to outdated representations of rapidly changing landscapes. This leads to inaccuracies, particularly in less populated or rapidly developing regions, posing challenges for users relying on the platform for navigation and planning.

Source: Hivemapper 

Left: Google Maps; Right: Hivemapper 

Part II: Why Crowdsourcing Mapping Alternatives Failed to Dethrone Google?

Over the years, various competitors – from big tech offerings to crowdsourced projects – have tried to solve these gaps and chip away at Google Maps’ dominance. However, none have made more than a minor dent in Google’s market share. Key reasons for these failures include data gaps, inferior user experience, fragmented ecosystems, and inability to achieve critical mass:

Exhibit: Analysis of Google Maps Alternatives

Exhibit: Market Share of Mapping Platforms (as of 2020) 

Source: Wall Street Journal 

Common Themes in Failure: Most of these companies couldn’t match Google’s full-stack approach – massive data collection, continuous updates, effective user experience and accurate data. Thus, despite a plethora of mapping solutions, Google Maps has remained essentially a monopoly in practice.

Part III: Can DePIN Decentralized Networks Challenge Google Maps?

A new wave of projects is emerging under the umbrella of DePIN – Decentralized Physical Infrastructure Network to fill the void in solutions provided by Google Maps and the other platforms we discussed in Part II. 

Exhibit: How blockchain solves three main problems in the current crowdsourcing platforms?

The core idea of DePIN is to allow anyone in the world to contribute data to achieve specific goals. Here all the users need to install a device in their cars and share the data of their surroundings to build accurate and real time data of the roads and nearby areas with minimal effort and cost as compared to the capital intensive techniques adopted by Google and Apple. Hundreds of millions of people drive everyday, especially drivers from ride hailing services, transport companies and e-commerce deliveries who spend 10-12 hours on the road everyday. They could enable us to build infrastructure 10-100x faster than previous solutions at a much minimal cost. 

While this could still be achieved with OSM or Mapillary, there was a lack of organized approach to reach scale and lack of incentivization to keep these drivers motivated to share the data everyday. With the integration of proper reward mechanisms, DePIN based mapping solutions can be a go-to-solution for solving this trilemma in the most effective way.

Big corporations like Facebook, Amazon have also shown their key interest in building crowdsourced decentralized maps. While the vision of platforms like OSM, Overture was more about users to get cheap access to mapping data and for humanitarian reasons, these companies had different objectives. Facebook owns a multitude of social media platforms like Meta, Instagram etc which requires location data for check-ins, events,  targeted advertising and addresses linked for business pages among others while Amazon requires it for their e-commerce deliveries. As the influence of these big corporations is increasing in these platforms with almost 17% of road edits being made by corporate accounts, most volunteers now fear it could threaten the map’s status as a free and open-source project. 

Exhibit: The rise of the corporate mapper on OpenStreetMap, 2014-2020.

Another reason why OSM and other crowdsourcing platforms couldn’t overpower Google was due to the quality of data collected. A smartphone has limited capabilities which lack the same level of hardware optimization and quality processing for mapping purposes. This is why most enterprises continue to pay millions of dollars to Google for its APIs like Uber, paying approximately $58Mn back in 2018. 

Web3 based DePIN platforms can solve problems in the earlier crowdsourcing platforms by enabling a much better infrastructure for reliable real time data, with proper incentives for stakeholder contributions and solving for monopolies in the hands of big corporations. 

When crowdsourcing data collection is involved, we need to consider 3 major factors to create a network effect - 

• Cost of the user contributing data 

• Efforts required by the contributor on a daily basis

• Potential rewards that can be earned for active contribution 
  

Generally, we’ve observed that DePIN networks scale more easily if the contributors pay a one-time cost (in time or money) up front, as opposed to an ongoing, continuous cost; passive networks with minimum effort are much more easy to set up, and therefore easier to scale

                                                                                                                  -  Multicoin Capital

Based on these metrics, I have analyzed these prominent DePIN mapping projects – Hivemapper and NATIX – and their approach towards data collection. 

Exhibit: Supply Side Analysis of current DePIN Solutions 

** All the information about rewards and user sentiments have been collected by directly connecting with users who have been active on the platform and analyzing their experience. The information about the amount of rewards you can earn is not fixed and may vary depending on a project’s policies and terms.

Most DePIN users are willing to pay an upfront cost on the hardware if they believe the ROI is good and cost can be easily covered up. It is important for a network to balance the rewards for early adopters while also ensuring that tokens are available to reward contributors in later stages of network maturity. 

In the case of Hivemapper, the cost is relatively higher but the rewards ratio seems good. In the earlier Dashcam version, all users in 3x multiplier areas like the UK were able to cover their initial costs in 2-3 months while for 1x zones it took relatively longer. Most users, especially full time drivers are getting attracted to it for an additional passive income source. Since the cost of the upgraded Bee version is higher, it is expected to have higher rewards allocation for data collected through these. While for Natix, the entry point is easier but relatively there are lesser rewards with some effort required on a daily basis to fix the smartphone. 

Analysis of Current Usage Metrics

Hivemapper 

~ Unique Contributors: 82,049 users (as of Jan’2025)

~ Data Collected in Kms: 268Mn (as of July’24)

~ Usage (Demand) in Tokens Burned: 11.3Mn (as of Jan’2025)

(75% of the tokens used to buy map data is burned) 

Natix Network (As of March’25)

~ Unique Contributors: 244K users

~ Data Collected in Kms: 153Mn 

~ Usage (Demand) in Tokens Burned: 190.7Mn

(40% of the tokens used to buy map data is burned/used for buyback) 

Both Hivemapper and Natix Network have been able to outperform Google Maps and other platforms in the pace of bootstrapping and building the supply side network as it took Google Maps 12 years to collect 16 million kms of data by employing heavy capital expenditures. 

Exhibit: Hivemapper collected data for 5Mn miles in just 8 months as compared to Google Map’s 5 Years 

The seamless onboarding to Natix network without the limitations of hardware requirements is reflected in its total unique contributors as compared to Hivemapper. 

Another factor to consider is the supply chain management when there is hardware involved. Usually it is ideal for a DePIN project to have control over manufacturing of hardware in the initial stages to ensure quality checks, an approach followed by Hivemapper. And later they can open it up to third party manufacturers and outsource once they are headed towards scaling. 

However, Hivemapper has been facing extensive supply chain issues in manufacturing and users have been waiting for over 1+ years now to receive their devices as indicated in their low number of contributors. Some of them have even applied for refunds due to the prolonged waiting time. But the team has been quite responsive on Discord and the users have received timely refunds. Overall the users are bullish about the project and some of them are only waiting for supply chain issues to get resolved to purchase the device and get started. 

Despite the issues, Hivemapper has managed to outperform in the data collected from across the world where most of the data collected comes from USA & Europe and now much efforts and partnerships are being built to expand in Asia. 

The subscription service offered at only $9 for fleet owners could also be a game-changer to create a network effect.Interestingly, many individuals who don’t drive much but have the means to invest are partnering with active drivers, paying for the device and having it installed in their cars. In return the drivers are paid about $60 or so a month, a mutual benefit for both. 

While both of these networks have actively managed to generate a network effect on the supply side, the real differentiator and success would be when they achieve threshold scale i.e. there is demand for their data and they’re able to take away some market share from the current platforms. 

 And this is only possible if the network provides good quality data which can compete with Google Maps. Though Hivemapper involves a relatively higher cost for a user to join in, its 360 degree cameras attached on the top of the car could give better and quality view of the roads that can compete with Google Maps as compared to Natix’s smartphone data. Moreover, they also added AI Map Trainer programs where the raw data collected by drivers worldwide is turned into a richly detailed map through human validation. The AI does the initial heavy work—spotting speed signs, traffic signals, and more—but it’s the trainers who verify and refine each object to ensure pinpoint accuracy.

Hivemapper has explicitly mentioned in their Q2’24 report about working with a Fortune 50 customer, two of the top 10 global online map providers, local municipalities and property management companies, and dozens of automotive companies. This implies data collected has reached threshold scale with around 11Mn Honey tokens burned as of Jan '25. 

Similarly Natix announced in a Nov19’ 2024 post on X, regarding securing its first data client and moving towards data monetization. No details have been disclosed regarding the buyers and have overall burned 190.7Mn Natix tokens till date. 

Decentralized Verification

In the current market sentiments, it is relatively easier for a DePIN startup to work on the supply side as people tend to be attracted to a project with handsome rewards mechanism, strategic marketing and prominent investors. However, the most difficult part would be verification. Another very important point to consider as the project will only materialize and be successful if there is also enough demand for its data. And the demand will only come if the supply side is consistent, accurate and reliable. 

Without robust verification, dishonest participants can “game the system”—collect rewards without doing real work or with the help of AI can also upload fake data. 

For instance: When Hivemapper's  AI trainer role first launched, spam bots flooded the system, hoping to snatch quick rewards with fake reviews. Fortunately, Hivemapper swiftly rolled out stricter reputation checks, blocking these bad actors and protecting the integrity of the data. Decentralized verification is thus the key to trust and quality control in a protocol.

Hivemapper employs multiple layers of verification including Proof of location via GNSS (Global navigation satellite system) or helium hotspots to record the position of the dashcam while mapping. Another major method involves vision based consensus where data is cross-checked from different vehicles riding in the same areas factoring in these metrics -  distinct hardware IDs, non-overlapping routes, and time-separated observations. Also data from 2 or more Bee devices fixed in 1 vehicle is de-duplicated and rewards are not issued multiple times which makes it difficult for a single player to inject false data. 

Similarly Natix employs a number of methods including random sampling where contributors are randomly sent challenge requests and enforces slashing in case they submit inaccurate data. Secondly, it has enabled Proof of Location in collaboration with the XYO network with which NATIX can validate that collected data originates from claimed locations. It also employs a Proof of Computation using zk-Proof which helps to prove that a device ran a particular ML model correctly. This is relevant for ensuring that the device’s output (e.g., “there are 3 cars in frame”) is true and is not guessed or faked. But the primary problem with this method is - “How do we ensure the real-world data that is uploaded is genuine in the first place? A ZK proof can confirm you correctly computed something on an input, but it cannot confirm the input (a sensor reading or camera feed) is honest or not tampered with, without additional trust assumptions or secure hardware.

Security of Data Collected

Both of these platforms extensively take measures to not invade the privacy of users and maintain anonymity of the identity of users sharing data with the network. The device automatically blurs faces, bodies, and license plates etc rather than uploading raw images. You can also manually pause data collection in places like your home or workplace giving you control over what gets mapped. 

Closing Thoughts

Having explored the ups and downs of mapping—both in centralized giants and emerging DePIN projects—I’m convinced that the decentralized route holds immense promise. The pace of data collection is already outstripping what we’ve seen from Google’s early years, and the global community of drivers and contributors stands to make these maps more accurate, dynamic, and up-to-date than ever before.

However, the real litmus test lies in whether these platforms can maintain quality standards at scale. On one hand, we have hardware supply chain issues, malicious players hoping to exploit reward systems, and on the other hand, smartphone-based data that might not always match professional-grade cameras. Ultimately, it’s a race to achieve Product Market Fit by balancing incentives, data quality, and real-world usability. I’m excited to see how these projects evolve in the coming years—and whether they can spark the real competition that this space has been missing for so long.