The Earth’s Surface Will Soon Be Searchable

Satellites
September 17, 2019
Author
Julian Caicedo
Jessica Holland
September 17, 2019
Authors
Julian Caicedo
Jessica Holland

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

The Earth’s Surface Will Soon Be Searchable
Click to Read More

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

Imagine a search engine that allows you to type in questions about any physical object on the Earth’s surface—How many solar panels are there in California? How many cars were on the road across the U.S. last Saturday? How many houses are being built in China?—and provides you with instant answers. This type of service would provide valuable strategic and operational insights for the logistics, resource management, finance, and insurance industries. It could become indispensable to university departments, journalists and humanitarian organizations, and spark the development of new applications we can’t even imagine yet, in a similar way to how ubiquitous GPS led to Uber and Airbnb.

This vision is likely to become a reality in the near future. The company that accomplishes this feat has the potential to rival Google in scale and impact. The physical and technical infrastructure required for such a shift is already in place. Satellites, drones and other observation platforms are increasingly able to pick up high spatial and temporal images, while the proliferation of cloud-based servers and AI software has cut the costs of storing and processing those images.

What’s missing is a product that bridges the gap between raw images of the Earth and end-users who may not be trained data analysts but want to pull out useful facts and figures without poring over maps and counting objects individually. Such software must be “trained” to recognize all kinds of objects, from oil rigs to sheep: a painstaking process. This training works best if someone takes the time to identify many examples of that object, as well as examples of things that aren’t the object.

Until recently, this process has been outsourced or performed in house by satellite operators and pure-play analytics providers. The labor-intensity and technical expertise required for this process lead to vertical and imagery specific analyses that aren’t easily scalable or transferable. The fees charged can also reflect the time-consuming nature of this process, as well as a lack of transparency and standardized pricing.  

EO data could become as valuable as GPS

Space Angels has been searching for a startup that has the ability to offer Earth-observation (EO) analytics in a way that’s more efficient, scalable and fit for the coming EO revolution, and we found it in Picterra, the latest addition to our portfolio. The Swiss startup, which spun out of EPFL, intends to make the Earth’s surface searchable, enabling EO data to become as valuable and ubiquitous as GPS. To do so, it has created a scalable model for training AI models to recognize objects of interest and a sustainable plan for gaining users and generating revenue as it evolves through several stages of growth.

Picterra allows users to easily train AI models to recognize features of Earth-observation imagery. Researchers, business analysts, students and anyone else can upload aerial or satellite images, pinpoint examples of the objects they want to detect and then run an automated search for these objects. (They can also use object detectors that have already been trained by others.) While users benefit from easy-to-use AI modeling, Picterra benefits by accumulating a database of trained models. 


Picterra well-positioned to dominate EO analysis 

This model gives Picterra several competitive advantages. The user interface and clear range of solutions offered help make sales and deployment frictionless, and the system enables a high level of customization. The fact that users are contributing data allows the company to cheaply and quickly address a wide variety of industries, and the size and diversity of this data set ensure that the AI models perform accurately and reliably in a range of contexts.

Above all, it’s a solution that’s scalable and that improves itself over time, without intervention. As user numbers grow and the database of trained models becomes richer, a flywheel is created and data network effects kick in. This is a virtuous circle that keeps turning once a critical point has been reached. More users mean more data, which means better algorithms, which leads to a better product, which leads to more users—and so the cycle continues.

The flywheel effect is what allows companies like Amazon and Facebook to break away from their competitors, and it has the potential to propel Picterra from the most user-friendly EO-insights platform on the market to the Google of the physical world. Once this library of geo-spatial information has been built by users, it can create immense value—just look at the $116b Google made in ad revenue in 2018; this is just one model for monetizing such a database.

The platform will also generate value for businesses who use it for strategic insight, and by opening up access to real-time information about global ecosystems, it will empower people to make better-informed decisions about how to protect the planet, and enable positive change.

ARE YOU LOOKING FOR MORE SPACE?

There is serious financial upside to space investing, and commercial space is inspiring a new generation of scientists, engineers and entrepreneurs. Your investment will be instrumental in making this happen. The fact is that space angels are investing in innovative ventures that will help launch humanity’s storyline off-planet.

Apply For Membership