Space Capital：2022地理空间情报策略手册（英文版）（58页）.pdf

《Space Capital：2022地理空间情报策略手册（英文版）（58页）.pdf》由会员分享，可在线阅读，更多相关《Space Capital：2022地理空间情报策略手册（英文版）（58页）.pdf（58页珍藏版）》请在三个皮匠报告文库上搜索。

1、The GEOINT PlaybookReport by Space CapitalAn investment thesis on the future of geospatial intelligence and our outlook on digitizing the physical world.2The GEOINT Playbook|Report by Space Capital Executive SummaryThis is largely due to the proliferation of spatial analysis;a critical feature of mo

2、dern enterprise and consumer applications.Yet despite the growth in geospatial data creation and,subsequently,the tools to conduct geospatial analysis,working with this type of information remains complex and the number of products limited.This is in large part due to the broader industry-wide trend

3、 of market consolidation amongst existing incumbents and well-funded technology startups.3 This points to a conflicting theme within the geospatial stack:companies that begin as a modular technology stack(unbundled)then evolve into a vertically integrated software solution(bundled).The new“As-a-serv

4、ice”paradigm has unlocked scalable modular development across compute,storage,application programing interfaces(APIs),etc.and is accelerating the pace of innovation.In this playbook we will walk through our thesis,which analyzes geospatial intelligence(GEOINT)through the lens of technology layers:In

5、frastructure,Distribution,and Applications.This framework helps us connect the dots from the origin and constraints of the geospatial stack to evolution and inflection within the market.It also helps us see the larger role that space-based technology plays in an ecosystem that intersects with the mo

6、dern tech industry and serves customers across a wide variety of markets.Key TakeawaysThe global geospatial1 market is expected to grow from$63.1 billion to$147.6 billion in the next five years.2 Application:A seemingly infinite number of venture-scale businesses are now being built in multi-trillio

7、n dollar global industries like Agriculture,Insurance,Climate Markets,and Augmented Reality,with the market potential looking similar to Location-based Services when the iPhone first embedded GPS.Infrastructure:A variety of geospatial sensor platforms now capture data at different altitudes,benefiti

8、ng from low-cost components,commoditized storage/compute,and decades of geographic information systems(GIS)product development.Distribution:Adoption of cloud,AI/ML capabilities,and increasingly powerful APIs and software development kits(SDKs)are expanding the user base and applications beyond data

9、engineers,GIS specialists,and subject matter experts.3The GEOINT Playbook|Report by Space Capital Executive SummaryNASA began to develop observation technology to better understand our planet at scale(remote sensing)in the 1960s.Landsat 1 was launched in 1972 and the resulting data proved capable of

10、 serving a range of applications including agriculture,forestry,mapping,geology,hydrology,coastal resources,and environmental monitoring.Around the same time,Esri,a geographic information system(GIS),became the first company to digitize mapping information for commercial use.This early form of quant

11、itative and computational geography enabled users to see the promise of working with large geospatial datasets to understand where things happen(patterns,clusters,hot spots),why they happen,when they happen(location decision),and where things should be located(optimization).As a result of this techn

12、ology,a standard for interacting with geospatial data was established:data as layers(raster and vector),where each layer can be stacked on top of one another and analysis can be done at any level.Geographic Information System(GIS).(Source:City of Newberg)4The GEOINT Playbook|Report by Space Capital

13、Executive SummaryThen,in the 1990s and 2000s,significant advancements in computational infrastructure enabled technical innovations that changed mapping.Although domain-knowledge was still required to interact with geospatial data,three technical innovations unlocked never before seen functionality:

14、1)GPUs were released and became a standard in rendering complex multidimensional scenes,2)APIs gained traction amongst developers for serving and requesting all types of data,and 3)the internet gained traction which educated the modern consumer on the advantages and convenience of web mapping(naviga

15、tion).In the 1990s,a group of engineers at Intrinsic Graphics unknowingly developed the core technology behind Google Earth;3D graphics libraries for video games.Shortly after the release of their inaugural demo,the team created a company called Keyhole in order to improve the product stack to strea

16、m interactive 3-D maps with satellite imagery and aerial photographs.4 Then,in 2004,Google acquired Keyhole,the then three-year old digital-mapping software company.5,6 Historically,Keyhole was primarily used for defense;this acquisition expanded Googles business line for commercial use cases(includ

17、ing advertising).The adoption of cloud,edge computing,AI/ML capabilities,and increasingly powerful geospatial APIs and SDKs are making the benefits of geospatial intelligence more accessible.Developers no longer need to be experts in image capture,data processing,or object detection,and instead can

18、focus on building specialized applications tailored to unique customer needs.Demand for geospatial data is being quantified through marketplaces and used to inform what sensors and platforms will be built in the future.The ability to collect,process,and analyze endless geospatial data is creating po

19、werful new applications that are helping to reshape how entire industries operate and transform our relationship with our planet.Today,the use of small satellites and commercial off-the-shelf components has made it more cost effective to capture timely geospatial data at a global scale.5The GEOINT P

20、laybook|Report by Space Capital Infrastructure6Distribution22Applications39Conclusion52Primary Research53Endnotes54ContentsThe GEOINT Playbook6The GEOINT Playbook|Report by Space Capital Infrastructure7The GEOINT Playbook|Report by Space Capital We refer to Infrastructure as the hardware and softwar

21、e to build and capture geospatial data.Today,a variety of geospatial platforms capture data at different altitudes,benefiting from low-cost components,commoditized storage/compute,and decades of GIS product development.Ground-Station and Edge-as-a-Service have fundamentally changed the way geospatia

22、l solutions are built and are delivering more timely,actionable data.The next evolution of geospatial companies will look and feel more like developer tools and deep tech companies focused on the power of computing and AI/ML.Key TakeawaysInfrastructurePrincipally,this includes networks of connected

23、devices that sense input from the physical environment(sensors ranging from satellites to smart phones).This information creates the foundation to understand our world,from the movement of people and goods to the health of our crops and changes in the weather.For more than 60 years,geospatial produc

24、ts that visualize information through maps and reports have provided a powerful tool to understand our physical world.Satellites transformed modern mapping from static approximations into dynamic reflections of our planet and the activity on it.In the early 2010s,the use of small satellites and comm

25、ercial off-the-shelf components made it more cost effective to capture timely geospatial data.However,downlink bandwidth was limited and prohibitively expensive,latency was considerable,and computer memory and processing power were scarce.Further,there were no common file formats that determined how

26、 data should be managed.Verticalized solutions were built to address the gaps in frameworks and shared infrastructure.This fragmentation stemmed from the fact that geospatial intelligence,and its builders,never intended their systems to be interoperable.Instead,the technology was designed to provide

27、 closed solutions,with controlled economics,serving a small subset of high value customers.This is now changing with the rapid growth in data collection and the highly specialized processing and analytics capabilities that remove the barriers to adoption.8The GEOINT Playbook|Report by Space Capital

28、InfrastructureIn the 1960s NASA began to develop observational technology,focused on the underlying scientific measurements required to inform technology development.The Landsat constellation began with the launch of Landsat 1 in 1972 and continues to this day,providing the worlds longest continuous

29、ly acquired collection of space-based land remote sensing data.7 Similarly,since 2014 the European Space Agency has launched Sentinel missions,clusters of two remote sensing satellites,including radar and super-spectral imaging.Although it was initiated as a research activity,data from the Landsat s

30、ystem soon proved capable of serving a range of applications including agriculture,forestry,mapping,geology,hydrology,coastal resources,and environmental monitoring.While the Landsat commercialization efforts were abandoned in the mid-1990s,the underlying data was utilized by numerous businesses,gov

31、ernments,and scientific organizations.Recent studies estimated the annual economic value generated from the Landsat archive at about$3.5 billion in 2017,up from$2.2 billion in 2011.8 The development of these space-based remote sensing platforms laid the foundation for new platforms to emerge and be

32、integrated into the geospatial stack.SatellitesThere has been exponential growth in the number and type of satellites launched over the past decade.In fact,there were 971 remote sensing satellites in orbit as of April 2021,a 42%increase in just 3 years.9 The global Satellite Earth Observation(EO)mar

33、ket was valued$3.6 billion in 2021 and is predicted to reach$7.9 billion by 2030.10 This rapid growth is being driven by the commoditization of launch services and computing that have lowered the barriers to entry for new data providers.Data at Global ScaleNumber of active satellites from 1957 to 20

34、21Earth Observation Data Cubes DATA AT A GLOBAL SCALE(Source:Jonathan C.McDowell)(Source:United Nations)9The GEOINT Playbook|Report by Space Capital InfrastructureNASAs Space Shuttle had a cost of$1.5 billion to launch to Low Earth Orbit(LEO)at$54,000 per kg and between 1970 and 2010,the cost to lau

35、nch a kilogram to space averaged around$18,500.12 SpaceX was able to reduce launch costs nearly 19 times by vertically integrating production to reduce costs,forcing competition through transparent pricing,and successfully achieving partial reusability.13 Today,a SpaceX Falcon 9 block 5 rocket costs

36、$62 million to launch at around$2,500 per kg to LEO,14 while the cost of a Falcon Heavy averages around$1,400 to LEO.15 SpaceXs Starship aims to become the worlds largest and only fully reusable launch vehicle,which will further lower the cost of launch to$10 to$20 per kg to LEO.16 This has made it

37、more affordable to launch satellites and has attracted new entrants.Processing capacity has been another key barrier to utilizing complex data captured by satellites.In 1971,Intel released their first complete general-purpose Central Processing Unit(CPU)to help advance integrated electronics.CPUs ar

38、e responsible for a computers main functions including input,processing,data storage,and output.In addition to handling critical functionality,they were also responsible for rendering graphics.However,a driving interest in more performant rendering abilities led engineers to develop Graphics Process

39、ing Units(GPUs).In 1999,NVIDIA introduced the first widely available GPU,which included a rendering engine capable of processing 10 million polygons per second;several years later that number increased to 38 billion per second.17 This step change in performance was made possible by the architecture

40、of the GPU-while a CPU handles jobs sequentially(better at calculation)a GPU parallelizes jobs(better at rendering).In 2009,researchers discovered the GPUs promise in building and training machine learning applications.This has led to a new paradigm of GPU-accelerated computing that makes large comp

41、lex data sets usable in real time.“By adding a CUDA powered GPU to the mix we were able to speed up processing capabilities 250 x.This turned looking at 10 years of observation data from an impossible task,to one that could be done in a week.”Brian Furtaw,Senior Data Scientist specializing in GPU ha

42、rdware and software solutions for ML/AI programs,NVIDIAHistorically,launch costs had been a major hurdle to access space.Satellites can weigh thousands of kilograms and be as large as a school bus.11DATA AT A GLOBAL SCALE10The GEOINT Playbook|Report by Space Capital Infrastructure“NVIDIAs platform i

43、s powering everything from simulation to design and operations to asset management on orbit.”Geoffrey Levine,Director of Worldwide AI Initiatives and Global Public Sector Defense,Climate Action and Space,NVIDIASimilar to that of GPUs,Application-specific Integrated Circuits(ASICs)have gained tractio

44、n in recent years.These chip designs apply hardware and computing configurations to solve a specific problem.One well-known example is Googles Tensor Processing Units(TPUs),an Application-specific Integrated Circuit designed to accelerate machine learning workloads.TPUs accelerate the performance of

45、 linear algebra computation by minimizing the time-to-accuracy when training large complex neural networks;training models that used to take weeks,in hours.18 Most ASICs have become focused on speeding up the time to train models;as advancements in hardware evolve to meet the demands for application

46、s development and better enable machine learning and deep learning workflows.Advances at the chipset level have made it possible to not only capture more geospatial data,but also derive insights from it.The changes in processing have also impacted consumer electronics,providing smaller,more powerful

47、 smartphones,and ultimately changing the way satellites are built.Since the early 2010s,satellite engineers began to adopt a new approach of using lean,low-cost,and highly integrated components made for automobiles,mobile phones,and other consumer electronics to develop small and cost-effective sate

48、llites.This opened up the possibility of using commercial off the shelf(COTS)components to create standardized satellite buses from a larger pool of technology suppliers,which significantly reduced the time and cost of development.These changes have allowed new satellite companies to experiment with

49、 state-of-the-art systems,de-risk technical challenges early,increase iterations with fewer expenditures,and achieve incremental revenues as a constellation is being built.GPU vs CPU on Matrix ManipulationDATA AT A GLOBAL SCALE(Source:Qubole)The founding of Skybox Imaging in 2009 is an early example

50、 of these trends converging.The company set out to build a fleet of satellites flying in low Earth orbit(LEO).The team recognized the serious drawback to long development cycles of large satellites and sought to build a cheaper,faster solution by assembling off-the-shelf parts and applying rapid ite

51、rative engineering.19 Space Capital Managing Partner,Tom Ingersoll,was brought in to lead the company and prove the technical viability of this approach.The team launched their first satellite in November 2013 and Skybox became the earliest commercial remote sensing company to deliver high-quality o

52、ptical imaging from a small satellite form factor.In 2014,Tom led the sale of Skybox to Google for$500 million,achieving one of the largest venture-backed exits in the space economy at the time.A wave of new remote sensing companies soon followed including Planet(NYSE:PL,founded 2010),Spire(NYSE:SPI

53、R,founded 2012),and ICEYE(founded 2014).The next generation of remote sensing companies are providing Satellites-as-a-service and developing highly-targeted scientific instruments to improve decision making for commercial,civil,and defense customers.Muon Space,a Menlo Park company founded by former

54、Skybox engineers,is at the forefront of identifying key signals that go beyond traditional sensors.Their approach builds on the scientific foundations developed at NASA while leveraging low-cost commercial satellite infrastructure to achieve higher sampling cadences and more rapid and responsive mis

55、sion development.Jonny Dyer,CEO and Co-founder,ultimately envisions his work will provide data-driven tools to respond to the most complex questions associated with climate change.The company will require multiple space-based platforms to experiment,iterate,and validate predictive signals capable of

56、 understanding and simulating Earths climate.InfrastructureSatellite Pros Global monitoring,best for macro view Large existing body of open source scientific imagery-based data products Low to high resolution:30m+10cm(VLEO)Satellite Cons High CapEx relative to alternative methods Expensive for resol

57、utions 2M Challenging to leverage data without geospatial expertise11+DATA AT A GLOBAL SCALEThe GEOINT Playbook|Report by Space Capital Skybox Imaging satellites being build in clean room.(Source:TechCrunch)12The GEOINT Playbook|Report by Space Capital These platforms rely on a variety of technologi

58、es ranging from stratospheric balloons and drones to vehicle mounted sensors and handheld devices.Each of these capabilities provide unique perspectives,benefits,and cost structures that can compliment satellite systems in geospatial analysis.New PerspectivesInfrastructureStratospheric BalloonsThe g

59、lobal High Altitude Platforms(HAPs)market is expected to reach$4.3 billion by 2026,20 an increase from$3.4 billion in 2022 and over 400%increase from$1.0 billion in 2016.21 This includes both balloons and airships that are able to operate in the stratosphere.These stratospheric platforms have a uniq

60、ue ability to provide persistent coverage over a localized area at a competitive cost when compared with existing alternatives(satellites,aircraft,drones).The data acquired from these assets have been used in industries with a large set of fixed real estate assets like insurance,transportation,energ

61、y,and conservation.Remote-controlled vehicles can also be used for climate science,disaster recovery and response,and military surveillance.NEW PERSPECTIVESToday,a variety of geospatial platforms capture data at different altitudes,benefiting from low-cost components,commoditized storage/compute,and

62、 decades of GIS product development.Geospatial platforms.(Source:Space Capital)13The GEOINT Playbook|Report by Space Capital WorldView captures geospatial data using its high-altitude balloon,called a Stratollite.It is designed to carry payloads of up to 4,500kg and provide the capabilities of a sat

63、ellite,but without the launch costs and development timelines.Hovering above Earths surface at controlled altitudes of 55,000 to 75,000 feet,the specialized stratocraft would float almost twice as high as a typical commercial airline,giving it a more global view than aircraft,but closer to the Earth

64、 than a typical satellite,and therefore able to provide higher resolution.The remote-controlled Unmanned Aircraft System(UAS)can be used for a variety of applications,such as,disaster recovery and response,communication,weather forecasting,and military surveillance.For its test flight,the solar-powe

65、red Stratollite carried multiple payloads,including a 50.6-megapixel camera to demonstrate its potential as a viable high-altitude remote sensing platform.HAPs Pros Monitor specific assets,best for persistent micro view Cheaper data collection than satellites Faster development and deployment timeli

66、nes vs satellites(does not require integration and launch)Ability to monitor a specific place for long periods of time High resolution data:10cmHAPs Cons Specific regions/geographies(not global)Experimental when compared with more traditional forms of collection like Drones and Aircraft+World View k

67、eeps one of its high-altitude balloons afloat for a record 16 days.(Source:The Verge)Worldview geospatial platform in-flight.(Source:World View)InfrastructureNEW PERSPECTIVES14InfrastructureAircraftAerial photography is a human-in-the-loop data collection method that includes collecting photographs

68、using an airborne camera.For this type of data collection,it requires well-trained interpreters and pilots who have experience flying over and capturing“fine-scale landscape features.”22 Until the popularization of drone technology,the pursuit of aerial photography was reserved for military,hobbyist

69、s,and commercial entities with access to aircraft.While most aerial monitoring use cases have been cannibalized by HAPs and drones,monitoring oilfields and conducting search and rescue missions are still well-suited for aerial surveillance.Aircraft are a great,affordable way to test sensors efficien

70、tly for other platforms,such as satellites.The data collected from aerial flights are then also often paired with satellite data as an alternative to ground-based monitoring to derive actionable insights.23 As of 2021,the aerial imaging market was valued at$2.0 billion and is expected to deliver a 1

71、5%CAGR in the next five years,reaching$4.6 billion in 2027.24One company that helped advance aerial imaging is Vexcel Imaging.Vexcel is based in Graz,Austria and was founded in 1992.It is a long-time leading provider of cutting-edge aerial imaging camera products and photogrammetry software.Their co

72、mprehensive lines of aerial cameras provide a wide range of imaging capabilities from wide-area mapping to nadir and oblique camera systems.Vexcels UltraCam Eagle was described as the only commercial sensor capable of capturing 10 cm GSD at 5,000 m above the ground.Vexcel was acquired by Microsoft i

73、n 2006 and became one of the underpinning technology providers of its Bing Maps web service and mapping platform.Microsofts UltraMap photogrammetry software solution helps customers generate aerial imagery and data such as high density point clouds,digital surface models and orthophotos efficiently.

74、Other than imagery,aircraft sensors are frequently used in capturing data for environmental monitoring.Montreal-based GHGSat has adapted their proprietary satellite sensing technology onto their patented aircrafts to accurately measure methane emissions and detect leaks.Aircraft Pros Monitors specif

75、ic assets,strong for micro view Cheaper to collect data than satellites Faster time to data collection than satellites Medium resolution data:6m-1mAircraft Cons Specific regions/geographies(not global)Expensive when compared to drones and HAPs+NEW PERSPECTIVESGHGSat(Source:GHGsats aerial sensors)The

76、 GEOINT Playbook|Report by Space Capital 15InfrastructureDrones&UAVsNEW PERSPECTIVESDrone/UAV Pros Monitors specific assets,best for micro view Cheaper data collection than satellitesDrone/UAV Cons Specific regions/geographies(not global)Requires certified drone operators+The market for drones is ex

77、pected to grow from$3.0 billion in 2021 to around$3.6 billion by the end of this year and reach$4.6 billion by 2027.25 Goldman Sachs has a more aggressive market sizing with estimates of the drone industry reaching$100.0 billion.While 70%of the market is still driven by military use,the remaining 30

78、%is fueled by the rapid growth in commercial use cases in construction,agriculture,insurance claims,and offshore oil/gas and refining activities.Remote-controlled drones have played an important role in lowering the cost and increasing the safety profile of collecting geospatial data in physically d

79、angerous environments.In agriculture,insurance adjusters are using tools that visualize the plant health in real time so that users can“evaluate damage at the fields edge”and receive quantifiable results.This enables software solutions provided by companies like DroneDeploy to act as a“mediation too

80、l for farmers and adjusters”by providing high-quality and reliable geospatial data.26 Another exciting technical advancement for drones has been the integration of GPS technology with these devices,which has made it possible to fly drones over much larger distances and beyond visible line of sight.T

81、his unlocks new use cases for operators and commercial customers that are certified to operate a drone under Part 107 guidelines.27 This ensures drone operators are abiding within set parameters for controlled(well-populated areas)and uncontrolled airspace.New FAA approved mapping products have emer

82、ged to help hobbyists and commercial operators meet the standards and criteria set across airspace.While most drones operate outdoors,startups like Ware have built autonomous drones for the warehouse.This curtails the need to adhere to regulation since drones are operating indoors.Built with proprie

83、tary software,Ware allows large Enterprise customers like Bosch and Honeywell to capture,process,and report deliveries to warehouse managers without a warehouse worker.28The GEOINT Playbook|Report by Space Capital 16The GEOINT Playbook|Report by Space Capital Ground SensorsGround collection is the o

84、ldest method of acquiring geospatial data and is still the most widely adopted.Sensors are deployed at fixed locations or mounted on mobile platforms.Fixed sensors tend to be in close proximity to an area of interest and can produce continuous monitoring at the highest resolutions unattended.Fixed s

85、olutions tend to be used in industries like agriculture and mining.The market size of the fixed sensors is expected to grow moderately at a 4.7%CAGR,from$2.0 billion in 2022 to$2.8 billion by 2028.29 Mobile sensors include those mounted on vehicles and hand-held devices.The most well known mobile gr

86、ound sensing project is Google Street View.Started 15 years ago,Google has collected street view imagery across the globe.This service is now accessible to users in over 100 countries and has accumulated over 220 billion images.Google relied on a variety of mobile platforms to collect the data inclu

87、ding driving,pedaling,sailing,and walking.The custom built cameras simultaneously collect images in multiple directions and are instantly geotagged.The imagery is then overlapped and stitched together into a single 360-degree image.Dashcams are emerging as a cost-effective way of capturing timely st

88、reet level data at scale.Hivemapper,a San Francisco based company,is aiming to solve this challenge by building a decentralized,self-updating map powered by individuals using proprietary dash cams.Contributors will be rewarded with Hivemappers own cryptocurrency for capturing data as they drive.This

89、 new approach represents a potential shift in how maps are built,who owns them,and potential for near real-time updates.In addition to cameras,car mounted LiDAR sensors are now being used in semi-autonomous and autonomous vehicles.These sensors can help capture detailed renderings of physical infras

90、tructure including buildings,roads,lights,etc.The processed data provides a 360 degree view of the assets location and condition,as well as any factors that are affecting visibility or accessibility,such as vegetation encroachment.InfrastructureHow Google Street View Gets Its Pictures.(Source:Reader

91、s Digest)Hivemapper Dashcam.(Source:Hivemapper)NEW PERSPECTIVES17The GEOINT Playbook|Report by Space Capital InfrastructureGround Sensors Cons Specific regions/geographies but can reach scale with relatively inexpensive sensors deployed Privacy concerns becoming more apparentGround Sensors Pros Moni

92、tors specific assets,macro view at scale Cheaper data collection vs satellites Faster time to data collection vs satellites Ability to monitor a specific place for long periods of time Highest resolution data:1mm+Google VPS illustrated in Google Maps(Source:9to5Google)Handheld devices,especially sma

93、rtphones,have become a powerful geospatial platform.NEW PERSPECTIVESModern smartphones are equipped with a variety of sensors and as a result,every person with a smartphone becomes a part of a larger crowdsourcing effort to create a detailed understanding of human activity at a micro and macro level

94、.Geotagged photos and videos provide a tremendous amount of spatial data that is useful in a variety of applications.Snapchat allows users to tag photos with friends and visualize where friends are on Snapmap via the mobile app.Another excellent example is Los Angeles-based dataPlor,that leverages m

95、achine learning to interpret smartphone photos in addition to AI callbots to update the Point-of-Interest(POI)data of local businesses at scale.dataPlor has collected over 150 million POI data available in 100 countries across the globe.Wearables,like Apple Watch,have also become a popular form of g

96、round sensors that is growing rapidly.Geospatial data collected by the wearables are being used by healthcare professionals to better understand the relationships between different physical activities and body health,monitor patient recovery conditions,and send out alerts in emergency situations.The

97、 market size for mobile sensors is much larger than the fixed ones,as the smartphone sensor market alone is forecasted to grow at a CAGR of 17%from 2022 to$379.0 billion in 2030.3018InfrastructureThe growing demand for timely access to data,particularly from satellites,has created bottlenecks that n

98、ecessitate innovation.Efforts to develop a shared ground station infrastructure and reduce costs date back to the late 1990s.Despite these efforts,entrepreneurial activity focused on capturing data outpaced innovations within the ground segment and the problem has persisted.Scalable solutions for sh

99、ared infrastructure are finally being built by the leading tech companies,enabling a new engineering paradigm.Among the key players developing Ground-Station-as-a-Service are Amazon,Microsoft,and Google.This transformation allows the entire processing chain to shift to the cloud,from the data acquis

100、ition phase to analytics and distribution,significantly lowering delivery latencies and increasing access to data.Resolving the BottleneckLearn more about the role satellite communication is playing to move internet traffic off the ocean floor and into space:Lessons from the History of Satellite Com

101、munications RESOLVING THE BOTTLENECKAs we have seen,satellites provide a powerful tool for collecting data and understanding our world on a global scale.But satellites are limited in terms of resolution and frequency,and benefit greatly from complementary data collected by balloons,drones,and ground

102、 sensors.As governments,enterprises,and customers increasingly rely on geospatial data for decision making,these platforms are becoming critical infrastructure.The GEOINT Playbook|Report by Space Capital 19The GEOINT Playbook|Report by Space Capital CompanyDateMilestone2018Amazon was one of the earl

103、iest Tech Giants to get involved in the space economy when in 2018,the company launched AWS Ground Station to provide a global network of ground stations in close proximity to the global network of AWS data centers.With AWS Ground Station,data providers no longer need to worry about buying,leasing,b

104、uilding,scaling,or managing their own ground stations.Later,the team launched a network of 3,236 satellites known as Project Kuiper to provide broadband internet access.2020Microsoft announced plans to compete with Amazon in connecting satellites to the Azure cloud for processing and storage.Azure O

105、rbital Ground Stations enable satellite operators to downlink data directly from a satellite to Azure Storage and then access additional services such as Azure AI and Azure Data Analytics.Their infrastructure also provides users with security and compliance built-in,reducing the risk for engineers t

106、hat were previously responsible for maintaining,building,and certifying their services.These ground stations are compatible with Non-Geostationary Earth Orbit satellites such as LEO and MEO satellites.322020Google is actively expanding ground station services to better support space companies.The co

107、mpany recently announced a partnership with Leaf Space providing Ground-Station-as-a-Service to the space ecosystem and further opening the doors for the development of the space economy.They also announced a partnership with SpaceX to connect SpaceXs Starlink satellite broadband network to the Goog

108、le Cloud infrastructure.33 This will enable the delivery of data,cloud services,and applications to enterprise customers at the networks edge.Ground-Stations-as-a-serviceAs previously noted,the reduction in launch costs has led to an increase in satellites in Low Earth Orbit(LEO)as opposed to geosta

109、tionary orbit(GEO).This is important because unlike GEO satellites,which are fixed,LEO satellites“move relative to the Earths surface.”This results in connectivity challenges since each LEO satellite requires line of sight to a ground antenna to facilitate space-to-ground communications.This has led

110、 to the entrance of companies attempting to address these problems by providing“global infrastructure of antennas,processing equipment,and software for satellite operators”.31InfrastructureRESOLVING THE BOTTLENECK20The GEOINT Playbook|Report by Space Capital Edge-as-a-serviceProcessing at the edge i

111、s becoming another important tool that can reduce delays,downlink bandwidth,and costs across geospatial platforms.Edge computing(or IoT edge processing)is the process of taking“action on data as near to the source as possible rather than in a central,remote data center,to reduce latency and bandwidt

112、h use”.34 Both hardware and software innovations are needed to unlock edge processing use cases given the challenges of processing data on device.CompanyDateMilestone2018Berlin-based Mobius Labs is building a new generation of AI-powered computer vision(CV)that can accurately detect objects using li

113、ghtweight models optimized for low-latency,low-data scenarios.Their solution is hardware agnostic and backwards compatible,providing the immediate benefits of edge processing across a wide range of geospatial platforms,from handheld cameras,to drones and satellites.Mobius supports both optical and s

114、ynthetic aperture radar(SAR)sensors,and outperforms other solutions using fewer annotated samples,reducing the time and cost to train new models for specialized object identification.2021An early stage Australian startup that is developing a proprietary edge computing technology that can reduce the

115、cost of downlinking satellite data by 5 to 20 x.They are developing the Space Edge computer-an onboard computing system that will give EO satellites the ability to process images captured on the satellite itself.2022Palantir(NYSE:PLTR),a company that builds data intelligence solutions,partnered with

116、 Satellogic(NASDAQ:SATL)to launch their first satellite with a specialized hosted payload using edge computing hardware to run Palantirs Edge Artificial Intelligence(AI)platform.This joint project allows customers to run a variety of models against an image and detect land,sea and air vehicles,build

117、ings,clouds,and terrain.The Edge AI platform also adapts the input image according to each models requirements and stores relevant inferences,allowing downstream post-processing services to use this information to generate additional outputs while sequencing multiple models in parallel.35 This shift

118、 to the edge allows customers to identify objects of interest,improve the accuracy of image segmentation,remove false positives,adjust projections based on their coordinate systems,and generate lightweight PNG thumbnails for quicker download all at the same time.InfrastructureRESOLVING THE BOTTLENEC

119、K21InfrastructureThis shift will be critical in solving the fundamentally hard problems of transforming geospatial data into insights that can then be applied to a broad set of applications across industries and technical abilities.Ground-Station and Edge-as-a-Service have fundamentally changed the

120、way geospatial solutions are built and are delivering more timely,actionable data.That is why we believe the next evolution of geospatial companies will look and feel more like developer tools and deep tech companies,focused on the power of computing and AI/ML.Mobius Labs accurately detects critical

121、 targets and removes unusable data,before ever downlinking.(Source:Mobius Labs)RESOLVING THE BOTTLENECKThe GEOINT Playbook|Report by Space Capital 22The GEOINT Playbook|Report by Space Capital Distribution23The GEOINT Playbook|Report by Space Capital We define Distribution as the hardware and softwa

122、re to structure,process,analyze,and disseminate geospatial information.DistributionAdoption of cloud,AI/ML capabilities,and increasingly powerful APIs and SDKs are expanding use cases and adoption beyond data engineers,GIS specialists,and subject matter experts.Data fusion is beginning to provide re

123、al-time situational awareness built on automated,algorithmic analysis leveraging the benefits of multiple platforms and sensors at scale.Marketplaces are allowing geospatial demand to be quantified and used to inform what sensors and platforms are to be built.There are currently over 280 types of da

124、ta in demand across a broad customer base,with only 10-15 available today.Key TakeawaysWhile geospatial products have been around for more than 60 years,the lack of automated,machine-to-machine solutions has kept costs high and adoption low.Recent developments have helped fuel new platforms,tools,an

125、d business models that enable engineers and end-users to better utilize geospatial information.Geospatial data comes from a variety of devices that speak fundamentally different languages.Historically,making use of this information has required the collaboration of highly trained professionals inclu

126、ding data engineers,subject matter experts,and GIS specialists to create even basic reports.Constraining matters further,only 5%of data science and other engineering fields are trained in the tools to analyze geospatial data.36 The need for human intervention and specialized knowledge has kept costs

127、 high and access low.24“Spatial is a multiplier across industries and there are many datasets still waiting to be spatialized.What we need are more spatial talent and a common language or framework to foster collaboration.”John P.Wilson,Professor of Spatial Sciences,University of Southern California

128、DistributionAdditionally,new and existing end users have become accustomed to a world where they are provided timely answers to specific questions with no prerequisite knowledge.In a geospatial context,consultants have traditionally bridged this gap by helping customers navigate a fragmented GEOINT

129、ecosystem.37 However,this approach merely served as a temporary solution,during a transitional period,to accelerate adoption by reducing the technical barriers to entry.A structural change would be necessary to unlock the enormous potential of the geospatial market.The GEOINT Playbook|Report by Spac

130、e Capital 25The GEOINT Playbook|Report by Space Capital DistributionData Fusion as the FoundationThe more effectively we can merge different spatial data to create unified and uniformly accessible data,for example vector and raster,the more types of spatial relationships we can derive.This is the co

131、ncept of spatial data fusion,a way to enable new cross referencing and data visualization to provide a better understanding of a situation or problem,which wouldnt have been possible otherwise.Data fusion is not limited to the merging of data from different sources,but from different time periods as

132、 well.With an increase in the variety,capability,and number of sensors,the amount of data(optical,spectral,multispectral,LiDAR,SAR,DEM,etc.)intended to enhance decision-making and human-system performance can lead to information overload and obscure the most relevant aspects of a situation,leading t

133、o decision fatigue.Fusing raw sensor dataCo-registering and overlaying sensor outputsExtracting features/key points of interest from each sensor independently and then fusing the feature setsApplying independent algorithms to each sensor(e.g.,a target detection algorithm on each sensor individually)

134、and then fusing the algorithm(e.g.,detection)resultsImplementing change analysis that occurs on the same type of spatial data over a period of time(multitemporal data)Data fusion can address this challenge by making sense of disparate data and leading to a richer understanding with less noise by:DAT

135、A FUSION AS THE FOUNDATIONOne of the many characteristics that makes geospatial intelligence so powerful is its ability to overlay and connect multiple categories and types of information.26The GEOINT Playbook|Report by Space Capital DistributionA recent thesis by Andreas Holmberg at Ume University

136、detailed how scientists combining multispectral and radar data enables more refined analyses over broader scales than either can alone for the mapping and analysis of habitat extent,vegetation health,land use change,and plant species distributions at various scales.38 The study showed how the combin

137、ation of radars structure data and multispectral sensors reflectance data are highly complementary to improve the accuracy of assessing and monitoring biodiversity at scale.A commercial example of data fusion is provided by Planet,who recently implemented a methodology to“enhance,harmonize,inter-cal

138、ibrate,and fuse cross-sensor data streams”for a variety of use cases.The companys Fusion Monitoring platform combines daily,global optical imagery with additional data sets to deliver a consistent stream of information about a customers area of interest.Planet then merges SAR data39 into the fusion

139、line,providing improved sensing to customers,particularly those in the agricultural sector.This additional data stream boosts the reliability of agronomic models that consume fusion data,especially in very cloudy regions where SAR data has the advantage of imaging through clouds and weather that obs

140、cure optical data.Through this fusion technique,Planet offers superior continuous monitoring of crop health and captures key events like harvesting or the effects of storm damage.This image below shows three image tiles of Planets quality assurance product that are built with fusion technique.The la

141、yer in the middle was fused together with ten different captured orthoimagery(satellite images that are geometrically corrected)and the images on the sides were fused together with scene data with different spectral bands.Planet Labs fusion quality assurance product layers(Source:Planet Labs)Data fu

142、sion can produce new insights that are indistinguishable with a single type of sensing.DATA FUSION AS THE FOUNDATION27The GEOINT Playbook|Report by Space Capital DistributionTrimbles accurate geospatial solutions can help farmers build out autonomous land preparation,planting and seeding,implementin

143、g control,input management,and harvesting tools.Fusing data provides real-time situational awareness that requires accuracy,consistency,and reliability.The migration to the cloud is helping to address these needs and build automated,algorithmic analysis from multiple platforms and sensors at scale.C

144、anopy changes observed through different data types on Trimbles eCognition software(Source:LiDAR Magazine)Trimble,an industry leader in geospatial tools,made their name by providing a wide variety of advanced GPS receiver systems to help improve efficiencies in surveying,measuring,monitoring,calibra

145、tions,and data fusion is a core feature inside many of their software solutions.In a geospatial analysis solution buildout,the separate steps in image interpretation like object creation,object classification,object detection,and objection modification benefit greatly from looking at a combination o

146、f different data sources.Trimbles eCognition software was developed with this specific understanding in mind.In order to further distill meaning from the data observation when segmenting and classifying physical semantics,eCognition examines the imagery as both pixels and contexts and includes lever

147、aging all relevant input data such as color,shape,texture,size of the object and its surrounding environment,which comes in different data types such as spectral raster data,3D point cloud data,and thematic data from vector layers.The image below shows how Trimbles software helps bring thermal imagi

148、ng and 3D point cloud data into one place to help scientists understand canopy changes dynamically over time.“One of the things that can take our GPS system to the next level is combining multiple technologies together,we are now bringing multi constellation GNSS receivers with IMU,looking closer to

149、 photogrammetry,visual odometry,all of these technologies are coming together to make it so surveyors can always get position information wherever they are working at.”Ron Bisio,Senior Vice President of Geospatial,TrimbleDATA FUSION AS THE FOUNDATION28The GEOINT Playbook|Report by Space Capital Dist

150、ributionAs-a-service ParadigmThis began in the late 1990s with efforts to develop shared infrastructure that would reduce capital expenditure.Since then,Big Tech companies have realized the value of hosting geospatial data on their cloud and,as an outcome,have taken strides to make it easier for com

151、panies to host,process,analyze,and visualize these heavy datasets.While developed back in the 1990s,cloud computing did not achieve broad adoption until 2006 when companies like Google began to use this term to describe the new paradigm of processing40 and Amazon launched Amazon Web Services.41 Anot

152、her decade would pass before cloud service providers like Amazon reached an inflection point,growing from$4.6 billion in 2014 to$25.7 billion in revenues in 2018.42 While many technology companies were quick to see the promise of this new architecture,the Satellite industry was slow to adopt.The hes

153、itancy to move from on-prem to the cloud was predominantly driven by data providers challenges associated with the technical migration.More specifically,an uncertainty about how to 1)migrate existing workloads to the cloud,2)unwillingness to walk away from the preexisting investment into costly infr

154、astructure equipment,3)security concerns about hosting sensitive customer data,and 4)federal customers refusing to have their imagery hosted.43 That is why when,in 2017,DigitalGlobe migrated its entire catalog of imagery to AWS-100 PB-it effectively set a new standard for the wider remote sensing in

155、dustry to evolve from large file-transfer protocols and delivery workflows to a cloud-first storage and analytics infrastructure.44DigitalGlobe moves to the cloud with AWS Snowmobile.(Source:MAXAR)The exponential growth in geospatial data has driven the need for new ways to process,transform,and ana

156、lyze information,which has resulted in a shift to the“as-a-service”paradigm.AS-A-SERVICE PARADIGM29The GEOINT Playbook|Report by Space Capital DistributionCloud-as-a-ServiceIn recent years,cloud computing has become an integral component in the management of geospatial data.According to NSR,a satell

157、ite and space market research and consulting firm,486 PB(petabytes)of raw satellite imagery will need to be downlinked to Cloud servers over the next 10 years.45 As mentioned in the Infrastructure section,one way companies are evolving to meet the deluge of data is moving processing to the edge.A se

158、condary effect is Big Tech companies are releasing additional products and features specifically for geospatial use cases.CompanyDateMilestone2018With Google Cloud,companies can establish more secure,reliable and scalable systems and save time on the development process using its automated code mana

159、gement and deployment tools that are tightly integrated with other essential services.Google Earth and Google Maps have also expanded their cloud services offerings by integrating directly with Google Cloud.This has led to the tech giant bundling products together in order to create a platform for g

160、eospatial developers.2020AWS is establishing a new unit called Aerospace and Satellite Solutions,led by former U.S.Air Force Maj.Gen.Clint Crosier who most recently directed the establishment of the U.S.This new department seeks to offer a wide array of cloud services that can enable high-performanc

161、e computing,otherwise known as,the provision of high-throughput capabilities for digital modeling and simulation.2020Azure Space is a new platform for space operators that promotes“ubiquitous connectivity,space data and AI,and developer tools.”The service aims to help Microsoft,along with government

162、s,scientists,enterprises,and other stakeholders access to a multi-petabyte catalog of analysis-ready global environmental data,geospatial APIs,and a flexible development environment to meet sustainability commitments by 2030.2021IBM Cloud Satellite is a hybrid cloud services architecture that promot

163、es a“layer of cloud services for clients across environments,regardless of where their data resides.”46 This offering is an extension of the IBM Public Cloud as each cloud satellite location hosts an instance of the IBM public cloud with the flexibility to“deploy and run apps across on-premises,edge

164、 computing and public cloud environments from any cloud vendor.”AS-A-SERVICE PARADIGM30The GEOINT Playbook|Report by Space Capital Technical wireframe that shows how using Amazon SageMaker improved DigitalGlobes cache by more than a factor of two,allowing them to cut their cloud storage cost in half

165、.(Source:Amazon AWS)Deployment of an Azure Orbital Ground Station at Quincy Data center in Washington.(Source:Azure)DistributionAs the shift to the cloud continues,geospatial stakeholders have started to identify patterns across technical reference architectures for geospatial tooling,like processin

166、g pipelines and geospatial platforms,as well as broader market trends in the push to host this data.In tandem,we are seeing a swell of changes in how geospatial practitioners leverage AI/ML to build on the foundational innovations across ground-stations,edge processing,and cloud infrastructure.47 AS

167、-A-SERVICE PARADIGM31The GEOINT Playbook|Report by Space Capital APIs-as-a-serviceSince 2005,geospatial APIs(Application Programming Interfaces)and SDKs(software development kits)have helped developers,geographers,and non-geographers make sense of geospatial data.Developers are now able to make use

168、of geospatial software libraries and developer tools with modern programming languages,broadening the user base dramatically,from geospatial experts to general software engineers.Integrations with modern software architecture are helping to build the future where geospatial data is accessible like a

169、ny other data in that there are four distinct phases to making it useful:ingest,process,store,and analyze.Now that each phase is primarily run on the cloud,the opportunities to refine processes,collaborate more effectively,and automate undifferentiated human interventions have increased significantl

170、y.Mapbox is a developer platform that sells APIs,Mobile SDKs,and Location-Solutions.The company spun out of Development Seed,an international development consultancy,which started out working with governments to visualize large geospatial datasets.In 2008,the team built a suite of developer tools on

171、 top of OpenStreetMap for customers that needed location data beyond the borders of the United States.These tools focused on creating new ways to ship geographic data updates to clients where the data can be rendered dynamically on their devices.As engineering teams began to see the added value of t

172、he Mapbox product stack,the company signed their first commercial client,Foursquare.In 2013,the team raised a$10 million Series A from Foundry Group to build the future of interactive,mobile maps,all powered by an API-first product platform.48 Mapbox was valued at$1.0 billion in 2018 after raising a

173、$160 million Series C from SoftBank.Recently,the company has faced significant competition from AWS as the tech giant has expanded its offerings in geospatial and location services.Interactive,thoroughly customizable maps in native Android powered by vector tiles and OpenGL.(Source:AndroidRepo)Distr

174、ibutionAS-A-SERVICE PARADIGM32The GEOINT Playbook|Report by Space Capital Iggy is another cloud-based location platform that licenses endpoints to customers interested in enriching their data.Their flagship product,Enrich,powers common use-cases including:real estate prospecting,training machine lea

175、rning algorithms,and enriching geospatial data.Hailing from Airbnb,the teams infrastructure makes it easy to extract location data from images and add metadata to points of interest.Their endpoints also enable users to select the features they want to enrich and then provide them with coordinates ov

176、er an API,which allows them to run joins,aggregation,and analyses.DistributionAS-A-SERVICE PARADIGMDall-E creating an AI-Generated Image based on a Text&Image Prompt.(Source:OpenAI)Learn more about the history and evolution of mapping technology in the digital age:Mapping Technology and the Google E

177、ffect Algorithms-as-a-serviceThe algorithms-as-a-service(AaaS)business model has garnered interest from a broad swath of the technical community due to the promised ease-of-use and ability to commercialize and sell IP in a new way.Traditionally,algorithms were hosted on the backend and then the serv

178、ices from those algorithms were made available to users.Now there is more interest in directly licensing algorithms to other companies and/or serving algorithms through a marketplace.So,algorithms are now at a point in which they are being exposed for use directly by companies given the interest in

179、building applications in-house,using other companies algorithms natively,and solving rudimentary and mathematically complex tasks with AI and ML.One example of an algorithm as a service is Dall-E,a neural network that“creates images from text captions”for a variety of“concepts expressible in natural

180、 language.”49 The team was able to use a 12-billion parameter version of GPT-3,a popular autoregressive language model,then train it to generate images from text descriptions which has become popular amongst designers,marketers,and other creatives.33The GEOINT Playbook|Report by Space Capital Platfo

181、rm-as-a-serviceThe evolution of AI was built on top of massive training data sets,but the expansion comes with real challenges.Finding,validating,and generating data for machine learning is a complex,often inaccurate task.Gartner estimates that,by 2022,85%of AI projects will deliver incorrect outcom

182、es due to biased data.50 There are several challenges that current machine learning researchers are facing.51Synthetic data is not a new concept and has been used by astrophysicists for decades.The historical approach to synthetic data was often seen as a lower-quality substitute,useful only when re

183、al data is inconvenient to get,expensive,or constrained by regulation.However,recent innovations in synthetic data may help data scientists unlock highly-valuable training data at scale without compromising on quality,balance,or accuracy.Mavericks research shows that synthetic data is actually on a

184、trajectory to go from an alternative data source to becoming the main force powering the future of AI.52 DistributionAS-A-SERVICE PARADIGMData is incomplete:AI needs both large and diverse datasets,but real data is often incomplete,excluding infrequent scenarios that are critical for AI performance.

185、Data is expensive:It is hard to collect,integrate,store,and maintain.Data is biased:Even if data perfectly reflects reality,it can encode biases present in the real world that we would like to remove.Data is restricted:Regulation is increasingly limiting data use for AI.(Source:COGITO)34The GEOINT P

186、laybook|Report by Space Capital DistributionAS-A-SERVICE PARADIGM(Source:Aldec)Synthetic Data can transform GEOINT based processing and simulation with a leap in AI efficacyJust as humans learn from both real world experiences and fictional scenarios,the ability to simulate and explore hypothetical

187、outcomes allows a unique type of data richness for complex AI models.Since synthetic data includes a higher degree of possible outcomes,it is a more efficient way to explore infrequent or unknown situations.In a case study that researched combining synthetic data with real data to improve detection

188、results in satellite imagery,using a mix of synthetic(95%)and real(5%)data produced algorithm improvements by up to 20%.53 While synthetic data has promising potential for a seemingly infinite number of use cases,geospatial analysis is an early adopter.A new approach to synthetic data leveraging phy

189、sics-based simulation models may address previous limitations.35The GEOINT Playbook|Report by Space Capital Rendered.ai has developed a Platform as a Service(PaaS)for synthetic data,creating a Common Application Framework to enable the production of physics-based synthetic datasets for AI training a

190、nd validation.The platform works with both 2D and 3D content and has the power to unlock more advanced processing capabilities such as spatial data fusion.Rendereds platform users are currently generating synthetic satellite imagery data for rare object detection,identifying land cover segmentation,

191、and building detectors for asset conditions.Users are also applying its tools to improve the ability to extract data from drone imagery,video,and other geospatial content types.54Rendered has partnered with Orbital Insight and UC Berkeley to support the National Geospatial-Intelligence Agency.55 The

192、 project team plans to demonstrate that:DistributionAS-A-SERVICE PARADIGM“When you talk about how big could this opportunity space be,its as big as AI and not just where AI is,but where it can go.”Nathan Kundtz,CEO,Rendered.aiRendered combined ArcGIS imagery services and 3D models of planes(Source:R

193、endered.ai)Synthetic data can be used with real data to increase the training performance of AI for rare and unusual objectsMuch less synthetic data can be used to train AI than would be required using real datasetsPotentially much less real data may need to be collected to be able to detect rare an

194、d unusual objects36The GEOINT Playbook|Report by Space Capital Governments have long been the primary customer and enabler of large-scale surveying,mapping,and remote sensing initiatives.In 1992,the Land Remote Sensing Policy Act permitted private companies to enter the satellite imaging business.Ea

195、rly movers quickly consolidated to raise barriers to entry and limit competition.Today,companies like Maxar have vertically integrated to further solidify their position in the market:Vricon,a subsidiary that provides 3D geospatial analytics through satellite data;DigitalGlobe,a vendor and producer

196、of high resolution satellite imagery;Radiant Solutions,a subsidiary that provides data modeling;and analytics for satellite data and SSL,a subsidiary that constructs satellites.56 Geospatial Information Systems,and their builders,never intended for their systems to be interoperable.Instead,these sys

197、tems were designed to be closed experiences with controlled economics designed to serve government customers.Satisfied with their low volume,high margin business,incumbent providers have failed to deliver broader access to commercial satellite imagery and the benefits of geospatial data.In 2009,new

198、commercial remote sensing providers like Skybox,Planet,and others were entering the market,deploying low cost satellites with the intent to expand the commercial market beyond government customers.In an effort to stimulate customer adoption,new entrants began hosting workshops to help customers unde

199、rstand how they might build geospatial applications on top of this new infrastructure.In 2019,Planet hosted their first-ever user conference,Planet Explore,with over 400 attendees from 28 different countries who were represented across dozens of industries.The company offered a single data type with

200、 specific spatial,spectral,and temporal resolution,albeit with a market-leading capability of imaging the entire Earths landmass daily.57 For many of these new users,one data type wasnt enough to answer the business questions modern enterprise customers were asking.The marketplace model is helping t

201、o address access and adoption issues by aggregating data from multiple commercial satellite imagery providers and providing customers with a single point of access.This has helped create a new customer segment,one that purchases highly targeted imagery across a broad swath of use cases at smaller co

202、ntract sizes,which has largely been ignored and is currently unserved by existing commercial providers.More data providers are beginning to sell through these marketplaces because that is where the customers are.Specifically,customers save time and money as they no longer need to survey the landscap

203、e of data providers,negotiate pricing and service agreements,manage multiple access points and integrations,and meet minimum purchase requirements.Unlocking Access through CodeDistributionUNLOCKING ACCESS THROUGH CODEMarketplaces provide customers a higher chance of finding the insights that meet th

204、eir needs.37The GEOINT Playbook|Report by Space Capital SkyWatchs EarthCache platform offers access to many types of imagery output(Source:SkyWatch)SkyWatch,a leading marketplace for satellite and other geospatial imagery,connects customers with imagery requests.58 The platform is a first of its kin

205、d in servicing the needs of commercial organizations,whose requests are fundamentally different from traditional imagery purchasers(government and defense).By aggregating data,streamlining the request process,and enabling users to purchase smaller swaths of imagery at a higher frequency,SkyWatch has

206、 solved the hardest part of working with remote sensing data:access.59 As the company has grown,SkyWatch has been able to fully automate the process of ordering an image all the way through delivery.60 This has removed multiple layers of human intervention and reduced the costs to access satellite d

207、ata.EarthCache,the companys first product,has successfully met their growing demand by leveraging an API-first business model.This enables a pay-as-you-use pricing strategy that resonates with their developer community.TerraStream,an extension of EarthCache,was subsequently launched to quantify cust

208、omer demand,automate tasking for their supplier network,and help the team better engage in a product-led-growth strategy;enabling customer demand to direct platform and sensor development.SkyWatch has identified more than 280 types of data in demand across their customer base,with only 10-15 availab

209、le today.This suggests that the market is currently significantly undersupplied and that additional geospatial platforms and sensors are required to meet customer needs.DistributionUNLOCKING ACCESS THROUGH CODE“You should be able to leverage any sensor in orbit with just a few lines of code.You shou

210、ldnt need to know anything about those sensors,who owns them,what capabilities they have.You just know you have a problem to solve in your business and you just need those sensors to solve those problems.”James Slifierz,CEO,SkyWatch38The GEOINT Playbook|Report by Space Capital Several established sa

211、tellite providers have experimented with launching their own marketplaces including Airbus who owns Up42 and OneAtlas,and Maxar who launched GBDX Notebooks to empower developers and data scientists to access their imagery and machine learning tools.However,these solutions have failed to gain broad a

212、doption,primarily due to their inability to aggregate data from multiple providersthe incumbents competitors were hesitant to contribute their own data.Google currently has a multi-petabyte catalog of satellite imagery and geospatial datasets with planetary-scale analysis.However,they continued to l

213、everage this data to generate in-house capabilities and expand their relationship with developers.The multiple layers of transformation,processing,and specialization make it clear how complex geospatial data is.The adoption of cloud,edge computing,AI/ML capabilities,and increasingly powerful geospat

214、ial APIs and SDKs are making the benefits more accessible.Developers no longer need to be experts in image capture,data processing,or object detection,and instead can focus on building specialized applications tailored to unique customers,similar to the democratization of GPS brought about by Apples

215、 App Store.We are now beginning to see a world where demand for geospatial data is being quantified and used to inform what sensors and platforms are to be built.DistributionWhile some barriers around pricing and education still remain,the rapid innovation across the Distribution layer has already e

216、xpanded the market for geospatial data and we believe the transformation will continue to accelerate as a new era of geospatial applications emerge.39The GEOINT Playbook|Report by Space Capital Applications40The GEOINT Playbook|Report by Space Capital We define Applications as any specialized hardwa

217、re and software that benefits from geospatial data to address industry and customer-specific needs.Technical applications have enabled machine-consumption of data and reduced the need for humans-in-the-loop,expanding the user base that can build on top of geospatial data,similar to Location-based Se

218、rvices when iPhone embedded GPS.The next evolution on geospatial will deliver integrated business intelligence and real-time automations based on a variety of specialized user experiences,including photorealistic 3D and immersive.Over time,geospatial applications could appear to be self-aware,automa

219、tically interpreting changes in the sensor ecology.The algorithms that reside within the mapping stack would be able to detect and understand critical decisions and initiate the appropriate response.Key TakeawaysWhile it is not new for geospatial consultants and platforms to provide tailored service

220、s,these solutions have remained bespoke and unscalable.The adoption and development of new services within the Distribution layer of the stack now make it possible for highly specialized industry experts to build scalable applications that dont look like a traditional geospatial company.These new ap

221、plications are inherently end-user facing and provide the most essential truths and contextual insights in a concise,actionable experience that leads to timely decisions with limited human intervention.Applications41The GEOINT Playbook|Report by Space Capital ApplicationsThanks to adoption of cloud

222、services,edge computing,AI/ML capabilities,and increasingly powerful geospatial APIs and SDKs,we are now seeing companies build solutions that leverage machine-consumption of data and reduce the need for humans-in-the-loop,which is enabling venture scale business models in GEOINT for the first time.

223、We are now witnessing the birth of a new market that could exceed the$36.0 billion Location-Based Services market.61 The pioneers of this new geospatial model are building solutions that look different from the mapping and analytics platforms of the past.Highly specialized Applications that prioriti

224、ze industry knowledge and customers needs are growing rapidly and addressing the needs of multi-trillion dollar global industries like Agriculture,Insurance,Climate Markets,and Augmented Reality with the potential to achieve venture-scale exits.Looking forward,we see tremendous opportunity for compa

225、nies that are building products that are immersive and reimagine our current conception of the user-interface to interact with spatial data.As governments,businesses,and consumers increasingly rely on geospatial intelligence to make decisions and automate operations,the underlying technologies will

226、form next generation digital infrastructure that serves as an“invisible backbone”to power our global economy.Today,GEOINT is evolving from a field dominated by mapping experts and geospatial consultants into a commercially available toolset that integrates seamlessly into existing workstreams,automa

227、tes complex tasks,and enables the development of highly specialized end-user applications.42The GEOINT Playbook|Report by Space Capital ApplicationsTHE“NEW”SPATIAL WORLD Regrow helps farmers examine historical and current crop conditions,identify interventions,and assess the impact of managements de

228、cisions on soil and ecosystem health.Regrow recently partnered with Remington,one of the worlds largest third-party seed producers,to help manage their seed production across millions of acres of land.Cross-pollination is one of the biggest threats for Remington.For crops like sunflowers,that are na

229、turally pollinated by bees,a nearby field can pose a real threat to the genetic purity of the seed.In the past,Remington has had to rely on word of mouth and local scouting to identify threats of cross-pollination.Regrows FluroSense platform provides the geospatial insight Remington needs to monitor

230、 crop growth from afar and proactively address any crop stress areas to mitigate potential yield loss.Using satellite imagery and AI,FluroSense can automatically identify specific crop types and retrospectively assess sunflower field locations to identify any threat from other crops.With this knowle

231、dge,Remington can react quickly by changing the crop rotation early in the season or isolating plots that may have been contaminated.Regrows FluroSense platform.(Source:Regrow)AgricultureThe global agriculture market is expected to grow from$11.3 billion in 2021 to$12.5 billion in 2022 at a compound

232、 annual growth rate(CAGR)of 11.1%.The agriculture market is expected to grow to$18.8 billion in 2026 at a CAGR of 10.7%.62 As the human population grows to a projected 10 billion people by 2050,feeding this number of people will require a step change in how we grow crops.With climate change,water sc

233、arcity,and soil erosion bringing on additional challenges,geospatial data is helping power agriculture innovation.Named Precision Sustainable Agriculture(PSA),this new method combines precision agriculture with advances in big data analysis and leverages the latest automation,AI,and sensor network t

234、echnologies.Popular PSA applications include soil moisture monitoring,fertilization and water management,crop growth optimization,and autonomous tractors.As a result,the AgTech market is growing at an accelerating pace;analyzing everything from water quality to wheat protein.The global agritech-as-a

235、-service market is expected to triple in size by 2025 compared to its state in 2020,while the broader agriculture technology market value is expected to grow 2.5 times from approximately$9.0 billion to$22.5 billion.63The“New”Spatial World43The GEOINT Playbook|Report by Space Capital InsuranceThe glo

236、bal insurance market was forecast to grow by 13 percent from 2020 to 2021,reaching$5.5 T as the insurance industry recovers from the global coronavirus(COVID-19)pandemic.Following this,forecasts suggest the industry will grow at a compounded rate of around 6%per year,reaching almost$6.4 trillion U.S

237、.dollars by 2025.66Insurance underwriters are constantly looking for ways to improve risk management.Geospatial models help insurers quantify exposure to risk factors by analyzing past events and determining an areas susceptibility to a given hazard.One of the more pertinent trends that has grown ou

238、t of the use of geospatial data is the growth of parametric insurance.Arbol has launched an insurance marketplace that connects weather exposed businesses with capital providers like insurers and asset managers.Arbols parametric insurance platform removes information asymmetry and reduces administra

239、tive costs by leveraging trusted data sources from NASA,NOAA,commercial satellites,and weather stations to verify and validate claims.Adopting a parametric insurance model means that Arbol pays out based on predetermined,verifiable,and objective metrics that customers choose at the beginning of thei

240、r contract,such as yearly deviation in rainfall amounts or temperature.Because parametric contracts are data-driven,they can achieve a level of precision,reliability,and cost effectiveness that traditional insurance cannot.The increased efficiency and global scale also means that they can provide se

241、rvices to small farmers and other micro-entrepreneurs that have historically been priced out of the market.ApplicationsArbol marketplace.(Source:Arbol)The global market volume of agricultural robots totaled around 9.3 billion units,this number is expected to grow nearly 4 times by 2030 and reach 36

242、billion units.64 Bear Flag Robotics,acquired by John Deere in 2021 for$250 million,has developed the technology to turn any tractor into an autonomous vehicle.By using geospatial data for route planning,farmers are able to maximize efficiency,save fuel,and eliminate excessive tilling.65 Ultimately,w

243、e believe this ties into the broader macrotrend to automate activities that require a high time-investment and low-output per hour.Field robotics is another AgTech area that is growing quickly.THE“NEW”SPATIAL WORLD44The GEOINT Playbook|Report by Space Capital Legacy insurers have been hesitant to pr

244、ovide property insurance policies in and around floodplains due to changing weather patterns and a lack of confidence in their ability to model risk.This represents a tremendous market opportunity as 75%of losses due to flooding are uninsured67,leading to approximately$29.0 billion in annual losses

245、for the past ten years.68 The amount of global insured losses from natural catastrophes driven by floods and storms climbed sharply in 2021 to$82.0 billion69,nearly tripled the average of the past ten years and this number continued to stay high as Swiss Re estimated the loss to be$38.0 billion in t

246、he first half of 2022 alone.70 Cloud to Street,a New York based analytics company,has become a leading flood mapping platform that provides real-time monitoring of flood risks worldwide using geospatial data derived from satellites.Traditional hydrology modelers rely on expensive and slow ground dat

247、a.This method has underperformed direct satellite observations because of the complex and dynamic nature of flood events.A static flood map becomes unusable after days or even hours as water rises and recedes very quickly,driven by topography,drainage systems,built structures and many other variable

248、s.At the same time,since flood changes are hyperlocal,small deviations in elevation and drainage systems can result in vastly different risk assessments for individual houses.This is where new types of geospatial data,such as SAR,have proven to be extremely useful to accurately measure the slightest

249、 changes and help the insurers and home owners asset changes in risk.Cloud to Street has fused together a global network of optical,radar,and microwave satellites to monitor floods anywhere in near real-time.Combined with 30+years of historical geospatial record,their platform can identify infrastru

250、cture failures,urbanization trends,and segment at-risk populations.Their algorithms can also smooth out difficult terrain,understand diverse flood types,capture human interventions,and see through clouds.These advances not only improve government response and insurer risk,but also help to save lives

251、.Applications(Source:Cloud to Street)THE“NEW”SPATIAL WORLD45The GEOINT Playbook|Report by Space Capital Artutro,a Chicago based company,is a deep learning spin-out from American Family Insurance.The company provides physical property characteristic data and predictive analysis for residential and co

252、mmercial properties for use in the Property&Casualty(P&C)Insurance,Reinsurance,Lending,and Securities markets.By examining historical satellite imagery,claims and inspection data from a variety of insurance carriers,Arturo is able to derive unique temporal insights and assess risks by understanding

253、how areas are affected by the changing weather and human activity.On the processing side,Arturo does not require any pre-fetching or pre-processing of the satellite imagery.This means they are able to drastically reduce the time from data to insights and generate millions of property records on-dema

254、nd within days.Key variables such as roof material,number of stories,first floor elevation characteristics,proximity analysis of fire zones and other disasters risks,presence of new outbuildings are all illustrated in their web-based underwriting and claims solution platform and can be delivered thr

255、ough 100%configurable APIs.This allows insurers to access only the data they need to power their model or workflows.The flexibility and timeliness of this solution allows insurers to proactively monitor their entire property portfolio and understand why changes happen over time.Arturos AI property a

256、nalytics provides on-demand property characteristics insights from multi-source imagery via configurable APIs(Source:Arturo)ApplicationsIt is clear that geospatial intelligence plays an enabling role in transforming the insurance industry,from a primarily human-driven service to a data-driven produc

257、t.By focusing the power of geospatial data that satellites help collect,insurers are able to provide personalized policies,quantify real-time changes,and proactively manage risk.THE“NEW”SPATIAL WORLD46The GEOINT Playbook|Report by Space Capital ClimateLarge public and private businesses are increasi

258、ngly motivated by customers,competitors,and shareholders to establish an environmental governance strategy to monitor and offset their emissions.The value of traded carbon dioxide permits grew by 164%to a record$851 billion in 2021.71 In the past,many have doubted the viability of forest carbon proj

259、ects around their ability to permanently capture CO2 by trees and traditional measurements have mostly been completed manually by field plot networks every few years.72 With forest carbon projects expanding to the size of an entire city or small country,it has become increasingly challenging to prov

260、ide continuous and expansive oversight.Carbon credit purchasers are demanding better information to verify the legitimacy of a projects claims and performance over time.With transparency a key roadblock to fostering a successful voluntary carbon market,satellite-based monitoring is the most scalable

261、 solution to CO2 measurements that are independently verifiable as well as understandable to consumers.Carbon stock estimation models are becoming increasingly accurate with access to multispectral imagery.73 Consistent monitoring ensures that carbon sequestration takes place and any incident will b

262、e identified immediately.74 Using these geospatial datasets,carbon traders are able to track governments,companies,and projects that reduce emissions and target actors that excessively pollute or inaccurately report their footprint.This level of transparency and scrutiny is needed to determine accur

263、ate,daily carbon offset prices.75 The Climate Market Flywheel.(Source:Space Capital)ApplicationsClimate Market FlywheelTechnologyMarketsDataDirect,verifiable measurements of emissions sources and sinksRules and accountability mechanismsSpace-based and terrestrial sensors targeting key scientific sig

264、nalsTransparent quantifiable risk managementPolicy and RegulationTHE“NEW”SPATIAL WORLD47The GEOINT Playbook|Report by Space Capital Monitoring and addressing methane emissions can yield a near-term reduction in the rate of warming,easing the efforts to decarbonize the worlds energy and transport sys

265、tems while also delivering air quality benefits.76 Montreal-based GHGSat collects high precision emissions geospatial data at a fraction of the cost of existing alternatives from its satellite and aircraft sensors.It is the only organization in the world that is capable of capturing facility-level e

266、missions from orbit.The unique measurements are collected at short-wave infrared light(SWIR)wavelengths compared to standard imagery within the visible light wavelengths.This raw output is then post-processed to generate a surface reflectance in the SWIR and greenhouse gas concentrations per pixel i

267、n the atmosphere,and presented to the customers in“concentration maps”to visualize the methane distribution.This constant monitoring helps companies plan maintenance,maintain production,mitigate risk,optimize resources,and respond to emergencies.As the SEC and other regulatory agencies started to ma

268、ndate standardized climate risk disclosure,companies are required to provide reliable,validated climate data,assessment and risk aversion,and response tools.However,high quality climate data is very hard to find and only 13,000 companies are disclosing CO2 emissions globally according to CDP.77 The

269、status quo of global climate data is highly fragmented with a mix of government and private companies providing the majority of the data with poor standardization and source tracking,which makes it very challenging for companies or researchers to get the data they need to produce reliable climate un

270、derstanding and actions.dClimate,the first decentralized,fully transparent marketplace for climate data is working to become the solution.dClimate has created the infrastructure for immutable climate data to live on a decentralized network and,as a marketplace,it connects climate data providers like

271、 researchers,government agencies,academic institutions directly with the data consumers.This makes credible climate information easily accessible for businesses and application developers to find,access,and utilize.78 With dClimate,companies and other stakeholders can build customized tools to make

272、informed decisions in response to the rapidly changing climate on top of transparent sources,and the data can be shared so other organizations can benefit from the knowledge to improve their own assessments.The use of blockchain helps increase the transparency and auditability of their climate data.

273、The organization has partnered with the government of Namibia to provide their official verification for renewable hydrogen and other sustainability initiatives.79 dClimate set up a blockchain native registry and verification system to quantify the carbon sequestration,CO2 emissions,and the carbon c

274、redits generated from these initiatives.ApplicationsGHGSat visualizes the layer of methane emissions overlaid on ground imagery to show per-pixel methane concentration data,including surface reflectance image(Source:GHGSat)Beyond carbon markets,another use case powered by geospatial data is emission

275、s monitoring.THE“NEW”SPATIAL WORLD48The GEOINT Playbook|Report by Space Capital ImmersiveNiantics Pokmon Go is the most successful example of a mobile game that uses real world geospatial information.Niantics Pokmon Go mobile game has surpassed$6.0 billion in global lifetime player spending from the

276、 App Store and Google Play,according to market researcher Sensor Tower.81 It represented a powerful shift in entertainment technology,and a step closer to mixed reality.82 This means that unlike other location based applications that simply record a players location,applications like Pokmon Go can i

277、nfluence where a player goes.83 We are also seeing this trend from other companies like Snap,which is building their AR product around Custom Landmarkers(landmarks).The Landmarker Template allows AR creators and developers as well as small businesses to build anchored,location-based AR experiences a

278、t places they care about.Googles recent platform release also allows developers to anchor and overlay content by specific coordinates.There has been a shift towards AR development platforms,such as Niantic Lightship,which provides an AR development kit to empower creators to seamlessly and realistic

279、ally bring virtual content and experiences into the real world.In May 2022,Google announced the new ARCore Geospatial API at Google I/O 2022.The new Geospatial API gives app developers access to global localization for their AR applications,games,and experiences.Landmarker Guide.(Source:Snap)Applica

280、tionsAugmented Reality(AR)uses the physical world as its backdrop while overlaying information to generate real world applications.With AR,geospatial data is being presented in a new way to help users instantly understand the environment they are in.This differs from Virtual Reality(VR),which delive

281、rs a computer generated simulation of a virtual world presented to the user with a headset.During the past few years,companies like Snap,Facebook,and Apple have increased their investments in AR,encouraging the proliferation of AR usage on mobile devices.The market capitalization associated with AR

282、is projected to grow 1,000 x from$1.0 billion today to roughly$1.0 trillion by 2030.80 THE“NEW”SPATIAL WORLD49The GEOINT Playbook|Report by Space Capital Game engines,such as Unity and Epic Games Unreal Engines,are proving to be indispensable tools for building digital twins of the physical world.Th

283、ese game developer tools are used to generate photorealistic environments to replicate real-world behavior.The first public version of Epics Unreal Engine 5(UE5)was released in late 2021.In the past this process was very fragmented,the building of 3D models,animation,sounds,game logic,lighting,physi

284、cs and a dozen other aspects needed to be done in different programs.Over time,game engines have grown to encompass more and more of this process in one place and UE5 is the most capable and integrated engine yet,putting high-end graphics and Applications3D geospatial imagery composed with Cesium(So

285、urce:Venture Beat)visual design together with all the other critical components.Cesium,an open-source 3D geospatial company based in Philadelphia,has partnered with Epic Games to provide“Cesium for Unreal.The open-source developer plugin provides users the ability to discover,in real time,the locati

286、on of a player in a given 3D space,using accurate real-world 3D content captured from satellites,cameras,sensors,drones,and autonomous vehicles.It includes access to curated 3D content like global terrain,3D buildings,photogrammetry,and satellite imagery.The technology is helping Komatsu,a Japanese

287、company,build its Smart Construction app that monitors construction sites with precise,near real-time measurements to compare target and accomplished building progress.THE“NEW”SPATIAL WORLD50The GEOINT Playbook|Report by Space Capital echo3D makes it very easy for developers to manage 3D content at

288、scale(Source:echo3D)Through 3D scenes,Cognitive3D can provide users with a complete picture of exactly how human movements interact with the surrounding environment along with proper context(Source:Cognitive3D)ApplicationsAnother excellent example of a 3D developer tool is New York-based company ech

289、o3D.The company provides both content delivery and management services that are tailored for lightweight 3D development experiences.Since streaming quality is critical to seamless AR experiences,echo3D can help developers stream 3D content faster and more reliably to a great number of devices in rea

290、l-time across networks with different performance including 5G and Satcom.At the same time,developers can use echo3D to compress,convert,manage,and publish all of their 3D assets and leverage echo3Ds massive library to create new content.Analytics and user testing technologies have defined the core

291、approach behind mobile app development and optimization.The rapid growth of immersive 3D has led to the need for sophisticated spatial analytics tools.Vancouver-based company Cognitive3D has built a spatial intelligence platform that leverages all virtual,augmented,and mixed reality devices as a veh

292、icle for spatial data collection.Cognitive3D measures human interactions and performance across cognitive behavior,biometrics,and eye tracking.These insights are being used to drive research,refine product development and training,and improve customer experience.THE“NEW”SPATIAL WORLD51The GEOINT Pla

293、ybook|Report by Space Capital The world of mapping has changed dramatically over the past twenty years and we now find ourselves in a position where we can clearly see how these technical innovations laid the groundwork for future use cases and product development.As geospatial data becomes immersiv

294、e and ever-present through advances in gaming engines and augmented reality,we expect that interaction with our maps will respond to our actions and the changing environment around us.New breakthroughs in deep learning are required to eliminate human intervention on the data side and create more pre

295、dictive intelligence for the end user.This next evolution towards semantic(self-aware)mapping learns about the physical world around it then offers users recommendations based on that nuanced understanding of the geography.While maps have always been subjective,semantic maps may offer a more objecti

296、ve view of the world.Today,geographers are concerned about the spread of fake,AI-generated satellite imagery.84 By moving away from humans-in-the-loop and leveraging a diversified sensor ecology,created and managed by multiple stakeholders,we may be better equipped to triangulate on an objective tru

297、th about our environment.We envision a future where maps teach us phenomena about the world.This is the grand vision of IoT or smart cities-sensors constantly updating a map with physical inputs,which results in proactive responses vs reactive monitoring.Mapping the FutureApplicationsMAPPING THE FUT

298、URE52The GEOINT Playbook|Report by Space Capital Throughout history,maps have provided a powerful tool to understand our physical world and the information layered within has swayed the balance of wars and allowed companies to succeed where others failed.Satellites transformed modern mapping from st

299、atic approximations into dynamic reflections of our world and the activity within it.In the early 2010s,the use of distributed networks of small satellites made it more cost effective to capture timely geospatial data at a global scale.The adoption of cloud,edge computing,AI/ML capabilities,and incr

300、easingly powerful geospatial APIs and SDKs are making the benefits of geospatial intelligence more accessible.Developers no longer need to be experts in image capture,data processing,or object detection,and instead can focus on building specialized applications tailored to unique customers,similar t

301、o the proliferation of GPS brought about by Apples App Store.Demand for geospatial data is being quantified through marketplaces and used to inform what sensors and platforms will be built in the future.New breakthroughs in deep learning could eliminate human intervention entirely,unlocking more pre

302、dictive intelligence that appears self aware.We are beginning to see a future where every government,industry,and consumer can benefit from real-time intelligence about the world around them.The ability to collect,process,and analyze endless geospatial data is creating powerful new applications that

303、 will reshape our industries and our planet.In the same way that every company today is a technology company,the integration of GEOINT into existing workstreams will see every company of tomorrow be a space company.Conclusion53The GEOINT Playbook|Report by Space Capital Primary ResearchAnastasia Vol

304、kovaCEO of RegrowAugust 13,2019Brian FurtawSenior Data Scientist at NVIDIA,specializing in GPU hardware and software solutions for ML/AI programsMarch 21,2022Chris AndrewsCOO and Head of Product of Rendered.ai,Former Group Product Manager of Geo-enabled Systems,Smart Cities and 3D at EsriMarch 21,20

305、22Diego Saez-Gil CEO of PachamaJune 24,2021Primary ResearchThree additional interviews were conducted with executives and researchers that wish to remain anonymous.Geoffrey Levene Director of Worldwide AI Initiatives and Global Public Sector Defense,Climate Action and Space at NVIDIA,Former Director

306、 of Global Business Development and Data ScienceMarch 18,2022Gregg WalshBusiness Development Manager at MicrosoftApril 7,2021James SlifierzCEO of SkyWatchSeptember 2,2022John P.WilsonProfessor of Sociology and Spatial Sciences and Founding Director at University of Southern CaliforniaMarch 21,2022Na

307、than KundtzCEO of Rendered.AIMarch 21,2022Peter Marquez Head of Space Policy at Amazon AWSApril 25,2022Ronald BisioSenior Vice President and Executive Team Member at TrimbleApril 25,2022Siddhartha JhaCEO of Arbol,Founding Partner of dClimateMarch 11,2022Sina KashukSVP of Data Science at Foursquare,F

308、ormer CEO of UnfoldedApril 12,202254The GEOINT Playbook|Report by Space Capital Endnotes1The National Geospatial Intelligence Agency(NGA),a combat support agency within the United States Department of Defense,whose primary mission is collecting,analyzing,and distributing geospatial intelligence,defi

309、nes geospatial intelligence(GeoInt)as“the intelligence about human activity on Earth derived from the exploitation and analysis of imagery,signals,or signatures with geospatial information,”while geospatial information is data and information having an implicit or explicit association with a locatio

310、n relative to Earth(a geographic location or geographic position).2Fortune Business Insights.“Geospatial Analytics Market Size,Growth|Global Report 2028.”Fortune Business Insights,Fortune Business Insights,2021,https:/ 9 September 2022.3Guy,Jean-Sebastien.“The Changing Geospatial Market Size Adds Co

311、mplexity.”Korem,18 March 2021,https:/ 9 September 2022.4Maney,Kevin.“Tiny tech company awes viewers.”USA Today,USA Today,21 March 2003,https:/ 14 September 2022.5“Google Acquires Keyhole.”Wall Street Journal,27 October 2004.Accessed 9 September 2022.6Goddard,Chelsea.“Insights|Mapping Technology and

312、the Google Effect.”Space Capital,11 June 2021,https:/ 9 September 2022.7The National Academies Press.“Assessment of Impediments to Interagency Collaboration on Space and Earth Science Missions.”The National Academies Press,2011,https:/www.nap.edu/read/13042/chapter/4#17.Accessed 9 September 2022.8St

313、raub,Crista L.,et al.“Economic Valuation of Landsat Imagery.”Open-File Report,2019,doi:10.3133/ofr20191112.https:/landsat.gsfc.nasa.gov/article/usgs-report-updateseconomic-value-landsat-archive.9“How many Earth observation satellites are orbiting the planet in 2021?”Pixalytics Ltd,18 August 2021,htt

314、ps:/ 9 September 2022.10Straits Research.“Satellite Earth Observation Market Size,Report,Forecast to 2030.”Straits Research,19 July 2022,https:/ 9 September 2022.11Miller,Alex.“How big is that satellite?A primer on satellite categories.”Viasat,27 April 2021,https:/ 9 September 2022.12Cobb,Wendy Whit

315、man.“How SpaceX lowered costs and reduced barriers to space.”The Conversation,1 March 2019,https:/ 9 September 2022.13Cobb,Wendy Whitman.“How SpaceX lowered costs and reduced barriers to space.”The Conversation,1 March 2019,https:/ 9 September 2022.14Duffy,Kate.“Elon Musk Says SpaceX Starship Flight

316、s to Cost Under$10M in 2-3 Years.”Business Insider,11 February 2022,https:/ 9 September 2022.15Sheetz,Michael.“Elon Musk:SpaceX Falcon Heavy costs$150 million at most.”CNBC,12 February 2018,https:/ 9 September 2022.16Zafar,Ramish.“Elon Musk Reiterates Insanely Low Starship Launch Costs Of$10/kg.”Wcc

317、ftech,https:/ 9 September 2020.17Estevez,Eric.“Graphics Processing Unit(GPU).”Investopedia,2021,https:/ 9 September 2022.18Google.“Cloud Tensor Processing Units(TPUs).”Google Cloud,2022,https:/ 9 September 2022.19Temple,James.“Everything You Need to Know About Skybox,Googles Big Satellite Play.”Vox,

318、11 June 2014,https:/ 9 September 2022.20Research and Markets.“High Altitude Platforms-Global Market Trajectory&Analytics.”Research and Markets,July 2022,https:/ 9 September 2022.55The GEOINT Playbook|Report by Space Capital Endnotes21Global Market Insights.“High Altitude Platform Market Size By Prod

319、uct(UAV,Airships,Tethered Aerostat Systems),By Application(Surveillance,Navigation&Remote Sensing,Communication,EO/IR System),By End-use(Government&Defense,Commercial),Industry Analysis Report,Regi.”Global Market Insights,Global Market Insights,Pct 2017,https:/ 09 September 2022.22Morgan,Jessica L.,

320、et al.“Aerial Photography:A Rapidly Evolving Tool for Ecological Management.”BioScience,vol.60,no.1,2010,pp.47-59.BioScience,https:/ 9 September 2022.23“DATA.AIR.”GHGSat,GHGSat,2022,https:/ 9 September 2022.24IMARC Group.“Aerial Imaging Market Size,Share,Trends and Growth 2022-2027.”IMARC Group,IMAR

321、C Group,2022,https:/ 9 September 2022.25Statista.“Drones-Worldwide.”Statista,Statista,2022,https:/ 9 September 2022.26“How To Use DroneDeploy For Crop Insurance.”DroneDeploy,4 March 2020,https:/ 9 September 2022.27Spoors,Gavin.“Drone Regulations:Everything you need to know.”S,15 October 2021,https:/

322、 9 September 2022.28Ware.“Welcome to the warehouse of the future.”Ware|Drones For Warehouse Automation,2022,https:/www.ware.ai.Accessed 9 September 2022.29360 Research Reports.“Global Unattended Ground Sensor Market Outlook 2022.”360 Research Reports,360 Research Reports,4 January 2022,https:/ 9 Sep

323、tember 2022.30Verified Market Research.“Smartphone Sensors Market Size,Share,Opportunities&Forecast.”Verified Market Research,Verified Market Research,August 2022,https:/ 9 September 2022.31“Leaf Space enables low Earth satellites with Google Cloud.”Google Cloud,9 April 2021,https:/ 9 September 2022

324、.32Microsoft.“Azure Orbital Ground Station Satellite Comms.”Microsoft Azure,2022,https:/ 9 September 2022.33Novet,Jordan.“Google wins cloud deal from Elon Musks SpaceX for Starlink internet connectivity.”CNBC-TECH,CNBC,13 May 2021,https:/ 14 September 2022.34Hazelcast.“Edge Computing Overview:IoT Pr

325、ocessing at the Edge.”Hazelcast,2022,https:/ 9 September 2022.35Palantir.“Edge AI in Space|Palantir and Satellogic.”Palantir Blog,1 April 2022,https:/ 9 September 2022.36Safegraph.“Challenges of Geospatial Data Integrations.”SafeGraph,2022,https:/ 9 September 2022.37Johnson,Mark.“Meditations:A Requi

326、em for Descartes Labs|by Mark Johnson|Aug,2022|Medium.”Mark Johnson,4 August 2022,https:/ 9 September 2022.38“Combination analysis of multispectral and radar satellite data.”Digitala Vetenskapliga Arkivet,DiVA,2021,http:/www.diva-portal.org/smash/record.jsf.Accessed 9 September 2022.39Planet Labs.“P

327、lanet Fusion Monitoring.”Planet Labs,Planet Labs,March 2021,https:/ 9 September 2022.40Regalado,Antonio.“Who Coined Cloud Computing?”MIT Technology Review,31 October 2011,https:/ 9 September 2022.56The GEOINT Playbook|Report by Space Capital Endnotes41Foote,Keith D.“A Brief History of Cloud Computin

328、g-DATAVERSITY.”Dataversity,17 December 2021,https:/ 9 September 2022.42Vailshery,Lionel Sujay.“Amazon Web Services revenue 2021.”Statista,2021,https:/ 9 September 2022.43Shirman,Johanna.“Top 4 Reasons for Slow Cloud Adoption and How Theyre Being Addressed.”RiverMeadow,22 October 2014,https:/ 9 Septe

329、mber 2022.44This has since provided the DigitalGlobe customer base with on-demand access to data.It has also enabled the DigitalGlobe team to train and deploy new caching and product usage algorithms,which have directly supported their efforts to conserve cloud costs and identify new patterns in acc

330、ess requests to specific image tiles amongst its customer base.45NSR.“Satellite Ground Segment:Moving to the Cloud.”NSR,NSR,2021,https:/ 9 September 2022.46IBM.“IBM Cloud Satellite Enables Clients to Deliver Cloud Securely in Any Environment Including at the Edge.”IBM Newsroom,IBM,1 March 2021,https

331、:/ 9 September 2022.47Geospatial artificial intelligence(GeoAI),an emerging scientific discipline,combines innovations in spatial science,artificial intelligence methods in machine learning(e.g.,deep learning),data mining,and high-performance computing to extract knowledge from spatial big data.48Go

332、ddard,Chelsea.“Insights|Mapping Technology and the Google Effect.”Space Capital,11 June 2021,https:/ 9 September 2022.49OpenAI.“DALLE:Creating Images from Text.”OpenAI,5 January 2021,https:/ 9 September 2022.50“Gartner Says Nearly Half of CIOs Are Planning to Deploy Artificial Intelligence.”Newsroom

333、,Gartner,13 February 2018,https:/ 9 September 2022.51Gartner Research.“Maverick*Research:Forget About Your Real Data Synthetic Data Is the Future of AI.”Gartner,Gartner,24 June 2021,https:/ 9 September 2022.52Gartner Research.“Maverick*Research:Forget About Your Real Data Synthetic Data Is the Future of AI.”Gartner,Gartner,24 June 2021,https:/ 9 September 2022.53OneView.“Combining Synthetic Data w