AI-RAN and Open RAN<\/p>O-RAN reference architecture<\/h3>The split of the RAN into O-RU, O-DU, O-CU, RICs, SMO and O-Cloud with defined open interfaces between them.<\/p>We design and integrate.<\/p>O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>O-RU (radio unit)<\/h3>The radio and low-PHY at the antenna, category A or B depending on where precoding runs.<\/p>We integrate, deploy and maintain.<\/p>O-RAN WG4 Control, User and Synchronization Plane Specification \u00b7 O-RAN WG7 white-box hardware reference designs<\/p><\/article>AI-RAN and Open RAN<\/p>O-DU (distributed unit)<\/h3>High-PHY, MAC and RLC running on servers at the cell site or hub, often with acceleration.<\/p>We integrate, deploy and operate.<\/p>O-RAN WG8 stack reference design \u00b7 Small Cell Forum SCF222 (5G FAPI)<\/p><\/article>AI-RAN and Open RAN<\/p>O-CU (central unit, CP and UP)<\/h3>RRC, PDCP and SDAP functions split into control and user planes and run centrally.<\/p>We integrate, deploy and operate.<\/p>3GPP TS 38.401 (NG-RAN architecture) \u00b7 3GPP TS 38.463 (E1AP) \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Open F1, E1, X2 and Xn interoperability<\/h3>Interoperability profiles so CU, DU and neighbouring nodes from different suppliers work together.<\/p>We integrate.<\/p>O-RAN WG5 open F1\/W1\/E1\/X2\/Xn interface specifications \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Shared O-RU and multi-operator Open RAN<\/h3>One radio unit serving several operators’ DUs, cutting site equipment in shared deployments.<\/p>We design, integrate and deploy.<\/p>O-RAN WG4 shared O-RU specifications \u00b7 O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI-RAN and Open RAN<\/p>O-RU (radio unit)<\/h3>The radio and low-PHY at the antenna, category A or B depending on where precoding runs.<\/p>We integrate, deploy and maintain.<\/p>O-RAN WG4 Control, User and Synchronization Plane Specification \u00b7 O-RAN WG7 white-box hardware reference designs<\/p><\/article>AI-RAN and Open RAN<\/p>O-DU (distributed unit)<\/h3>High-PHY, MAC and RLC running on servers at the cell site or hub, often with acceleration.<\/p>We integrate, deploy and operate.<\/p>O-RAN WG8 stack reference design \u00b7 Small Cell Forum SCF222 (5G FAPI)<\/p><\/article>AI-RAN and Open RAN<\/p>O-CU (central unit, CP and UP)<\/h3>RRC, PDCP and SDAP functions split into control and user planes and run centrally.<\/p>We integrate, deploy and operate.<\/p>3GPP TS 38.401 (NG-RAN architecture) \u00b7 3GPP TS 38.463 (E1AP) \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Open F1, E1, X2 and Xn interoperability<\/h3>Interoperability profiles so CU, DU and neighbouring nodes from different suppliers work together.<\/p>We integrate.<\/p>O-RAN WG5 open F1\/W1\/E1\/X2\/Xn interface specifications \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Shared O-RU and multi-operator Open RAN<\/h3>One radio unit serving several operators’ DUs, cutting site equipment in shared deployments.<\/p>We design, integrate and deploy.<\/p>O-RAN WG4 shared O-RU specifications \u00b7 O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI-RAN and Open RAN<\/p>O-DU (distributed unit)<\/h3>High-PHY, MAC and RLC running on servers at the cell site or hub, often with acceleration.<\/p>We integrate, deploy and operate.<\/p>O-RAN WG8 stack reference design \u00b7 Small Cell Forum SCF222 (5G FAPI)<\/p><\/article>AI-RAN and Open RAN<\/p>O-CU (central unit, CP and UP)<\/h3>RRC, PDCP and SDAP functions split into control and user planes and run centrally.<\/p>We integrate, deploy and operate.<\/p>3GPP TS 38.401 (NG-RAN architecture) \u00b7 3GPP TS 38.463 (E1AP) \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Open F1, E1, X2 and Xn interoperability<\/h3>Interoperability profiles so CU, DU and neighbouring nodes from different suppliers work together.<\/p>We integrate.<\/p>O-RAN WG5 open F1\/W1\/E1\/X2\/Xn interface specifications \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Shared O-RU and multi-operator Open RAN<\/h3>One radio unit serving several operators’ DUs, cutting site equipment in shared deployments.<\/p>We design, integrate and deploy.<\/p>O-RAN WG4 shared O-RU specifications \u00b7 O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI-RAN and Open RAN<\/p>O-CU (central unit, CP and UP)<\/h3>RRC, PDCP and SDAP functions split into control and user planes and run centrally.<\/p>We integrate, deploy and operate.<\/p>3GPP TS 38.401 (NG-RAN architecture) \u00b7 3GPP TS 38.463 (E1AP) \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Open F1, E1, X2 and Xn interoperability<\/h3>Interoperability profiles so CU, DU and neighbouring nodes from different suppliers work together.<\/p>We integrate.<\/p>O-RAN WG5 open F1\/W1\/E1\/X2\/Xn interface specifications \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Shared O-RU and multi-operator Open RAN<\/h3>One radio unit serving several operators’ DUs, cutting site equipment in shared deployments.<\/p>We design, integrate and deploy.<\/p>O-RAN WG4 shared O-RU specifications \u00b7 O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI-RAN and Open RAN<\/p>Open F1, E1, X2 and Xn interoperability<\/h3>Interoperability profiles so CU, DU and neighbouring nodes from different suppliers work together.<\/p>We integrate.<\/p>O-RAN WG5 open F1\/W1\/E1\/X2\/Xn interface specifications \u00b7 3GPP TS 38.473 (F1AP)<\/p><\/article>AI-RAN and Open RAN<\/p>Shared O-RU and multi-operator Open RAN<\/h3>One radio unit serving several operators’ DUs, cutting site equipment in shared deployments.<\/p>We design, integrate and deploy.<\/p>O-RAN WG4 shared O-RU specifications \u00b7 O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI-RAN and Open RAN<\/p>Shared O-RU and multi-operator Open RAN<\/h3>One radio unit serving several operators’ DUs, cutting site equipment in shared deployments.<\/p>We design, integrate and deploy.<\/p>O-RAN WG4 shared O-RU specifications \u00b7 O-RAN WG1 Architecture Description<\/p><\/article>AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI-RAN and Open RAN<\/p>Deployment scenarios and white-box hardware<\/h3>Reference placements of RAN functions (site, hub, regional) and open hardware designs for each.<\/p>We design and integrate.<\/p>O-RAN WG7 deployment scenarios and base station hardware architecture<\/p><\/article>AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI systems<\/p>EO foundation models<\/h3>Large pretrained models over optical, SAR and multi-temporal imagery, fine-tuned for many downstream tasks.<\/p>We design, integrate and operate.<\/p>OGC Training Data Markup Language for AI (TrainingDML-AI) \u00b7 CEOS Analysis Ready Data framework<\/p><\/article>AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>
AI systems<\/p>Training data and labels<\/h3>Documented, licensed training sets with known label quality and geographic coverage.<\/p>We design and operate.<\/p>OGC TrainingDML-AI \u00b7 ISO 19157-1 (data quality)<\/p><\/article> <\/section> 02<\/b>Architecture<\/p> Where it sits, and what it speaks.<\/h2> Benchmarks draws on 4 domains and 32 layers of the spectrum; the highlighted layers carry this area. It speaks 3 of the protocol’s 8 core functions.<\/p> Edge inference, model serving and optimisation<\/p>10<\/p> Foundation models for Earth observation and geospatial<\/p>9<\/p> Time-series and sensor AI<\/p>8<\/p> Multimodal fusion and agentic operations<\/p>7<\/p> Federated and privacy-preserving learning<\/p>7<\/p> Physical AI and robotics (within red lines)<\/p>7<\/p> MLOps<\/p>7<\/p> Evaluation, model risk and AI governance<\/p>10<\/p> O-RAN architecture and open interfaces<\/p>7<\/p> Fronthaul (7.2x split) and transport<\/p>8<\/p> RAN intelligent controllers, xApps and rApps<\/p>8<\/p> SMO, O-Cloud and accelerated L1<\/p>8<\/p> AI-RAN: shared accelerated infrastructure<\/p>7<\/p> AI\/ML in the air interface and network management<\/p>8<\/p> Energy saving and efficiency<\/p>6<\/p> Testing, integration and security<\/p>8<\/p> 01Join<\/b>Identity, attestation quote, signed policy.<\/p>02Admit<\/b>Admitted for a stated scope, nothing wider.<\/p>03Describe<\/b>What, where, how and by whom, with calibration.<\/p>04Time<\/b>One clock across the grid, with its error bound.<\/p>05Observe<\/b>Every value with quality and provenance.<\/p>06Qualify<\/b>Tested against the specification, state on record.<\/p>07Federate<\/b>Signed exchange across grids and trust domains.<\/p>08Orchestrate<\/b>Software and configuration by signed plan.<\/p> <\/section> 03<\/b>On the grid<\/p> Every site joins one observing grid.<\/h2>The open networks on the grid today, over real relief and the geodesic cells we count them into, filtered to the space domain this area serves. A site built here enrols into the same grid.<\/p> <\/section> 04<\/b>In the field<\/p> Where this runs, sector by sector.<\/h2> 3 deployment scenarios from the sector research that rely on benchmarks.<\/p> Scientific Research-infrastructure gap-filling nodes<\/h3>Observatory-as-a-service for research infrastructures that must extend coverage under their own quality rules.<\/a>Transport Non-federal automated weather observing for small airports<\/h3>Aviation-grade stations under the civil-aviation authority’s advisory circular; pricing not researched.<\/a>Defence-adjacent (civil dual-use) Research and development route through the defence science programme<\/h3>A science-and-technology programme on all-domain situational awareness.<\/a> <\/section> 05<\/b>Open and enterprise<\/p> Open by default. Enterprise by contract.<\/h2> The open counterpart lives on the main site for anyone to read and run. The enterprise service is designed, deployed and operated with you, under contract.<\/p> Open, on the main site<\/p>Benchmark methods<\/li>Pilot reports, published either way<\/li><\/ul>Enterprise, on this site<\/p>Research pilots<\/li>Prototype engineering<\/li><\/ul> <\/section> 06<\/b>The rest of the line<\/p> Frontier labs, area by area.<\/h2> Frontier labs Pilots<\/h3>New technology, a real site, a written plan.<\/a>Frontier labs AI-RAN and 6G<\/h3>Sensing and communication on shared compute.<\/a>Frontier labs Space and NTN<\/h3>Direct-to-device and new payloads, tested.<\/a>Frontier labs Physical AI<\/h3>Drones and robots, inside published red lines.<\/a>Frontier labs Into the product<\/h3>From a result to a qualified release.<\/a> <\/section> Bring us the site.<\/h2>Survey, design, deployment and operation for benchmarks. The form names who receives your details.<\/p> Talk to us<\/a>contact@observatory.global<\/a> <\/section> <\/main>