GoSmart

**Product and the concrete problem it solves.** GoSmart addresses a costly information gap in aquaculture: farmers often decide how much to feed, when to harvest, and whether a stock is healthy using periodic sampling, experience, and visual guesswork, even though the fish are underwater and conditions change continuously. That gap creates three linked losses: feed is wasted when fish are not eating, water quality deteriorates when excess organic matter accumulates, and disease or stress can remain invisible until mortality or growth loss is already material. GoSmart's product is an integrated precision-farming system built around underwater cameras, environmental sensors, a local processing unit, and a cloud application. The current official site describes a Bio Unit that estimates fish weight and population distribution while measuring oxygen and temperature, plus a Feeding Unit that connects a camera to a feeding boat or operations center and recommends daily feed amounts. The farmer receives real-time and historical views of biomass, feeding, dissolved oxygen, temperature, and operational trends rather than a single manual sample. The practical promise is better feed-conversion performance, earlier intervention, lower labor intensity, and more predictable harvest planning. The company reports 95% detection accuracy and 10% feed reduction on its site, but those are company-stated figures whose species, baseline, sample size, and independent validation should be requested in diligence.

**Core technology and how it actually works.** GoSmart's system combines computer vision, embedded sensing, energy management, remote communications, and a SaaS decision layer. The underwater camera captures fish imagery in cages, ponds, or tanks; the biomass algorithm uses those observations to estimate individual or average weight and the distribution of fish across the farmed population. Oxygen and temperature sensors add the environmental context needed to distinguish a feeding problem from a water-quality problem. The company says its algorithms can analyze multiple characteristics in parallel, and public reporting describes initial training with manually labeled data covering thousands of fish. Its engineering choices are directed at the hardest operating environments rather than at a laboratory demo. The 2023 technical account describes an NVIDIA Jetson edge-compute module, lithium-ion batteries charged by solar panels, and power-management software that can sleep, wake, and process data autonomously. The more recent company site describes a fully autonomous, solar-powered Feeding Unit that can operate in all weather and be controlled remotely through a web app. The Fish Site reports that early hardware measured biomass distribution, created growth curves, and recorded oxygen and temperature in ten-second intervals, while the next version adds rugged aluminum housing and a second behavior-analysis camera. The planned behavior and appetite model is important because weight alone cannot determine whether more feed is useful: changes in group movement and response during feeding can identify the point at which additional feed becomes waste. GoSmart is also developing health monitoring with local universities, using swimming patterns and group behavior to flag early disease signals. These capabilities are technically credible as a sensor-fusion and applied-vision stack, but the public record does not establish accuracy by species, performance in turbid water, uptime, cybersecurity controls, or interoperability with every farm system.

**Market, customers, and go-to-market.** The initial market is commercial aquaculture, especially marine farms producing seabass, seabream, salmon, trout, meagre, and related species where feed is a major operating cost and offshore or remote sites make continuous human observation difficult. GoSmart's target buyer is a farm operator, producer group, feed company, or aquaculture integrator that wants better feed conversion and stock visibility without installing a large fixed infrastructure project. Its hardware is deliberately compact and off-grid, which lets the company sell into cages far from shore as well as ponds and recirculating systems. The likely business model is recurring software and monitoring revenue attached to affordable hardware: CEO Josef Melchner told The Fish Site that farmers pay monthly, broadly from the savings created by reduced labor and optimized feed. The company's commercial route also includes strategic distribution and data partnerships. Skretting, a major aquafeed supplier, worked with GoSmart for more than three years on Skretting 360+, a package combining nutrition, farm-management support, underwater cameras, and IoT sensors. Skretting's own Spanish publication says the resulting service monitored biomass and environmental parameters and calculated more precise feeding protocols for Mediterranean producers. GoSmart's official site lists deployments or operating presence in Turkey, Spain, Croatia, Chile, Italy, Malta, and Albania, while The Fish Site reports deployment in nine countries and names work with one of Chile's large salmon producers and Eden Farm's arid Israeli salmon operation. These are meaningful market signals, but the public material does not disclose customer contract values, recurring revenue, unit shipments, renewal rates, or the proportion of deployments that are paid production rather than trials.

**Traction, funding, and third-party validation.** GoSmart has stronger evidence of real-world use than a concept-stage aquaculture startup, but its financing and scale remain unusually opaque. Startup Nation Finder lists the company as founded in 2020 in Kfar Saba, with six employees, a released product, an undisclosed 2023 non-equity round involving Horizon Europe and the Israel Innovation Authority, and an undisclosed April 2024 SAFE associated with OurCrowd. An Israel Innovation Authority company profile lists GoSmart Precision Farming Ltd. as an Israeli agritech company established in 2020 with seven employees and an initial-revenues stage, and describes the integrated biomass camera, oxygen and temperature sensors, real-time feeding connection, and fish-population analysis. The Fish Site separately reports that a 2017 European Commission SME Instrument application provided €2.35 million for an early prototype; because that funding predates the company's 2020 public founding record, it should be treated as prototype or project funding until the corporate relationship is clarified rather than counted automatically as company equity. External validation includes the 2023 SeafoodSource report on GoSmart's use of NVIDIA Jetson and its technology being used by Skretting to create feeding regimes, the Skretting partner account of commercial-environment evaluations, the company's inclusion in HiCenter's BlueTech portfolio, and a 2026 Israeli government summit agenda that lists Melchner speaking on real-time underwater intelligence at a blue-food-security event. The evidence supports an active, productized startup with initial revenues and international deployments. It does not support claims of profitability, a large venture round, a patent portfolio, regulatory certification, or a large operating organization.

**Founders and team background.** GoSmart's founder-market fit is unusually direct. The Fish Site describes Josef Melchner as CEO and co-founder with more than two decades in aquaculture. He joined Gili Ocean Technology in 2004, progressed from aquaculture technician to CEO, and worked on offshore projects including a farm 15 kilometers offshore in 82 meters of water that reportedly withstood waves above 11 meters without damage or losses. His decision to start GoSmart followed a specific operating observation: farmers were increasing feed portions based on estimates of fish size because affordable, continuous biomass data was unavailable. That background is relevant because the product is designed around farm workflow, not merely around an off-the-shelf camera. The official team page lists Tamir Anavi as head of innovation; Karin Karter Gavriel as operations and production manager; Yoav Zilbertzan as algorithm engineer; Ebraheem Saleh as technical lead for computer vision and system integrations; Netanel Malihi in cloud engineering; Menachem Zeigfinger in R&D and systems engineering; Doron Yechskel in embedded software; and Yehonatan Levy in customer success for precision aquaculture. The board listing includes Nir Gilad as chairman and Zeev Kalimi in affairs and commercial strategy. This is a compact interdisciplinary team covering domain operations, algorithms, embedded systems, cloud, production, and customer success. It is also a concentration risk: the public record does not establish the depth of the bench, the founders' equity or retention, the size of the current engineering group, or whether the listed names are full-time employees, contractors, advisers, or board members.

**Competitive dynamics.** GoSmart competes with a mix of precision-aquaculture specialists, farm-management software, equipment vendors, and manual operating practice. Observe Technologies competes directly on AI-assisted feeding optimization from underwater observations; Aquabyte applies computer vision and machine learning to fish biomass, welfare, and farm decisions; Innovasea combines aquaculture monitoring, sensors, feeding, and farm-management infrastructure; AKVA group supplies established cage, feeding, camera, and farm-management systems with strong industry distribution; and Manolin provides seafood-production intelligence and farm data workflows. Traditional sampling, sonar, underwater inspection, feed-company advisory services, and farm-specific spreadsheets remain powerful substitutes because they require less new hardware. GoSmart's claimed edge is not a new general-purpose AI model. It is the combination of an affordable all-in-one underwater unit, solar-powered off-grid operation, integrated oxygen and temperature data, a remote web interface, species-specific models, and a subscription model designed for producers that cannot justify expensive systems built for the largest salmon farms. The company's experience in Mediterranean and lower-margin farming environments may also be a useful wedge against systems optimized for Norway's highly instrumented salmon industry. That advantage will be durable only if the company proves that the camera and algorithms remain reliable across species, water visibility, cage depth, weather, and feeding methods; that customers act on recommendations; and that the savings exceed installation, communications, maintenance, and subscription costs. Larger competitors can bundle hardware and software, while feed suppliers can internalize analytics through existing customer relationships.

**Defense, security, and resilience dual-use relevance.** GoSmart's dual-use case is a food-security and infrastructure-resilience adjacency, not a demonstrated military capability. Aquaculture is a strategic source of protein that can reduce pressure on wild fisheries, but offshore farms and arid inland farms are vulnerable to disease, oxygen depletion, weather, power loss, labor shortages, and disrupted logistics. A solar-powered, remotely managed monitoring layer that detects biomass, feeding behavior, temperature, oxygen, and abnormal movement can help a distributed food-production network continue operating when shore infrastructure or routine inspection access is limited. The Israeli Eden Farm example is relevant because it adapts aquaculture to extreme heat and uses the country's water-treatment and desalination expertise; the same remote sensing and decision-support principles could support protected food production in water-stressed or crisis-affected regions. Potential strategic deployments include remote aquaculture farms supplying local communities, emergency food-production sites, autonomous monitoring of defense or civil-resilience farms, and government food-security programs that need auditable stock and environmental data. The core technology also has a limited security-adjacent transfer path to environmental monitoring, but no public source identifies a defense customer, Ministry of Defense program, classified deployment, military contract, or security certification. Any future defense use would require hardened communications, access control, supply-chain assurance, cyber testing, and validation under degraded connectivity. The appropriate score therefore reflects real resilience value and dual commercial/strategic applicability, while explicitly avoiding the inference that GoSmart is already a defense supplier.

**Growth stage, trajectory, and key diligence risks.** GoSmart is best classified as mid-stage on product and field evidence, but early in organizational and financing scale. It has a 2020 company record, a released product, an initial-revenues classification, a small named team, deployments across multiple countries, an established relationship with Skretting, and customer or partner references in Chile and Israel. The next growth step is to convert those deployments into repeatable, standardized production packages for more species and farm types. The company says its second-generation hardware is moving toward trials at three farms and that it is raising $3 million to $5 million for production, international deployment, and additional species and health algorithms; the amount is a current fundraising target reported by The Fish Site, not closed capital. The diligence priorities are concrete: verify paid versus pilot deployments, monthly recurring revenue, gross margin after field support, hardware failure and replacement rates, sensor calibration, communications costs, customer retention, and the independent basis for feed-reduction and detection claims. Technical diligence should test models in turbid water, different cage depths, night and storm conditions, species transfer, false disease alerts, data labeling, and cybersecurity of remote controls. Commercial diligence should examine dependence on Skretting, the Chilean customer concentration, the economics of selling into lower-margin farms, and competition from integrated feed and equipment suppliers. Funding risk is material because marine hardware and international support consume capital, while the public financing record is undisclosed and the company is reportedly raising. If GoSmart can turn domain expertise and a useful data flywheel into reliable autonomous feeding decisions, it could become a strategic monitoring layer for resilient blue-food production; if not, it may remain a technically promising but service-heavy equipment business.