At 685 meters, water pressure is roughly 70 bar. MOBY Robotics landed its Mantis crawler on the seafloor at that depth on 23 October, kept it there through a three-day offshore campaign, and brought it back. The company's own account puts the jump from pool testing to 685 meters at 72 hours.
The robot worked. The business case is still unproven.
Deep-sea mining has spent a decade caught between engineering promises and permitting fights. MOBY's answer is architectural. Replace the industry's single giant collector and vertical riser with a fleet of small autonomous crawlers, a skip-lift, and an autonomously operated surface platform. The 685-meter test is the first public evidence that those pieces can run together in salt water.
It is also the easy part. A machine that survives 70 bar is a procurement decision. A machine that can legally remove nodules from the Clarion-Clipperton Zone, show that it did not wreck the habitat, and sell the metal into a regulated supply chain is a decade-long institutional project.
Two architectures, one cost curve
The incumbent design is an industrial site moved underwater: one large collection vehicle, a vertical riser, and a production vessel holding station above it. MOBY inverts the stack. Many small crawlers work the seabed, a SmartLift carries material upward, and a dynamically positioned platform handles 25-ton payloads at the surface before a transport vessel takes them ashore.
| Layer | Conventional riser system | MOBY swarm system |
|---|---|---|
| Collection | One large tracked vehicle | Multiple autonomous crawlers |
| Vertical lift | Riser and pump | SmartLift skip system |
| Surface interface | Production vessel | 25-ton Mineral Production Platform |
| Capital profile | Large fixed commitment | Modular, added by campaign |
| Failure mode | One blockage stops the site | One crawler lost, fleet continues |
System comparison from MOBY Robotics disclosures and Offshore Engineer reporting, 2026.
The pitch is that smaller units lower the entry hurdle. Andrew Lipman, MOBY's chief operating officer, told Offshore Engineer that the system has the lowest capital and operating requirements in the category. That claim has not been independently audited. What is visible is the shape of the bet: MOBY is selling optionality, not a mine.
We've actually taken the paradigm and turned it on its head. The robots themselves will figure out the optimization of collection while minimizing the environmental impacts.— Andrew Lipman, Chief Operating Officer, MOBY Robotics
Autonomy is doing double duty here. It is a cost story and an environmental story. A swarm that senses marine life and adjusts its path can claim a lighter footprint than a single heavy collector. The same claim is also the hardest one to prove, because it requires before-and-after data from sites that regulators have not yet opened at commercial scale.
Certification is the product
On 23 September, the American Bureau of Shipping issued an Approval in Principle for MOBY's commercial-scale Mineral Production Platform. The approved configuration is the 25-ton version: dynamically positioned, designed to stay on station year-round, and built to handle multiple payload transfers per day without returning to port between campaigns.
The distinction matters. An AiP confirms that a conceptual design is, in principle, compliant with ABS rules for novel offshore concepts, subsea mining, and dynamic positioning. It is not a classification certificate. It is the difference between a design that a classifier will review and a design that a bank can underwrite.
MOBY hired its way into that conversation. Lipman spent 17 years at ABS, most recently as director of subsea and mining operations. A startup that wants to sell capital equipment into a conservative offshore market needs someone who knows how the review queue works. MOBY has one.
Alex Petersen, the co-founder and chief executive, frames the approval as a construction gate. The platform, he said, gives the company a foundation to move from a validated design into construction and offshore deployment. That is the correct reading. AiP is the first signature on a long document.
The launch and recovery chain works in open water, not only in a test tank.
The crawler, the SmartLift, and the vision system can operate as one sequence.
ABS has a reviewable platform design to classify, which is a different asset from a working robot.
A December mapping job in the Clarion-Clipperton Zone
The commercial proof is scheduled for December. MOBY has a contract to run advanced seabed mapping in the Clarion-Clipperton Zone, the abyssal plain that holds most of the world's explored polymetallic nodule fields. The customer is confidential. The work is mapping, not extraction.
That distinction is the point. The campaign is meant to demonstrate a platform rated for depths of up to 6,000 meters and to feed high-resolution imagery into MOBY Maps, the company's AI-native ocean intelligence layer. The output is a 3D map, a searchable mission record, and a set of predictions about terrain, mineral distribution, and site conditions.
Every deployment adds to the same proprietary dataset. That is the flywheel MOBY is actually selling: each survey makes the next survey cheaper to interpret, and the interpretation becomes the product that mining companies, cable operators, and offshore developers buy before they buy a robot.
The company also sells the hardware as a turnkey system, from seabed collection through surface handling, before it moves into operating its own leased concessions. That sequencing keeps MOBY asset-light while it waits for the legal regime to settle.
Vietnam is the commercial hedge
On the same day as the ABS approval, MOBY signed a memorandum of understanding with Saigon Telecommunication and Technologies, acting through its SoiHub open innovation hub. The agreement was signed in New York during Vietnamese General Secretary and President To Lam's visit to the United States. It gives MOBY a route into Vietnam's Ministry of Science and Technology and into local industry partners.
The scope is broader than mining. The MOU covers oil and gas infrastructure, seaports, offshore wind, submarine cables, and deep seabed mapping, including an assessment of critical minerals and rare earth elements. MOBY leads on the technology and training. Saigontel and SoiHub supply sites, manufacturing capacity, and introductions.
For an investor, the MOU is a hedge against a single regulatory outcome. If nodule extraction stays blocked, MOBY can still sell inspection and survey capacity into offshore energy and cable markets. If extraction opens, the same vehicles and the same mapping layer are already qualified in a new jurisdiction.
The risk is that an MOU is not revenue. MOBY says it is already engaged in commercial opportunities with oil and gas companies in Vietnam. No contract value, customer name, or start date has been disclosed.
The gap between a demo and a duty cycle
685 meters is roughly 11% of the platform's rated 6,000-meter depth. The remaining 89% is not a linear scale-up. At full ocean depth, acoustic positioning degrades, launch and recovery windows narrow with sea state, and every connector becomes a maintenance item. A three-day campaign can be staffed by a concentrated engineering team. A year-round operation cannot.
The harder number is throughput. MOBY says the platform is designed to handle multiple 25-ton payloads per day. That is a design target, not a measured production rate. The December mapping contract will produce imagery and survey data, which exercises the sensors and the navigation stack. It will not exercise the collection sequence at commercial tempo, because mapping does not require collecting nodules.
Duty cycle is where deep-sea hardware usually fails. Biofouling, corrosion, and the fatigue of repeated launch and recovery accumulate quietly. A crawler that works for 72 hours needs a different maintenance model from a fleet that works for months. The same applies to the software: collaborative autonomy has to keep working when one unit is lost, when communications drop, and when the seabed map is wrong.
None of this disqualifies the design. It defines the diligence list. An investor should want mean time between failures, recovery rates from the 685-meter campaign, and the operating envelope for sea state. Those are the numbers that turn a demonstration into an asset.
What the 2026 record does not show
The company was founded in 2025 and lists 11 to 50 employees. It has not disclosed a funding round, a valuation, or a revenue figure. Those omissions are normal for a young hardware startup, and they are also the first questions a private investor should ask.
No disclosed extraction permit, and no named commercial mining customer.
No published environmental baseline from the 685-meter campaign.
ABS approval covers a conceptual platform, not a classified production system.
The December CCZ work is mapping, a service contract with no mineral revenue.
The competitive field is moving in the same direction. Impossible Metals is developing selective nodule collection robots and signed an exploration partnership with Eco Minerals in October. Eco Minerals, through its Deep Sea Vision subsidiary, sells autonomous survey capacity into environmental impact work. The robots are becoming a commodity faster than the permits are.
Environmental science is not standing still either. A study published in October found that deep-sea creatures help prevent polymetallic nodule burial, which adds another variable to every baseline study. The more the science learns, the more expensive it becomes to prove that a collection campaign leaves the seabed in acceptable condition. That cost lands on the operator, not the robot vendor.
As we wrote in October, Reliable Robotics reached a $1 billion valuation with its first certificate still 18 months away. The same pattern is visible here: capital is pricing the certification path, not the finished product. MOBY's ABS approval is a step on that path, and the 685-meter dive is the evidence that the path is real. Neither is a license to mine.
The bet is the dataset, not the crawler
The most defensible asset in this story is not the Mantis hull. It is the mission record. A company that maps the same seabed repeatedly, with the same sensors, builds a longitudinal dataset that no one else can reproduce quickly. That dataset is useful to miners, to environmental reviewers, and to insurers. It is also the part of the business that does not need a mining permit to generate revenue.
If MOBY converts the December campaign into a repeatable survey product, it becomes an ocean-intelligence company with a mining option attached. If it waits for a commercial extraction permit before it builds a customer base, it becomes a capital-intensive hardware bet with a single regulatory dependency.
The next twelve months are the test. Watch whether the CCZ operation runs on schedule in December. Watch whether ABS moves from an Approval in Principle to full classification. Watch for the first named commercial customer and the first published environmental monitoring dataset. Those four items will say more about MOBY than the depth rating of its crawler.