Exploring compact floating data centers. Connected to shore. Designed around a useful first pod—and a wider network to follow.
Read the journalOffshore Pod explores nearshore data-center modules that bring computing, single-phase immersion cooling and shore-connected infrastructure together. The aim is to find where a floating system can offer a practical alternative as demand for compute grows.
Ideas, questions and the work ahead.
Stay close to the infrastructure that makes a data center useful: power, fiber, people and maintenance.
Our starting point is a sheltered coastal site with a real customer requirement. Moving onto water does not remove the need for a grid connection, permissions or reliable operations. We want to find the sites where a floating module can make practical and commercial sense.
Compatible electronics sit in a dielectric liquid. The liquid absorbs heat; a cooling system carries it away.
In single-phase immersion, the working fluid stays liquid during normal operation. Offshore Pod is exploring commercially available systems. The server, fluid and tank must be compatible, and the complete heat-removal system needs validation in its intended operating conditions.
Immersion reference · OCP ↗17%
2025 annual growthGlobal data-center electricity demand
485TWh
2025 estimateGlobal data-center electricity use
950TWh
2030 projectionIEA central outlook for electricity use
Global data-center electricity demand. A growing need for power, sites and carefully tested infrastructure.
IEA · published 16 Apr 2026 ↗DNV reported an independent technical review of Fred. Olsen 1848’s BRIZO system, covering design methods, wave-load assessment, structural behavior and testing. This is industry context for the marine engineering work ahead.
Read the source ↗OCP’s Version 1.5 guidance addresses server-component and immersion-fluid testing. Its revision history records a 10 June 2026 update. The outcome still depends on the hardware, fluid and operating conditions.
Read the source ↗The IEA’s 2026 outlook projects data-center electricity use at around 950 TWh in 2030—about 3% of world electricity demand. The outlook reinforces the need to assess power access alongside computing capacity.
Read the source ↗From one useful module to a connected system.
I’m Yash Jaiswal, the founder of Offshore Pod. My focus is to understand the customer need, define the first useful deployment and bring the right engineering and operating partners together.
Meet the peopleConcept development: visual layouts, research and system scoping. The next step is a customer-defined pilot with supplier and site inputs. No operating prototype is claimed.
A proposed architecture for nearshore compute: repeatable floating pods, power and fiber from land, and separate solar floats with small spiral wind turbines where local conditions support them.
The initial architecture uses shore electricity as the main supply. Separate floating solar arrays and small Archimedes-style turbines are additional sources to evaluate, connected through suitable protection, conversion and distribution equipment.
A protected route connects the pod to an approved shore landing. Available grid capacity, connection lead time, fiber routes and backup arrangements shape the site choice.
Separate panel floats connect through suitable power conversion and protected cabling. Solar area, local sunlight, waves, moorings and access all need to be designed together.
The visual concept uses Archimedes-style turbines on dedicated service spaces. Output depends on measured wind, turbine power curves and spacing. A spiral shape alone does not establish higher energy production.
Solar and wind connect through suitable converters, protection and power controls. Battery storage or a UPS may support the selected operating strategy; capacity and backup duration remain to be engineered.
A site energy assessment will establish actual generation and storage needs. The concept does not claim continuous off-grid operation.
Small-wind guidance ↗ Floating-solar guidance ↗Our proposed single-phase system circulates dielectric liquid around compatible IT equipment and transfers heat through a heat exchanger. A site cooling circuit then rejects that heat. Fluid containment, marine corrosion, water temperature and any environmental effects must be considered before selecting the final arrangement.
OCP immersion requirements ↗Single-phase describes the liquid’s operating state. It does not, by itself, mean pump-free cooling or zero environmental impact.
Our long-term vision is a network of regional floating compute sites. Nearby pods could share high-capacity fiber and support infrastructure. A common software layer would manage capacity, operations and customer access across sites.
Define a workload and a site. Select compatible equipment, then validate the complete system with a focused pilot.
Add pods as demand grows. Engineer power, fiber, servicing and redundancy for the cluster, rather than assuming capacity is unlimited.
Connect regional sites through a common service layer. Network latency, data location and workload needs determine where jobs should run.
A global service would still consist of separate sites. Connecting them does not automatically create one tightly coupled GPU-training cluster or guarantee lower latency.
At suitable coastal sites, floating modules could add an option where land availability constrains development. The comparison must include marine infrastructure and connection costs.
Repeatable modules could let an operator add capacity alongside demand, and explore redeployment where permissions, transport and connections make it viable.
Local solar and wind could offset part of grid consumption where the energy yield and full life-cycle economics support the investment.
If technically and commercially validated, the service could support AI inference, scientific computing and other digital workloads. Affordability and access are goals to test with customers.
The immediate work is concept development: visual layouts, public research and system scoping. Next come supplier-backed requirements, a site energy budget, a thermal and marine review, and a customer-defined pilot. Physical performance will need to be measured before deployment claims are made.
Cost, deployment time, water use, emissions and availability will be assessed against a comparable land-based option. Site-specific approvals remain part of that work.
Reference snapshot checked . Sources inform the concept; they do not constitute validation or an affiliation.
A thesis for modular vessels providing compute capacity. It informs the direction of this project; it is not an endorsement of Offshore Pod.
The IEA projects global data-center electricity use rising from 485 TWh in 2025 to about 950 TWh in 2030. This is a sector forecast, not an Offshore Pod performance claim.
A reference for immersion terminology, fluid requirements, thermal interfaces and system documentation.
Guidance for evaluating component and fluid combinations. Following a reference document does not itself certify an integrated product.
Floating solar requires engineering for its floats, moorings and environment. Sheltered-water guidance should not be assumed to cover harsh offshore conditions.
A review of another floating-solar system illustrates the importance of hydrodynamic loads, structural behavior and testing. It is not a review of Offshore Pod.
Annual energy estimates depend on the turbine’s power curve, site wind, tower height and local conditions. Nameplate power alone is insufficient.
A published study combining simulation and field testing. It does not establish the performance of the proposed Offshore Pod layout.
YC’s “Compute at Sea” thesis describes modular vessels working together as compute infrastructure. Offshore Pod is exploring its own nearshore approach. There is no claimed YC affiliation or approval.
Read the original RFS ↗A project spanning computing, cooling, marine engineering, power and customer operations. Offshore Pod is led by Yash Jaiswal, with technical and commercial roles being developed as the concept takes shape.
Roles, focus and contributions.
Founder
Concept development & customer discovery
Building Offshore Pod’s direction: understanding the customer, researching the system and bringing together the people needed to evaluate it.
Interested in a pilot, a suitable coastal site, equipment integration or an engineering collaboration? Start a conversation about the workload, the location and the problem worth solving.
We are developing the concept, the customer case and the requirements for a first pilot. This site shares that direction and the work ahead.