Atomarine
Building floating data centers powered by offshore vessels for scalable, location-agnostic AI compute.
Website: https://atomarine.co/
Cover Block
Open sources
| Field | Detail |
|---|---|
| Company | Atomarine |
| Tagline | Building floating data centers powered by offshore vessels for scalable, location-agnostic AI compute. [Atomarine, retrieved 2026] |
| Headquarters | Cambridge, Massachusetts, US [Crunchbase, retrieved 2026] |
| Founded | 2026 [Y Combinator, retrieved 2026] |
| Stage | Seed [Mission Critical Engineers, November 2026] |
| Business model | Hardware + Software |
| Industry | Deeptech |
| Technology | Hardware |
| Geography | North America |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (2), Dimitris Koutentakis and Emile Germonpré [Mission Critical Engineers, November 2026] |
| Funding label | Seed |
| Total disclosed funding | ~$19,000,000 [Mission Critical Engineers, November 2026] |
Links
Open sources
- Website: https://atomarine.co/
- LinkedIn: https://www.linkedin.com/in/emile-germonpr%C3%A9-766045245/
What an Investor Needs First
PUBLIC Atomarine is building floating data centers for AI workloads, a concept that has drawn investor attention because it proposes to shift capacity buildout off land and around grid and siting constraints at a moment when large-scale compute demand remains a live infrastructure bottleneck [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. Founded in 2026 at MIT by Dimitris Koutentakis and Emile Germonpré, the company is still early, but the founding story is unusually aligned with the technical problem: Koutentakis has MIT training across electrical engineering, mechanical engineering, and business, while Germonpré is reported to hold a PhD in nuclear engineering focused on maritime applications [Mission Critical Engineers, November 2026] [MLQ.ai, November 2026] [LinkedIn, retrieved 2026].
The product thesis is straightforward in outline and ambitious in execution: standardized compute barges anchored roughly 10 to 12 nautical miles offshore, initially powered by gas and later by ship-borne small modular reactors, with seawater cooling as part of the operating model [Y Combinator, retrieved 2026] [Crunchbase, retrieved 2026]. Atomarine argues this architecture can compress deployment timelines and bypass local opposition, land scarcity, and interconnection queues, although the headline performance claims, including 4x faster deployment and sub-1.1 PUE, remain company-sourced at this stage [Atomarine, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026].
The company disclosed a $19 million seed round in November 2026 led by Y Combinator, with Pantera Capital, Soma Capital, Seven Seven Six, Liquid 2 Ventures, Audacious Ventures, and E14 Fund also named in coverage, giving Atomarine enough early backing to pursue design, partnerships, and pilot preparation [Mission Critical Engineers, November 2026] [Blackridge Research, November 2026]. The business model is best understood today as hardware plus software infrastructure for large-scale compute deployment, but there is no public evidence yet of signed customers, contracted capacity, or named commercial partners beyond references to early barge-builder and SMR-vendor relationships [Y Combinator, retrieved 2026] [MLQ.ai, November 2026].
Over the next 12 to 18 months, the main questions are less about narrative appeal and more about execution proof: whether Atomarine can turn its offshore design and power partnerships into a credible 2028 gas-powered pilot path, and whether it can translate a novel infrastructure concept into bankable deployment milestones [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. For investors, the near-term watchpoints are partner specificity, regulatory progress, engineering hires, and any public sign that the company can narrow the gap between concept architecture and project development discipline [Y Combinator, retrieved 2026] [dmthefounder.com, retrieved 2026].
Claim stands unchecked -- This section relies heavily on company and accelerator descriptions, with partial corroboration from Mission Critical Engineers, MLQ.ai, Crunchbase, and LinkedIn.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Deeptech |
| Technology Type | Hardware |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | Seed, total disclosed about $19,000,000 |
Inside the Company
PUBLIC
Atomarine is unusually early, but the company has put a specific infrastructure thesis into the market: move AI data center capacity offshore rather than fight for land, grid access, and community approvals onshore [Atomarine, retrieved 2026] [Crunchbase, retrieved 2026]. Public materials place the company in Cambridge, Massachusetts, describe it as founded in 2026, and frame the product as floating compute platforms supported by offshore power infrastructure [Crunchbase, retrieved 2026] [Atomarine, retrieved 2026].
The public chronology is still short. Atomarine appears in Y Combinator's Summer 2026 cohort, then surfaced more broadly with a disclosed $19 million seed round reported in November 2026; Crunchbase also lists the company and its financing profile, while the company website outlines a first gas-powered pilot targeted for 2028 and standardized offshore platforms sized at 75 MW to 100 MW each [Crunchbase, retrieved 2026] [Atomarine, retrieved 2026].
The legal entity is not clearly established in the cited public records used for this section, so the cleaner read is to stay with what is corroborated: a newly formed Cambridge-based startup, founded by Dimitris Koutentakis and Emile Germonpré, pursuing floating data center infrastructure as a venture-scale hardware and software business [Crunchbase, retrieved 2026] [Atomarine, retrieved 2026]. At this stage, the company overview is defined less by operating history than by the specificity of the build plan it has chosen to publish.
Partially corroborated -- Confirmed primarily by Crunchbase and the company website; legal entity detail was not corroborated in the cited public records.
Under the Hood
Reported and inferred
The product story is straightforward in concept and unusually hard in execution. Atomarine says it is building floating data centers as standardized compute barges, intended to sit roughly 10 to 12 nautical miles offshore and host AI workloads, with power supplied separately rather than drawn from a terrestrial grid [Y Combinator, retrieved 2026]. Public descriptions are consistent on the core architecture: the first deployment is planned to use gas-powered generation [PUBLIC], while the longer-term design pairs floating compute platforms with nuclear power-delivery ships carrying small modular reactors [PUBLIC] [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026].
The technical pitch rests on three public claims, each material and each still company-led. First, Atomarine argues offshore siting can bypass land, community, and interconnection constraints that slow conventional data-center buildouts [Y Combinator, retrieved 2026] [Pivot News, November 2026]. Second, it says seawater cooling can support power usage effectiveness below 1.1 [PUBLIC], and Crunchbase separately describes the system as using seawater cooling [PUBLIC] [Atomarine, retrieved 2026] [Crunchbase, retrieved 2026]. Third, the company frames the hardware as standardized and manufacturable, claiming 75 to 100 MW per compute platform [PUBLIC], up to 1.5 GW deployed per shipyard per year [PUBLIC], and faster deployment than land-based facilities, stated as 4x on its site and 2x in YC-related materials [PUBLIC] [Atomarine, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026]. Those figures are useful as design intent, but there is no verified demo, operating site, or third-party performance validation in the cited public record.
One hiring signal adds texture without settling the engineering risk. A third-party listing for a Hardware and Infrastructure Engineer points to work spanning naval architecture, marine power delivery, seawater cooling, floating structures, shipyard partners, classification societies, and maritime regulators, which is directionally consistent with the product described in company and press materials [dmthefounder.com, retrieved 2026]. That supports a reading that the effort is as much maritime and regulatory systems integration as data-center design, though any deeper tech-stack discussion would be an inference rather than a documented product fact.
Claim stands unchecked -- The section relies primarily on company descriptions and secondary coverage, with limited independent corroboration and no verified public demo or operating deployment.
Market Research
PUBLIC
The market matters now because AI infrastructure demand is forcing buyers and developers to look beyond conventional land-based data center buildouts, while power, permitting, and grid interconnection constraints are tightening in parallel [Y Combinator, retrieved 2026] [Pivot News, November 2026].
Public sources on Atomarine do not provide a third-party TAM, SAM, or SOM, so the cleaner approach is to frame the addressable market through the adjacent categories the company is trying to combine: AI compute infrastructure, utility-scale power delivery, and maritime industrial capacity. What is visible in the record is the problem statement the company is aiming at. Atomarine says its barges are designed for AI workloads and intended to move data center capacity offshore to avoid land constraints, local opposition, and interconnection delays [Y Combinator, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026]. That does not size the market, but it does identify a large existing spend pool in data center development and power procurement, rather than a new budget line that would need to be created from scratch.
The demand drivers in the available reporting are unusually concrete, even if most remain company-described. First, Atomarine is targeting a setting where deployment speed appears to matter as much as steady-state operating cost. The company claims 4x faster deployment than on land on its website, while YC-related materials describe 2x faster deployment, a gap that suggests the speed thesis is central but not yet consistently framed across public materials [Atomarine, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026]. Second, power availability is positioned as the bottleneck. The architecture uses standalone generation, first gas and later ship-borne SMRs, to avoid waiting on terrestrial grid expansion and utility interconnection queues [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. Third, cooling efficiency is part of the economic case, with the company citing seawater cooling and sub-1.1 PUE, though that figure is company-only and should be treated as an engineering target rather than an observed operating result [Crunchbase, retrieved 2026] [Atomarine, retrieved 2026].
The adjacent and substitute markets are clearer than the market size. Atomarine sits beside conventional hyperscale and colocation data center development, on-site power systems for data centers, and emerging nuclear-enabled infrastructure for compute. The substitute for Atomarine is not another floating platform in the public record, but the incumbent path of building on land and pairing sites with utility power, gas generation, or behind-the-meter generation. That matters because the company is not asking customers to buy compute demand they do not already have. It is asking them to accept a different delivery model for capacity, one shaped by marine construction, offshore operations, and power-plant-style risk rather than standard real estate development [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026].
Regulation and macro conditions cut both ways. The positive side of the thesis is that offshore siting is explicitly meant to reduce local land-use friction and bypass some terrestrial constraints, and the use of standardized barges built on a manufacturing line points to a supply-side industrialization argument rather than one-off project development [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026]. The harder side is that the later nuclear roadmap depends on maritime regulation, reactor licensing, vessel or platform classification, and long lead-time execution across shipyards and power vendors, none of which are evidenced in detail in public materials. Even the nearer-term 2028 gas-powered pilot implies a permitting and construction path that differs from software-style scaling [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. In practice, the market opportunity is large if buyers increasingly value guaranteed power and deployment speed over conventional site familiarity, but the route to that market is conditioned by infrastructure regulation rather than pure demand.
| Cited market-related claim | Figure | Context |
|---|---|---|
| Standardized compute platform capacity | 75 to 100 MW | Company-described per platform design range [Atomarine, retrieved 2026] |
| Two-platform capacity ambition | 150 MW | Reported eventual output across two platforms [Pivot News, November 2026] |
| Claimed deployment speed | 2x faster | YC launch materials [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026] |
| Claimed deployment speed | 4x faster | Company website claim [Atomarine, retrieved 2026] |
| Claimed shipyard throughput | 1.5 GW per year | Company-described output per shipyard [Atomarine, retrieved 2026] |
The table shows why the market case is interesting but still early. The figures suggest Atomarine is aiming at utility-scale infrastructure from day one, yet the public numbers remain largely company-supplied and are better read as design intent than validated market proof.
Claim stands unchecked -- This section relies primarily on company and company-amplified claims about deployment speed, capacity, cooling, and market drivers, with limited independent corroboration from Mission Critical Engineers, Pivot News, and Crunchbase.
Competition and Substitutes
Market structure
MIXED Atomarine is positioning against a set of land-based data center, on-site power, and alternative-compute infrastructure models rather than against a clearly disclosed list of same-design offshore peers [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026].
The competitive map is easiest to read in three layers. First are incumbent land-based data center developers and operators, which still control the default deployment path for AI infrastructure even as interconnection queues, local permitting, and community opposition have become a more visible constraint in industry coverage [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026] [Pivot News, November 2026]. Second are challengers trying to pair compute with dedicated power, including nuclear-adjacent and other self-generation concepts, where Atomarine's own public framing sits: standardized compute barges, initially gas-powered, later paired with ship-borne SMR capacity [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. Third are adjacent substitutes, namely any approach that gets AI clusters online faster without moving offshore, including conventional terrestrial campuses with alternative power procurement or modular buildouts. The public record does not name Atomarine's direct peers, so the comparison has to stay at category level rather than company level.
That matters because Atomarine is not trying to win on software features or customer workflow. It is making an infrastructure siting and power-delivery argument: move the campus offshore, standardize the platform, use seawater cooling, and reduce dependence on local grid upgrades and local land politics [Atomarine, retrieved 2026] [Crunchbase, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026]. If that proposition holds in practice, the relevant competitors are less likely to be other early startups and more likely to be whichever incumbent developers can deliver large blocks of powered capacity onshore within similar timelines.
Sources of advantage
MIXED The clearest edge visible today is founder-market fit around the technical stack, but it is still an early and perishable advantage.
On public evidence, the team has unusual overlap across AI, naval architecture, and nuclear engineering. Dimitris Koutentakis is identified as founder and CEO, with MIT training across electrical engineering, mechanical engineering, and business, and prior roles at Amazon and Arrowstreet Capital according to MLQ.ai [MLQ.ai, November 2026]. Emile Germonpré is identified as co-founder and CTO, with a PhD in nuclear engineering and research tied to offshore nuclear reactor design for maritime applications according to Mission Critical Engineers and LinkedIn-sourced profiles [Mission Critical Engineers, November 2026] [LinkedIn, retrieved 2026]. For a company proposing floating compute barges powered by maritime energy assets, that interdisciplinary profile is more specific than what many infrastructure startups can show at seed.
There is also a capital and network signal. Atomarine's $19 million seed round was reported in November 2026, led by Y Combinator, with Pantera Capital, Soma Capital, Seven Seven Six, Liquid 2 Ventures, Audacious Ventures, and E14 Fund also named in coverage [Mission Critical Engineers, November 2026] [Blackridge Research, November 2026]. Y Combinator's profile also says the company is part of the Summer 2026 batch [Y Combinator, retrieved 2026]. That does not prove commercial traction, but it does improve access to follow-on financing and technical recruiting for a concept that will likely need both in quantity.
The durability question is straightforward. Founder expertise is durable at the idea and design stage, but much less so once incumbent operators, shipyards, utilities, or reactor vendors decide the category is worth pursuing. Atomarine also says it has early partnerships with offshore barge designers and builders, plus multiple SMR vendors, but no public source names those counterparties, so the moat value of those relationships cannot yet be tested from the outside [Y Combinator, retrieved 2026] [MLQ.ai, November 2026].
Where the company is exposed
MIXED The company's largest exposure is that its strongest claimed advantages remain company-stated and future-dated, while the competing substitute, conventional onshore deployment, already has mature buyers, contractors, and operating playbooks.
Public materials describe several performance claims: 75 MW to 100 MW per standardized compute platform, less than 1.1 PUE through seawater cooling, 4x faster deployment than on land, and 1.5 GW deployed per shipyard per year [Atomarine, retrieved 2026]. Another company-related source cites 2x faster deployment, and Pivot News reported a long-run ambition of 150 MW across two platforms [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026] [Pivot News, November 2026]. Those figures are directionally central to the pitch, but they are not independently validated in the sources provided. That leaves Atomarine competitively exposed to any incumbent or adjacent challenger that can show real delivered megawatts sooner, even with a less novel architecture.
The other exposure is regulatory and supply-chain complexity. Atomarine's stated roadmap begins with a gas-powered pilot in 2028 and shifts to nuclear deployment in 2031 or 2032 [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. That means its category leadership, if it develops, is likely to be won first on execution through marine engineering, classification, permitting, and partner coordination, not on branding. A category built on unnamed barge-builder and SMR relationships can move quickly if those ties are firm, but it can also prove fragile if counterparties reprioritize, if reactor timelines slip, or if customers prefer the lower coordination burden of terrestrial alternatives.
Likely near-term scenario
MIXED The next 18 months are likely to be decided less by end-state nuclear economics than by who can turn a concept into a credible development path.
If Atomarine can convert its November 2026 financing into disclosed technical milestones, partner specificity, and evidence that the 2028 gas pilot remains on schedule, it has a plausible claim to be the winner if speed-to-capacity becomes the governing customer criterion for frontier AI infrastructure [Mission Critical Engineers, November 2026] [Y Combinator, retrieved 2026]. The company has already defined a crisp wedge around avoiding land, community, and grid constraints, which is more concrete than a generic "AI infrastructure" label [Pivot News, November 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026].
The most plausible loser if execution slips is Atomarine itself, not because the core idea is disproven, but because adjacent substitutes can absorb demand while it is still de-risking marine and nuclear dependencies. Public evidence does not identify a named startup rival to assign that role to, and that absence is itself informative: the immediate competitive threat is the incumbent path buyers already understand, not necessarily a better-publicized offshore startup. In other words, the burden is on Atomarine to prove that unconventional siting yields conventional reliability, faster than onshore alternatives can close the same capacity gap.
Opportunity
Open sources If Atomarine can turn offshore power-and-compute infrastructure from a concept into an operable product, the prize is not a niche marine project but a new supply path for large-scale AI data center capacity at a moment when land, grid interconnection, and local permitting are all binding constraints for conventional builds [Y Combinator, retrieved 2026] [Pivot News, November 2026].
The headline opportunity is straightforward: become a category-defining developer of offshore AI compute campuses, where power generation and data center capacity are packaged together and deployed independently of constrained onshore sites. That outcome is still early and highly conditional, but the public case for reachability rests on three cited facts. Atomarine has already assembled a $19 million seed round led by Y Combinator with participation from Pantera Capital, Soma Capital, Seven Seven Six, Liquid 2 Ventures, Audacious Ventures, and E14 Fund [Mission Critical Engineers, November 2026] [Blackridge Research, November 2026]. It has also put forward a concrete deployment sequence, gas-powered pilot in 2028, then nuclear deployment in 2031 or 2032, rather than a purely open-ended research agenda [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026]. Just as important, the company is not describing a one-off custom installation, but standardized compute barges and power-delivery ships, which is the architectural choice that makes repeat deployment and manufacturing scale at least conceivable [Y Combinator, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026].
The public evidence supports a few distinct paths to scale, and they do not all require Atomarine to solve the full nuclear roadmap at once.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Gas-first offshore capacity provider | Atomarine proves the 2028 pilot and begins selling offshore capacity before nuclear readiness, using ship-based gas generation to serve AI workloads that cannot wait for onshore grid upgrades | A successful first pilot in 2028 and conversion of early infrastructure partnerships into build contracts [Y Combinator, retrieved 2026] | The company has already framed gas as the initial power source and positioned offshore deployment as a way around land and interconnection bottlenecks [Y Combinator, retrieved 2026] [Mission Critical Engineers, November 2026] |
| Offshore nuclear campus standard | Atomarine becomes an early integrator for SMR-powered floating compute, pairing nuclear power-delivery ships with standardized barges and setting the template for later entrants | A named SMR partnership or first nuclear deployment in 2031 or 2032 [Mission Critical Engineers, November 2026] [Y Combinator, retrieved 2026] | The company says it has partnerships with multiple SMR vendors and has designed the architecture around ship-borne nuclear power from the outset [Y Combinator, retrieved 2026] [MLQ.ai, November 2026] |
| Maritime manufacturing platform | Atomarine uses a standardized barge design and shipyard production model to scale capacity more like marine fabrication than like bespoke campus construction | Demonstrated repeatability of a standard platform and proof that shipyard throughput can compress deployment timelines [Atomarine, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026] | Its public materials repeatedly stress standardized, plug-and-play barges built on a single production line, with claimed deployment-speed advantages over land-based builds [Atomarine, retrieved 2026] [Launch YC: Atomarine: Floating Nuclear Data Centers, retrieved 2026] |
What compounding looks like here is less a consumer network effect than an infrastructure flywheel. One successful platform would not simply produce revenue, it would validate a template: offshore siting, marine construction partners, cooling design, classification and regulatory workflows, and a standardized compute module that can be copied across additional deployments [Y Combinator, retrieved 2026] [Crunchbase, retrieved 2026]. If that template works, each subsequent project could get easier to finance and faster to deploy because the core design is reused rather than reinvented. Atomarine's own public claims, including 4x faster deployment than on land, 1.5 GW deployed per shipyard per year, and sub-1.1 PUE via seawater cooling, remain company-only and should be treated cautiously, but they at least point to the mechanism by which an initial engineering win could become a manufacturing and cost-of-capacity advantage over time [Atomarine, retrieved 2026].
The size of the win is potentially very large even before assuming full category dominance. Atomarine says each standardized compute platform is designed for 75 MW to 100 MW, and outside reporting says the architecture could eventually deliver 150 MW across two platforms [Atomarine, retrieved 2026] [Pivot News, November 2026]. If a future version of the company were operating several such platforms, it would not resemble a small infrastructure startup, it would resemble an owner-operator of utility-scale digital infrastructure. Public sources provided for this report do not include a third-party market sizing study or a direct public comparable with matching economics, so a precise valuation benchmark would overstate the evidence. Still, if the "offshore nuclear campus standard" scenario plays out and Atomarine becomes one of the first repeatable suppliers of large-format AI compute capacity outside the land-constrained grid queue, the company could plausibly grow into a multi-billion-dollar infrastructure platform (scenario, not a forecast), because the underlying unit of deployment is measured in tens of megawatts rather than racks or single enterprise contracts [Atomarine, retrieved 2026] [Pivot News, November 2026].
Claim stands unchecked -- This section relies materially on company statements and company-linked reporting about future deployments, platform capacity, partnerships, and performance claims, with limited independent public corroboration.
Sources
Open sources
[Atomarine, retrieved 2026] Atomarine | https://atomarine.co/
[Crunchbase, retrieved 2026] Atomarine - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/atomarine
[Y Combinator, retrieved 2026] Atomarine | https://www.ycombinator.com/companies/atomarine
[Mission Critical Engineers, November 2026] Atomarine raises $19 million seed round for floating data centres with ship-borne power | https://research.missioncriticalengineers.com/news/atomarine-floating-nuclear-dc
[MLQ.ai, November 2026] Atomarine raises $19 million for floating AI data centers it plans to power with ship-mounted nuclear reactors | https://mlq.ai/news/atomarine-raises-19-million-for-floating-ai-data-centers-it-plans-to-power-with-ship-mounted-nuclear-reactors/
[LinkedIn, retrieved 2026] Emile Germonpré - Co-Founder @ Atomarine (YC S26) | https://www.linkedin.com/in/emile-germonpr%C3%A9-766045245/
[Blackridge Research, November 2026] Atomarine's $19 Million Floating Data Center Plan | https://blackridgeresearch.com/news-releases/us-based-atomarine-raises-usd-19-million-to-develop-nuclear-powered-floating-data-centers
[dmthefounder.com, retrieved 2026] Atomarine Hardware and Infrastructure Engineer | https://dmthefounder.com/jobs/atomarine-hardware-and-infrastructure-engineer-8a05b6
[Pivot News, November 2026] Atomarine takes $19m seed round for floating nuclear data centers | https://pivotnews.ai/five/atomarine-takes-19m-seed-round-for-floating-nuclear-data
Articles about Atomarine
- A Floating Data Center for Every Coast: Atomarine's 2028 Pilot — The YC-backed startup plans a gas-powered pilot by 2028, with nuclear-powered ships to follow, betting that the ocean solves land, grid, and community constraints.