Panthalassa's floating, wave-powered data centre technology

One of our first big milestones — and we mean big — came with Ocean-2, a full-scale prototype deployed off the Washington coast, built by our team of inventors, programmers, welders, physicists, and engineers.

Panthalassa is pioneering a new kind of data centre which floats in the deep ocean, powered by wave energy and connected by Starlink satellites. 

 Credit: Panthalassa 

From Sustainability Mag by Charlie King

Panthalassa is pioneering a new kind of data centre which floats in the deep ocean, powered by wave energy and connected by Starlink satellites.

 
AI’s appetite for energy is pushing investment into unchartered territory, with US start-up Panthalassa attracting millions for its floating data centres

If proof were needed that necessity drives innovation, the race to power AI has made the case undeniable.

Since the artificial intelligence boom accelerated in 2022, one of the defining challenges for the global economy has been how to supply the vast amounts of energy the technology demands. Companies across the world are now competing to develop viable solutions.

It is widely recognised that AI systems – along with the data centres that support them – consume enormous quantities of energy.

Indeed, the IEA projects that the sector’s energy consumption will rise by 30% annually through to 2030, when AI is expected to account for 3% of global energy use.

This surge in demand has triggered a wave of unusual and inventive approaches, from restarting nuclear facilities to exploring satellite-based solar power and even investing in fusion.

 

The IEA's projections for data centre energy consumption. Credit: IEA

Among the more unconventional ideas gaining traction is wave energy – a frequently overlooked renewable resource.

Panthalassa, an Oregon-based start-up that has spent the past decade refining its wave energy technology, has recently positioned itself at the forefront of this space.

The company takes its name from the “superocean” that once surrounded Pangea before the continents formed as we know them today – a fitting reference for a business built on harnessing ocean power.

Earlier this year, Panthalassa raised US$140m in a funding round led by venture capitalist Peter Thiel, known for backing Palantir and PayPal.

 

In just 10 years, Panthalassa has achieved a valuation of US$1bn. 

Credit: Panthalassa

The strength of investor interest has pushed the company’s valuation to nearly US$1bn, including the new capital.

Alongside Peter, the round attracted high-profile supporters such as Salesforce CEO Marc Benioff, PayPal and Affirm Co-Founder Max Levchin, and veteran investor John Doerr, an early backer of Google, Amazon, Uber and Netscape.

 

Peter Thiel, Co-Founder of Palantir and PayPal.

Credit: Gage Skidmore

How Panthalassa’s floating data centres work

From an engineering standpoint, Panthalassa’s approach is intentionally simple.

Its “nodes” are 85-metre-long steel structures – roughly the height of London’s Big Ben – designed to sit mostly below the ocean surface.

As waves move the structure, water is forced through a turbine, generating electricity that powers AI chips housed within a sealed, seawater-cooled container.

Manufacturing the Panthalassa nodes. Credit: Panthalassa

The system avoids complexity: there are no hinges, flaps or gearboxes, and few components are exposed to the harsh realities of open-ocean conditions.

Each node continuously recirculates water internally to drive its generator, producing no direct emissions and requiring no engine.

Once towed out to sea horizontally, the structure rotates upright and travels to its designated location using only the hydrodynamic properties of its hull.

Panthalassa's wave-powered data centres do not require any connection to land. Credit: Panthalassa

Because these floating data centres operate far offshore, all AI queries are transmitted via SpaceX’s Starlink satellite network.

According to the company, this satellite link is the only land-based connection the system requires.

The energy argument

Panthalassa Co-Founder and CEO Garth Sheldon-Coulson, formerly an AI and energy researcher at hedge fund Bridgewater, has been a vocal advocate for wave energy’s potential.

"Energy from open-ocean waves is low-cost, sustainable, abundant and now we have the technology to make it accessible for people," he recently told the Financial Times.

By positioning its nodes in remote deep-ocean locations rather than near coastlines, Panthalassa aims to avoid the limitations that have hindered earlier wave energy projects, such as grid connections and inconsistent wave conditions.

 

Garth Sheldon-Coulson, Panthalassa's Co-Founder and CEO. 

Credit: Garth Sheldon-Coulson

"The waves are like a battery for sunlight and we can be capturing from it 24/7," he adds.

The company argues that wave and wind power, alongside solar and nuclear, are among the few clean energy sources capable of generating “tens of terawatts” – the level of output increasingly required by AI infrastructure.

With global demand for clean energy accelerating, this proposition is likely to attract attention from both public- and private-sector stakeholders.

‘Go where the energy is’

A defining feature of Panthalassa’s model sets it apart from traditional offshore energy projects: it does not aim to transmit electricity back to land.

Panthalassa's motto is "go where the energy is". Credit: Panthalassa

Its guiding principle – “go where the energy is” – reflects this strategy.
"One of the key insights that we had was that it's very important to use the electricity in place," Garth Sheldon-Coulson explained to the FT.

This approach eliminates one of the biggest challenges in offshore renewables: the cost and complexity of subsea transmission cables and interconnectors.

It also shifts the company’s business model. Rather than selling electricity, Panthalassa sells compute capacity, simplifying its commercial proposition in several respects.

Panthalassa’s plans for the future

The newly secured funding will support the development of a pilot manufacturing facility in the US, with initial commercial deployments expected as early as next year.

 

 

We’re heading to the middle of the ocean, the planet’s most energy-dense resource, to harness the cleanest, cheapest energy to meet humanity’s growing needs.

 

Panthalassa states that its nodes are constructed using “earth-abundant materials” such as steel – a point that may ease pressure on global rare earth and critical mineral supply chains already strained by the energy transition.

The company also maintains that its supply chains are sufficiently resilient to support rapid scaling.

Peter, who has long supported the libertarian “seasteading” movement advocating for floating communities in international waters and has backed Panthalassa since 2018 through Founders Fund, views the company’s work as part of a broader expansion of technological frontiers.

Panthalassa says its technology only relies on abundant materials and minerals. Credit: Panthalassa

"The future demands more compute than we can imagine. Extraterrestrial solutions are no longer science fiction. Panthalassa has opened the ocean frontier," he told the FT.

While some critics have labelled such ideas as a form of “tech neocolonialism”, Peter’s endorsement is likely to accelerate Panthalassa’s path toward mainstream adoption. 

Trump Administration to remove hundreds of deep-ocean observation instruments, dismantling $368 Million program

 

A surface mooring deployed by the Ocean Observatories Initiative.
Credit: Sheri N. White © WHOI, Image from work supported by the U.S.
National Science Foundation Ocean Observatories Initiative

 From EOS by Grace van Deelen

 

Research & Developments is a blog for brief updates that provide context for the flurry of news regarding law and policy changes that impact science and scientists today.

The Trump administration’s National Science Foundation (NSF) has begun dismantling the infrastructure of a $368 million deep-ocean observing program critical to monitoring marine ecosystems, global currents, marine heat waves, and more, according to a 21 May announcement.

The Ocean Observatories Initiative (OOI), funded by the NSF, has been collecting long-term oceanographic data at multiple deep-ocean sites since 2016.
The information about ocean temperature, chemistry, currents, biological conditions, and more is used by scientists to understand a multitude of marine research questions including the activity of the Atlantic Meridional Overturning Circulation (AMOC), a critical ocean current.

“I worry that … we’ll be losing this enormously valuable site where we could really contextualize and detect these changes going forward.”

“There’s a real danger that we lose the ability to keep looking for long-term changes [in the ocean]” as climate change alters Earth systems, said Hilary Palevsky, a marine biogeochemist who has used OOI data for a decade to study how the ocean absorbs carbon dioxide.
“I worry that … we’ll be losing this enormously valuable site where we could really contextualize and detect these changes going forward.”

The NSF plans to remove all in-water arrays and infrastructure—including hundreds of deep-sea instruments—from four of the five currently-operating sites within the project: the Global Station Papa Array (in the Gulf of Alaska), Coastal Endurance Array (off the coasts of Oregon and Washington), Global Irminger Sea Array (southeast of Greenland), and Coastal Pioneer Array (off the coast of North Carolina).
The removal is expected to occur over the next 15 months, though the process has already begun at the Endurance Array. 

 

 

The National Science Foundation’s planned descoping of the Ocean Observatories Initiative will include dismantling four of the five currently operating arrays of equipment.

Credit: NSF/OOI

The Trump administration attempted previously to downscale OOI operations, proposing to cut its funding in 2025 and 2026, though Congress never approved the cuts.

The administration’s decision to dismantle the arrays “aligns with NSF’s wider strategy to have a nimbler approach to prioritizing support for evolving scientific priorities and emerging technologies as well as a deliberate approach to smart life cycle management within its portfolio of research infrastructure,” Michael England, an NSF spokesman, told the New York Times. 

 

Buoys used to gather data at the Ocean Observatories Initiative's Coastal Pioneer Array off the coast of Martha's Vineyard in 2021. 

The array has since moved to off the coast of North Carolina.

Véronique LaCapra/WHOI/AP

 

A Dearth of Data

As each array is dismantled, data streams will end, though all previously collected data from OOI networks will remain accessible, Jim Edson, principal investigator for the OOI, wrote in a letter to the oceanographic community.

Palevsky said there’s “a lot of real concern” among the oceanographic community that the Endurance Array is being dismantled just as an intense El Niño event—and associated marine heat wave—is expected this summer.
“It would be especially important to be able to document the effect that [El Niño] is having on coastal physical circulation and ecosystems,” she said. 

“We encourage the community to use the ten-plus years of OOI data by including it in proposals, publications, presentations, and conversations with colleagues.
Continued engagement demonstrates the scientific impact and wide-ranging applications enabled by the OOI and its data, underscoring its importance as a resource for the oceanographic community,” the 21 May announcement stated.

There are other sources of data that researchers like Palevsky can use.
But oceanographic research often requires stitching together different data sets, including OOI observations, satellite observations and observations from the U.S. research fleet.
Many of these other sources of data are also facing uncertain futures.

Palevsky also worries about the loss of expertise that will occur as the program scales down.
Installing these deep-sea observing networks was a huge achievement for U.S. science that will not be easy to replicate, she said.
“If, in five years, we as a community decide we want to again be able to deploy this kind of complicated infrastructure in places that have really difficult oceanographic conditions … it’s going to be a lot of reinventing the wheel to figure out how to put things out again.”

“The complete cessation without community input or a community conversation about what’s going to happen to all this equipment and what’s going to happen with all of the expertise,” she said, “feels like a huge loss.”

 

Links :

• TheNYTimes : Trump Administration to Dismantle Ocean Monitoring System 
• CNN : The oceans are in deep trouble. The Trump administration is ditching a vital deep-sea monitoring system 
• Le Monde : Le système mondial d’observation des océans fragilisé par l’arrêt d’un réseau de surveillance américain 

• Announcement on NSF OOI Descoping
• NSF OOI Descoping Update for the Community
• Trump Administration to Dismantle Ocean Monitoring System
• Democrats Pledge to Fight Trump’s Removal of Ocean Monitors
• Get Involved: AGU’s Science Policy Action Center 

The past and future surveyor

Remote control centre for USV.

(Image courtesy: Exail)

From Hydro by y Huibert-Jan Lekkerkerk

How past and current developments may impact the surveyor

The world is ever changing.
So is the profession of the hydrographic surveyor.
But how will current technological and societal changes impact hydrographic surveying? Will this be a matter of historical recurrence or are we on the brink of something completely new? Let us look at some past developments, taking a line from George Santayana (1905): “Those who cannot remember the past are condemned to repeat it.”

This article will therefore first consider some historical developments to see what may happen in the future to our profession.
Please note that this overview is neither complete nor can developments be pinned to an exact time period.

Until 1720: the age of discovery

Before the 18th century, hydrographic surveying was a very imprecise business with data unsystematically gathered from a great number of sources.
Bathymetry was mostly absent in early charts and positioning relied on relatively crude latitude measurements and dead reckoning.
Charts were either state or company secrets or were created and distributed by commercial printing houses.
Very often, information was not only inaccurate but also outdated.
New charts often copied old charts with a new look and name on them.

Hydrographic surveyors as we know them today were virtually non-existent, and chart information came from sea captains and explorers who wrote down what they witnessed.
They collected their data using whichever sailing ship they happened to be on, were often away for years on end and had to rely on navigators onboard, the education of whom was mostly in the hands of individuals.
Knowledge was transferred orally using hands-on experience or a select number of standard works on navigation that often were kept in print for decades.

Frontispiece of 'De groote lichtende ofte vyerighe colom' showing the state of the art of navigator education in the 17th century.
(Image courtesy: Allard Pierson Museum) 

1720–1920: hydrography becomes scientific

From the 18th century onwards, hydrography for the safety of navigation became more state-institutionalized, starting with the French Depot des Cartes et Plans de la Marine in 1720.
It was not until the second half of the 18th century that systematic chart updates based on proper hydrographic surveys were performed.
This was also the start of hydrography as we know it today.
First, land survey work was performed to set the geodetic network.
The development of the sextant and chronometer allowed relatively accurate determinations of latitude and longitude, which greatly improved the accuracy of data.
The hydrographic data was then systematically collected using resection from two sextant angles combined with depths from the lead and line along survey lines.
The charting itself relied, depending on the preference of the hydrographic service, on instruments such as the station pointer or on newly developed formulas.
Navigational charts were issued by governments rather than commercial companies.

Late 19th-century survey sextant.

(Image source: collection author)

As exploration became more systematic and institutional, specialized tools and training appeared.
Ships for exploration and surveys required their own outlay and specialized crew.
Engine power was adopted relatively quickly.
Hydrographic training was still ‘on the job’ and most hydrographic surveyors started their career as navigators.
Surveyors would be away for long periods but could often rely to some extent on existing infrastructure to support them.
As most surveying was government business, hydrographic surveyors became part of navies and were trained at navy institutes.
In the field they were supported by a survey crew that was trained on the job by the same surveyors who would oversee their work.
Books specifically devoted to hydrographic survey started to appear, such as those of Murdoch Mackenzie and Beautemps-Beaupre.

 

Illustration from Beautemps-Beaupre’s ‘Introduction to the practice of nautical surveying’ showing the resulting chart of a ‘modern’ survey.

(Image source: archive.org) 

1920–1970: hydrography goes electronic

The methods described were further refined during this period but essentially remained unchanged.
With the development of underwater acoustics, the single-beam echosounder made its introduction into hydrography and was quickly adopted as a standard tool.
In the United States, radio acoustic ranging was developed, a system that can be seen as a predecessor of long baseline positioning.
During WWII, electronic positioning systems were conceived which, after the war, were transformed into a multitude of high-accuracy hyperbolic and range-range positioning systems.

Photogrammetry for topography became mainstream, which also saw the introduction of aircraft into hydrography.
These new technologies were used side by side with the ‘old’ technologies.
Chart plotting did not change much and still required manual labour but chart printing was modernized.
Near the end of this period, the first of the ‘modern’ survey technologies such as multibeam echosounders, sidescan sonar, underwater acoustic positioning and sub-bottom profilers saw the light of day.

The establishment of the International Hydrographic Bureau (IHB, now IHO) in 1921 can be seen as the formal start of international cooperation which lasts until today.
With respect to hydrographic training, not much changed.
The publication of the International Hydrographic Review by the IHB and the hydrographic conferences held by the same did much to spread knowledge across the field.
Surveyors were now away for periods of a few months to maybe a year or so and could rely on existing infrastructure with relatively quick communications.

Radio acoustic ranging principle. 

(Image courtesy: NOAA) 

1970–1990: hydrography is automated

With the more systematic exploration and production of oil and gas, hydrographic surveying became a private, commercial, enterprise.
Though initially navies supported exploration, it became quickly clear that more capacity was required.
What also became clear was that project requirements were different although survey technologies were essentially the same.
Hydrographers were quick to adopt computers into the work process, allowing for faster data collection and processing.
As computer capacity increased, software became more elaborate and complete.

With the greater need for capacity, the training of surveyors could no longer be handled by just the navies, although many of the early commercial surveyors obtained their knowledge through the respective hydrographic services.
Specialized, civil training emerged with the IHO setting the standards for training programmes, which led to the Cat-A and Cat-B recognized courses we still see today.
This period also saw the establishment of hydrographic societies and new periodicals and congresses to continuously educate a much wider hydrographic audience and allow them to network and cooperate.

The surveyor from this period had to be skilled in both the ‘old’ manual techniques but also in the ‘new’ digital and electronic technologies.
The commercial environment also required faster turnaround times, and the surveyor could no longer afford a few months delay between surveying and delivering the final product.
Teams became smaller as automation did not require as many people.
The surveyor would generally be away for no longer than a few months and could rely on structured organizations and immediate communications with experts to help solve issues.

Crew at work on a survey launch in 1969.
(Image source: De Hollandse Cirkel) 

1990–2015: data revolution

GNSS and specifically dGPS were quickly embraced by the hydrographic world and almost fully replaced electronic positioning as they were about as accurate but much faster and cheaper to use.
At the same time, systems such as the multibeam echosounder and bathymetric Lidar became commercially available.
This changed the sparse data from single-beam to full bottom coverage, high-density datasets.

This period also saw the development of the geographic information system and of modern survey software to support the new data streams.
Charting became part of survey software supporting relatively quick turnaround of data to final product.
At the same time, the electronic navigational chart and electronic chart display system were defined and developed, allowing safety of navigation data to be distributed in digital form.
New platforms became more elaborate, with the ROV becoming the standard offshore tool.
The first autonomous underwater vehicles were developed but the main tool remained the survey vessel / launch and aircraft / helicopter for photogrammetry and airborne Lidar (bathymetry).

New technologies require new standards and commercial and civil institutes started to develop these standards, notably the European Petroleum Surveying Group (EPSG, now IOGP) and the International Marine Contractors Association (IMCA).
The new surveyor had to be able to handle the high data volumes and increased accuracy with tools that were still being developed and improved.
Survey crews became even smaller, but knowledge was easily disseminated through the internet and digital publications.
Hydrographic surveyors would be away for weeks to months now.
The hydrographic surveyor was responsible for a wide variety of systems using technology that was still under development.
As a result, training also became wider in subjects and more detailed in content with a focus on specific techniques and applications. 

2015–today: remote, autonomous and artificial

Most of the technologies we use today are still the same as in the previous era.
Systems have become easier to use if set up correctly.
However, clients also keep asking for more and higher quality data while setting stricter tolerances for construction.
Further miniaturization of electronics and the improvement of computing power have allowed the development of smaller and faster electronics.
Additionally, communications have become significantly faster and less expensive.

This has allowed the development of autonomous, uncrewed and remote systems.
The uncrewed aerial vehicle equipped with Lidar and photogrammetric cameras is standard on many construction projects.
The next step with remote control and remote processing of survey data with lightly or uncrewed and sometimes autonomous survey vessels is in full swing.
This has also changed the work environment; remote working does not require the remote surveyor to be away and, for the first time, some surveyors can work from behind their desk in the office and be home in time for supper.

With the increase in computing power, data processing has also become more automated.
Machine learning and artificial intelligence are out of the research phase and are slowly becoming mainstream in data processing.
Compiling data in databases and integrating it with other datasets is now standard for many clients for whom bathymetric data is just one aspect of their daily operations.

 

Cat-A students at work with a USV. 

(Image courtesy: Maritiem Instituut Willem Barentsz / NHL Stenden) 

Towards the future

What will hydrography look like in the next 10 to 20 years? No one can say for sure, but it is clear from history that new technologies will keep emerging.
Historically, hydrographic surveyors have shown themselves to be technocratic and flexible enough to be the early adopters and absorb new technologies quickly.
On the other hand, society has changed.
We can see this in the job rotation duration, which has gradually been reduced from years to weeks and for some no longer than a working day.

As the industrial revolution changed the way we propelled our survey vessels, the age of automation changes the way we collect our data.
Looking towards the future, we can see two types of surveyors emerge.
The first is very skilled in the higher theoretical and technological details of mobilization, data acquisition and processing.
This type of surveyor will possibly travel from site to site, mobilizing systems and troubleshooting them in the field.
Once the system has been set up and running, we will see another type of surveyor, more of an operator, taking over the operation.
These operators will most probably work increasingly remotely, and their main function will be overseeing the operation of highly automated systems.
When they detect an issue, they will involve a troubleshooting surveyor to analyse and resolve the issue, either through a change in the system or through corrective action with the automation.

Looking at data processing and products, we have seen a gradual change from pure safety of navigation products on paper to electronic products / data for a much wider use with integration into other datasets.
At the same time, the processing of data without major issues is becoming increasingly automated.
This will possibly create a similar division for data processing / charting as described for data acquisition.

The above translates to a potential paradigm shift in our industry that we have not seen for decades.
We will (again) require survey operators who can be trained relatively quickly and without all the theoretical details as well as more technical surveyors who can oversee the operations and can analyse and troubleshoot the system based on in-depth knowledge.
At the same time, there are so many systems around that it is impossible to be trained in detail on each system and method.
Education will need to give a basic understanding of all technologies and techniques with specialization occurring through additional formal training geared towards the application.

The above may be seen as a bad thing, but considering that it is becoming harder to obtain and retain personnel for many companies, it may also provide a way out.
The big challenge will be to sustain training programmes for the specialized surveyors if the volume drops even lower than it is today.