Every turbine you have ever looked at keeps its generator in the same place. Blades turn a shaft, the shaft runs back into a housing, and the generator sits inside that housing making electricity. The arrangement only works because something heavy is holding the housing still while the blades try to twist it out of position.<\/p>\n
A small Dutch company decided that the housing is the expensive part. In July it lowered a scale model of its alternative into the Marsdiep, the tide race between the island of Texel and the mainland at Den Helder, and on August 25 it said how the week went. The model dived on command. It surfaced on command. It started and stopped its rotor and it made electricity.<\/p>\n
Equinox Ocean Turbines calls the machine Mobula 5, a 1:10 model of a turbine it wants running at full size by 2030. In its announcement<\/a>, chief executive and CTO Pieter de Haas said the tests “confirmed key aspects of the system’s design and performance.” That puts the design at Technology Readiness Level 6, the rung where a prototype has to work in a real environment instead of a tank.<\/p>\n Picture the full-size machine. Two blades, one rotor, 164 feet (50 meters) tip to tip, turning slowly the way a ship’s propeller turns slowly when it is barely under way.<\/p>\n Slow blades produce enormous torque and very few revolutions per minute. A generator that has to eat torque instead of speed gets big, gets expensive, and usually needs a gearbox in front of it to make the numbers work at all. Every marine turbine developer runs into this. Most of them accept the gearbox and the giant nacelle and move on.<\/p>\n Equinox went the other way. It hung a small three-bladed turbine off the tip of each big blade, and those two little turbines are the only things on the machine that generate anything.<\/p>\n The tip of a 164-foot rotor moves quickly even when the rotor itself is loafing. Fast water arriving at a small rotor means high speed and low torque, which is what a cheap generator wants. Equinox has described the big blades as concrete, picked to handle the loads and vibration a long, slim underwater blade lives with. Concrete is an odd word to find next to the word turbine, and I get why it stops people, but the loads down there are nothing like a wind blade’s.<\/p>\n I had to read the American paperwork twice on this one, because it is stranger than the press release makes it sound.<\/p>\n The Department of Energy has a program called TEAMER, short for Testing Expertise and Access for Marine Energy Research, which hands federal lab time to marine energy companies. Equinox got a slot. The work went to the National Laboratory of the Rockies in Golden, Colorado, the lab that used to be called NREL, and the report is public<\/a>.<\/p>\n Here is the line that matters. Inside the hub of that big rotor, the lab found a bearing and nothing else. Zero watts leave the middle of the machine. Both tip turbines together account for every electron.<\/p>\n Which means the big concrete rotor is a delivery mechanism. It exists to drag two small generators through the water faster than the water is actually moving.<\/p>\n The lab spends more time on the second consequence. Almost no reaction torque builds up at the center of the machine, so the whole thing can hang from a single-point mooring instead of a fixed seabed foundation. Put a conventional turbine on that same mooring and the rotor just spins its own housing around instead of turning against the water, which the report says rules out generating anything at all. One anchor point instead of a piled foundation is a lot of money in deep water, and it is arguably the real product here.<\/p>\n The modeling itself did not go smoothly, which is unusual to see written down this plainly.<\/p>\n Hannah Ross, Will Wiley and de Haas ran high-fidelity computational fluid dynamics on the lab’s high-performance computer using Siemens STAR-CCM+. The mesh for the main rotor on its own came to 4.68 million cells, the tip turbine on its own to 3.16 million, and the combined machine to 20.59 million. The point of all that was to build a cheaper mid-fidelity model that Equinox could run on its own afterward, using the open-source OpenFAST code.<\/p>\n They swept 2,700 parameter combinations, narrowed to 115, then to 31, then to one. The combined-stage version still “was not successful in this stage of the project,” in the report’s words, and the full set of planned design load cases never got run.<\/p>\n They found the reason. The tip vortex shedding off the end of the big blade twists the water arriving at the little turbine, and the wake trailing behind the big blade punches a sharp velocity hole through it on top of that. Those two things decide how fast the small blades fatigue, how hard the drivetrain gets hit and whether the controller can hold a steady speed. A vortex model has no way to see the hole at all.<\/p>\n One number is missing from all of this. Equinox has not published a rotor diameter or a power rating for the Mobula 5 itself, and the naming runs in meters, so a tenth of 50 gets you roughly 16 feet. The company also said in July that the testing was happening in Friesland and in August that it happened at Marsdiep near Texel, which is in North Holland. Small stuff, but you notice it when you are trying to place a machine on a map.<\/p>\n Tidal races are violent. The Alderney Race off Normandy, where France’s six-turbine FloWatt farm is stalled on paperwork<\/a>, runs at around 11 mph. The Bay of Fundy, where Canada cleared a 680-ton floating turbine<\/a>, is worse.<\/p>\n The Gulf Stream is not like that. It runs closest to the American coast off southeast Florida, and in the stretch where turbine demonstrations have been staged, off Broward County and Palm Beach County, the water moves at roughly 3.5 to 5.5 mph and never stops. You get a fraction of the punch and none of the four-times-a-day slack. A design built to pay for itself at low flow speeds is interesting here for that reason specifically.<\/p>\n Equinox says it is chasing exactly that. Three regions are named in its plans: the American Gulf Stream, Japan’s Kuroshio and South Africa’s Agulhas. The company puts its opening addressable market at roughly 30 gigawatts, inside a worldwide resource it estimates at over 700.<\/p>\n American estimates for the home stretch are more modest. A federally funded Georgia Tech assessment set the theoretical ceiling across the whole Florida Current at about 5.1 gigawatts, or 45 terawatt-hours a year. Widen the extraction zone to everything within 200 miles of the shoreline and that ceiling rises to roughly 18.6 gigawatts. Nobody gets the ceiling. Florida Atlantic University’s marine energy center, renamed the Southeast Marine Energy Center on July 1<\/a>, holds a federal lease for floating test berths 13 miles off Broward County rated for prototypes up to 100 kilowatts. That is a two-percent-scale sandbox for a 2-megawatt machine.<\/p>\n Six people are listed on the Equinox team page, which is a small number of humans to be aiming at a 164-foot concrete rotor. The Colorado report closes with what still has to be modeled before anyone builds one: platform motion in six degrees of freedom, rotor speeds free to vary, moorings and non-uniform inflow. All of that, it says, is the subject of future proposals.<\/p>\n Equinox is planning the half-scale Mobula 25 for 2028 and the 2-megawatt Mobula 50 for 2030. If you want to check the American half of the story yourself, the report by Ross, Wiley and de Haas is dated March 26, 2026, and it has been sitting on the Department of Energy’s marine energy repository under a Creative Commons license since March 4.<\/p>\n","protected":false},"excerpt":{"rendered":" Every turbine you have ever looked at keeps its generator in the same place. Blades turn a shaft, the shaft … <\/p>\nSo why bolt the generators onto the blade tips?<\/h2>\n
The big rotor does not generate a single watt<\/h2>\n
A Colorado lab ran it through a supercomputer and hit a wall<\/h2>\n
The Gulf Stream is the slow one<\/h2>\n
Mobula 25 goes in the water in 2028<\/h2>\n