Seen from across Walvis Bay, the windswept patch of Atlantic Ocean known as Speed Spot is barely more than a sparkle of whitecaps against a long, low sandbar. As we get closer to what is one of the world’s most perfect speed-sailing areas, I scan the shore. It’s featureless save for two small shelters. We motor our zodiac toward the remote beach until we have to kill the outboard and tilt it up to spare the prop. The five of us jump overboard into the waist-deep water, following our guide, Paul Larsen, who is wading toward the shore. The wind howls in our faces, blowing so much sand that it runs down the beach in rivulets, like rain across a windshield. We climb up on the beach, jellyfish at our feet as thick as paving stones. “This is it. This is the Bonneville Salt Flats of speed sailing!” Larsen shouts, gesturing to the water just off the sandbar. The flying sand sticks to our teeth, turning the insides of our mouths to 600-grit with every word. “We’ll have to shovel out the timing hut,” Larsen says, peering into the primitive shelter he built years ago and pointing out animal tracks inside. “Jackal,” he concludes.

There is a shipping port on the far side of the bay, but over here the landscape is so desolate, so extreme, that we could be on an alien planet—Frank Herbert’s Arrakis, George Lucas’ Tatooine. In fact, we are in Namibia, a roughly Texas-sized country at the southwest corner of the African continent. Walvis Bay is one of the Atlantic’s great natural harbors, but it’s surrounded by emptiness: 31,000 square miles of desert. The dunes march right into the sea, setting up an elemental cycle that repeats itself nearly every day of the antipodal summer. Mornings break as clear and sunny as a Baywatch shoot, but in the afternoon, near-gale-force winds descend on the bay. The desert heat meeting the cool Benguela Current coming up from the Cape of Good Hope creates a powerful natural wind machine. It arrives like clockwork, steady and relentless. “No ruffles,” Larsen says, feeling the wind with his hand. The featureless landscape—no vegetation, no terrain, no fences, no buildings apart from the shelters—makes for perfectly organized air. “Attached flow,” he calls it, using the jargon of an aerodynamicist evaluating a successful wind-tunnel test.

Larsen is originally from Australia, but he searched the world for years to find this spot, a perfect natural runway to test a sailboat so radical that it is more at home in an airplane hangar than in any harbor—a futuristic craft that, if he can make it work, will not only capture the outright world speed-sailing record but also open up a new, no-limit era in sailing. “A hundred knots, maybe?” Larsen speculates from inside one of the huts, looking through a sandblasted window at the watery speed-sailing course just beyond the beach. The current record is just over 50.

Floating behind the zodiac is the boat that has brought Larsen to Speed Spot for the tenth time in as many years in pursuit of sailing’s speed record: the Vestas SailRocket Mark 2. Its aeronautical DNA is obvious at a glance. There’s a rigid carbon-fiber “wing” that functions as a sail, an ultra-streamlined 40-foot-long “fuselage,” and even something like landing gear—three pod-shaped floats that keep the wing and fuselage above the chop. Yet what looks at first glance like a water-striding sailplane is, on closer inspection, pure crazytown. For one thing, its wing is inclined at a 30-degree angle to the water and is nowhere near the fuselage. Instead, it’s mounted on the end of a 30-foot-long beam. The pole is, in a sense, an odd sort of mast—except that it runs horizontally. On the opposite side of the boat is a bladelike carbon-fiber fin. Technically this is the keel, or as Larsen calls it, the foil. SailRocket’s foil sprouts from the side of its fuselage, then turns to cut 3 feet down into the water. Critical to any sailboat, a keel keeps a boat from blowing over—or, in this case, from flying away.

“She’s 50 percent plane, 50 percent boat,” Larsen explains. Indeed, if SailRocket were dropped from a great height, it would glide down rather than fall. Larsen designed in aerodynamic stability as a safety measure. “If for some reason she lost the keel at speed,” Larsen explains, “than she really would be a plane, wouldn’t she?” The prototype version of SailRocket, Mark 1, actually did take off into the air, and Larsen survived what may be the most spectacular crash in sailing history.

It was 2008 and he was at Speed Spot putting the Mark 1 through its paces when a gust got under the boat and launched it clear into the sky. The half-plane/half-boat hit an altitude of between 40 and 50 feet while cartwheeling through a flip—before crash-landing upside down and backward. “It just kept going up and up,” Larsen said at the time, “then it hit bloody hard on my head.”

Larsen is confident enough about the stability of his revised design, the Mark 2, that he included a passenger cockpit behind the driver’s seat. It’s never held a passenger, however. “I haven’t installed the seat yet,” Larsen says, “but I’m going to have to test it out sooner or later …” He cocks an eyebrow in my direction. “Would that be good for your story?”

In 1947, Chuck Yeager strapped himself into the experimental Bell X-1 “bullet with wings” and broke the sound barrier 8 miles above the Mojave Desert in Southern California. Larsen sees himself as following squarely in Yeager’s footsteps. To become the fastest sailor in the world, he’s going to have to break through the nautical equivalent of the sound barrier—the so-called 50-knot barrier (about 57 miles per hour). The name is something of a misnomer, because the barrier can drift slightly higher or lower depending on things like salinity and water temperature. But it doesn’t move too much, and when it kicks in, “it’s like driving your car into a wall,” Larsen says.

That wall is caused by a phenomenon, still not fully understood, called cavitation. It’s familiar to every powerboater who has ever over-revved a propeller. At a certain point the prop will begin to “cavitate,” the water around it literally starting to vaporize. There’s a classic high school physics experiment where you put a glass of water under a bell jar and then pump out the air. Eventually the water bursts into a boil. The physics are similar with cavitation: The pressure drops at the trailing edge of a spinning propeller, and the water around it boils. When a prop cavitates, it stops generating thrust and starts generating bubbles.

Boats rarely move through the water as fast as a spinning propeller, but when they do, they also cavitate. The enormous amount of drag created by bubbles attacking a boat’s keel will stop it from accelerating, in the same way that the piling up of sound waves can halt a plane’s acceleration. The hydrodynamics of the 50-knot barrier and the aerodynamics of the sound barrier are analogous—they’re both limits imposed by the physics of their respective mediums.

Yeager pierced the sound barrier with brute force: He used a rocket engine. Under the right conditions, the brute force method can also get one through the 50-knot barrier. This is not just theory. The Russians proved it with their supercavitating torpedo, the Squall. Darpa has doled out millions of dollars in R&D contracts for the Navy’s Underwater Express program, which aims to build a supercavitating submarine. Supercavitating bullets shot from underwater rifles have even cracked the speed of sound. But a sailboat that can cruise at over 50 knots? It’s absurd. There is no such thing.

Granted, some powerboats do go faster than 50 knots, but they’re simply sidestepping the cavitation problem. Speedboats, unlike sailboats, don’t have keels—they skip over the water instead of plowing through it. Most vessels, however, need some sort of keel in the water to stay upright, and sailboats need a keel to move. That keel functions as a hydrofoil, resisting the lateral pressure of the sail and creating forward motion. The challenge for Larsen is to somehow force a keel past 50 knots using wind power alone.

With a radical new keel shape, Larsen thinks he has cracked cavitation. If he’s right, he not only takes the speed record for sailing, he also opens the door to an entirely new class of travel. “Just like with the sound barrier, once you’re through, you’re through,” he says, “and the equation for doing 100 knots or greater will have been written.”

Back at Speed Spot, Larsen has just offered me a chance to ride shotgun on SailRocket’s test run, and I don’t have to be asked twice. I grab a GoPro camera and wade over to the boat. There is no ladder and no deck; the body of the boat is a long, low missile held a few feet above the water’s surface by its three floats. Getting to the cockpit is a challenge. It’s a matter of clambering up onto the front float, swinging a leg over the nose cone, and then using both hands to scooch down the fuselage, a couple of inches at a go.

As I clamber aboard Sailrocket, I feel like Slim Pickens astride the H-Bomb in Dr. Strangelove

SailRocket is painted orange in homage to Chuck Yeager’s Bell X-1, but as I slide along, it’s Slim Pickens astride the H-bomb in Dr. Strangelove that comes to mind. Eventually I reach the auxiliary cockpit and climb inside. I’m in a coffin-sized carbon-fiber tube, with control ropes and coaxial cables running through the inside. It’s a tangle—half rigging, half wiring—but most of it eventually makes its way from Larsen’s cockpit past me and then down into the blackness of the tapering fuselage. The ropes control most of the boat’s moving parts. They maneuver the wing in and out and engage the hinge of the scythe-like foil, lowering it to its proper depth in the water. The cables send information back: the angle of the wing, the angle of the rudder. Larsen’s cockpit is wired with digital readouts, since he can’t actually see either the wing or the rudder from where he sits up front. My cockpit has the GPS unit, which will track our speed.

Larsen straps himself in. SailRocket is probably the only sailboat in the world that comes with a five-point seat belt harness as standard equipment. He’s angled low, hands on a tiny steering wheel nearly between his knees, his helmet just peeking over the edge of the cockpit. He might as well be driving a top-fuel dragster. He looks out over the nose cone, which is angled slightly down like that of the Concorde jet, the better to keep the boat glued to the water. I position myself facing aft like a tail gunner so I can see how the boat works. The wing is on the port side of the boat (Larsen’s left, my right), and the foil is to starboard. Larsen and I are back-to-back. “The main thing,” he says over his shoulder, “is to stay clear of the ropes.”

As Larsen mentioned, there is no seat installed in my cockpit (much less a five-point harness), so I’m on my knees, being careful not to touch any of the lines. They’re vibrating with tension and a clear hazard. The cables I’m not so worried about, because a gremlin has already gotten into the wiring harness—Larsen’s wing and rudder readouts are on the fritz. He’ll be piloting this run blind, sailing by feel alone. As I crouch down, I consider what would happen if we ended up in one of the catastrophic crashes that plagued Mark 1. From where I sit, a high-speed bailout seems like a really bad idea. Hitting the beam to port would obviously break every bone in my body, while the foil on the other side looks like it could slice me in half.

While SailRocket is being dragged upwind and into position by the team on the zodiac, Larsen gives me the preflight briefing. The wind is blowing at 30 knots or so, which is ideal. However, my extra weight means that it may be hard to get SailRocket‘s front float to “plane”: That’s the point where buoyancy is replaced by hydrodynamic lift and SailRocket‘s floats stop bobbing in the water and start to skim across it. So there are no guarantees. But with a thumbs-up, Larsen signals to the zodiac to let us go, and he starts the launch procedure.

“The wing is fully stalled,” Larsen reports. “We’re being pulled along like a square rigger.” Since every element of SailRocket—the wing, the foil, the aerodynamic streamlining—is optimized for high speed, the real challenge to sailing it is getting the boat started. Moving the rudder doesn’t have much effect at low speed, yet to accelerate you need to get the boat pointed in the right direction.

“I’m at full lock with the steering,” Larsen says after a frustrating minute or so. “There’s just too much drag to let the nose come around.” In other words, there’s too much cargo for this bird to fly. We’ve drifted nearly halfway down the course in a dud start. I’m wondering about lunch. But Larsen says, “Let’s try it again.”

Paul Larsen designed and built SailRocket with little more than a high school education. He was raised in the rural southeast corner of Australia. “We were thick in the bush—30 acres, no phones, no TV, and stinking-hot summers,” Larsen says. “We grew up climbing trees and going down wombat holes.” But his favorite thing to do was to play in the damned-up catchment that held the family water supply. “One day Dad got a piece of wood, sharpened the nose, stuck an ice cream container on it, and sat it in the water. It blew to the other side,” he recalls. “That was it.” The 10-year-old Larsen was hooked.

“The amount of stuff you can learn about sailing from just a sharp piece of wood,” he marvels. “Put a big sail rig on it: It falls over! Put a sail rig on the front or the back and watch it turn away. Put a heavier keel on and it sinks. Use a lighter keel, but bend it out at an angle so you have a canting keel …” When Larsen gets going, he talks a nautical mile a minute. “Put a keel on it with a little wing off a model airplane—that made sense to me.”

The obsession grew from there. In high school in the ’80s he raced the newfangled Hobie Cats and reached the national championships in 1984, followed by the world championship in 1985. “I would hitchhike down to the coast with my wet suit on Friday night,” he says, “and sail all day Saturday and Sunday.” Postgraduation, various odd jobs financed a life of freelance adventuring—big-wave surfing, hang gliding, ultralighting. Eventually Larsen’s motormouth and hail-fellow personality landed him an unpaid gig delivering one of the first maxi catamarans, an 86-foot-long version of a Hobie, to Japan. A strategy of simply turning up at the right place and time, ready to work—”with a paintbrush in hand,” as Larsen puts it—got him more, culminating in a paid job crewing another maxi-cat in “The Race,” the first-ever nonstop, round-the-world sailing event. The hick from the sticks ended up as a professional sailor, “a yachtsman,” Larsen laughs.

At the end of the 20th century, the world of high-end offshore sailboat racing was an especially dangerous game. Multihull catamarans and trimarans were pushing out monohulls, carbon-fiber construction was supplanting fiberglass—and the experiments with new designs and materials led to a lot of failures. Get Larsen going and he’ll regale you for hours with harrowing stories about what it’s like to be rescued from a boat that’s breaking up in 50-foot seas and what happens when a crew starts to mutiny in the middle of the ocean. But the truth is that when things are going well, there’s a lot of free time during a crossing. “We’d be belting along in the Southern Ocean or parked in a high-pressure zone,” Larsen says, “and I’d be just sitting there, drawing out versions of speed-sailing boats.”

The basic design he kept returning to was one proposed 50 years ago by an American rocket engineer named Bernard Smith. It never caught on, and Smith died in relative obscurity. But Larsen stumbled across Smith’s book in the back of a dusty yacht chandlery just after high school. Discovering it, he says, was like “getting a book on jet engine flight from 1917.” Smith posed a question few had ever thought to ask: “Most sailors think of a sailboat as something that reaches up to grab a bit of air,” Larsen says, echoing Smith. “But why not build it like a plane that reaches down and grabs the water?”

In more technical terms, what Smith proposed was a sailboat design that aligned the forces of sail and keel so that they opposed each other more directly. From a physics perspective, every sailboat has two “wings”: the sail and the keel. Without the keel, a boat could only be pushed directly downwind. The keel permits it to run across the wind or even point up close to the wind, so the sail becomes, aerodynamically speaking, a wing. Pushed through the water by the sail, the keel also generates lift and thus, in hydrodynamic terms, is also a wing.

Yet virtually every sailboat ever made stacks sail over keel. The sideways lift of the wind on the sail is opposed by the sideways lift of the water pushing on the keel, and the boat is propelled forward. But at the same time there is also a rotational force. That “turning moment,” to use a term common to both sailing and physics, is why sailboats lean when they’re cruising. Monohulls “heel,” while multihulls “fly a hull.” Sailing is a balancing act. Too much sail will overpower and eventually capsize a boat; too little and it will never go anywhere.

Working from first principles, Smith proposed a radical redesign. With both an inclined mast and a similarly inclined keel, and neither keel nor sail anywhere near the hull, it looked bizarre. But from a physics perspective, Smith’s design is simplicity itself. Sail and keel are parallel, facing each other from opposite sides of the boat. Draw arrows through their respective centers of lift and they point directly at each other. With the forces so aligned, there is no turning moment. Since the boat can’t tip, it can’t dump wind. More wind means more power. And more power means more speed. “Let it go,” Larsen says, and “it’ll rev until it blows up.” Writing in obscurity back in 1963, Smith couldn’t imagine anything holding together at speeds over 40 knots (46 mph). But Larsen realized that if the boat were built with modern materials and technologies—carbon fiber instead of wood, a hard wing instead of a fabric sail—Smith’s design would be burly enough to bang up against the 50-knot barrier all day long.

While crewing on a maxi-cat that managed to set a new 24-hour speed record in the Pacific, Larsen decided that the next step was to form a speed team of his own. “It’s time,” he told himself. “I am the person to do this.” That was 10 years ago.

A high-speed bailout seems like a bad idea. Hitting the wing beam would obviously break every bone in my body.

It’s been a decade of monomaniacal sacrifice. For a time, Larsen was scrimping by, living in a steel shipping container to save money. But he has gradually attracted enough talent and sponsorship to build two boats based on Bernard Smith’s design. The first was SailRocket Mark 1. “For me it was going full circle, back to playing on the dam with little models and stuff.”

The Mark 2 is the first to be kitted out with the radical new foil that Larsen himself helped design to beat cavitation. The job of the boat is to drag the foil up to the 50-knot barrier. The job of the foil is to glide past it. Fifty knots is the place where it’s all going to happen—or not.

Back in SailRocket after our dud start, Larsen looks out at what remains of the watery runway. He’s humoring me by trying again, and this time, to our surprise, the launch procedure goes as planned. We’re being blown sideways off the course, so Larsen loosens the mainsheet. The force of the wind, with a little help from the rudder, slowly brings the nose of the vessel around so that we’re pointing straight toward the beach. The wing is now pointed at the wind—attached flow. She moves forward.

At first our progress is excruciatingly slow. The front float is plowing, and the trick foil is producing buckets of drag. But after what seems an eternity, that front float pops up on a plane at about 8 knots. This is the first crucial transition: Static lift gives way to hydrodynamic lift. SailRocket is now at its natural ride height and has dumped its first load of drag. The sound changes from a kind of glug-glug to a swoosh. There’s a bit of spray.

We’re accelerating. With the foil down and the wing working, the pedal is to the metal. The boat has no choice but to surge forward. The swoosh of water becomes louder still, like the static between TV channels, volume turned all the way up. “It’s got its skates on,” Larsen says. We’re running straight for the shore at nearly 20 knots. The sandbar is dead ahead.

Larsen drops the mainsheet and grabs the wheel with both hands. He’s concerned, but avoiding the sandbar is not his first priority. Now that he’s got SailRocket started, his race instincts take over. His aim is to get it to go as fast as possible. He turns the wheel slightly to port so that we’ll avoid hitting the beach; more important, it angles us closer into the wind so that the boat pulls harder.

Despite the fact that his instrument cluster isn’t working, Larsen carves the turn perfectly. We accelerate the whole way. Except for the lack of engine noise, I can’t believe we’re not in a speedboat. SailRocket is starting to vibrate, hard. The boat’s three pods are sending up three Jet Ski-like rooster tails of spray. And then, suddenly, there are only two rooster tails—the wing has pulled the rear pod completely free of the water. I know this is what’s supposed to happen at 26 knots—the second crucial transition—but it’s alarming to feel the boat surge ahead once again. We’ve just shed a full third of our hull drag.

Larsen completes the turn somewhere between 30 and 40 knots, just in front of the sandbar. We’re out of the chop now, and the flat conditions give us another burst of speed. The zodiac, which had been chasing us, has now fallen far behind. Another few seconds, and a glance to the boat’s opposite side reveals that we’ve made the third crucial transition: the point very close to 45 knots when the leeward pod, the one under the wing, lifts free and takes to the air. SailRocket is now more plane than boat. We’re bombing along on just the front pod and the foil. There’s not much more wetted area than you’d find on a boogie board, and yet cantilevered behind that front pod is a massive machine—40 feet long and 40 feet wide.

SailRocket again surges ahead. We’re pressed up against the 50-knot barrier. The wing is lifting so hard that the leeward pod is no longer flying a few inches above the water, as it should be, but is gyrating several feet in the air. The beam that connects the hard sail to the boat is warping and torquing. It’s whipping around so fast that the wing seems to be literally flapping.

My eyes, however, are glued to the foil. At slower speeds it’s the hydrodynamic equivalent of a brick, but Larsen’s fluid-dynamics modeling predicts that by the time we reach 50 knots, the flow of water will switch into a slippery mode. When that final transition happens, SailRocket will shed the last of its drag and leap into speeds unknown. I can hear it whine—an awful, unexpected sound like a cat yodeling in a blender—but I can’t see much. Everything is obscured by massive amounts of spray. The foil is pulsing rhythmically, blasting spindrift and sea foam into the air like a water cannon on full automatic.

I glance down at the GPS unit in front of me: 54.4 knots. This is no joyride, I suddenly realize. This is the real thing. Has he done it? Has he cracked the 50-knot barrier?

I look up just in time to see the photographer on the beach receding far into the distance. “Oh shit” I think, “I’ve missed my cue, my Facebook moment.” But figuring what the hell, I get up on my knees, one hand on an improvised saddle horn (the GPS unit) and the other waving in the air to give my best impression of Major Kong riding the H-bomb to oblivion: “Yeeeeehaw!”

And then we crash.

The speed course is about a mile long, and we have reached the shallow end of the lagoon. There are no brakes on a sailboat, so Larsen does the only thing he can: He cranks the wheel over hard and pops open the wing, attempting the sailing equivalent of a bootlegger’s reverse. As the boat lurches to starboard, the g-forces throw me to port and, fortunately, back into the cockpit. I see the world rotate around me as we burn off speed and then plow, with a carbon-fiber crunch, onto the shoal. We’ve run aground. The tip of SailRocket’s foil is jammed into the gravelly sand.

Hydrodynamically speaking, hydrofoils are all pretty much the same. Viewed in cross section, keels, propeller blades, and even fish usually have an elongated teardrop shape. Thick in front and skinny in back is just the most efficient design—up to a point. At around 50 knots, cavitation kicks in. And when it occurs it always attacks the trailing edge, because that’s where the pressure is lowest.

There are no breaks, so Larson cranks the wheel and pops open the wing. We crash anyway.

Larsen’s trick foil—the one we have just rammed into the ground—is different. It is V-shaped in cross section. It’s a simple wedge. At low speed, water generally flows around it just as one might expect. The seas are parted by the foil’s leading edge and then collapse again into a turbulent mess once it passes. At higher speeds, however, something interesting starts to happen. The foil actually digs a trench, creating, just for a split second, a hole in the water. The faster it goes, the deeper the hole.

The hole is important, because the water flows around it. The funny-looking wedge foil now has a tail—made of air. In cross section it suddenly takes on the familiar teardrop shape. But cavitation can’t take hold of this teardrop, because its trailing edge is made of nothing. That’s the theory, at least.

Larsen desperately needs to get his foil unstuck before it’s damaged or, worse, snaps. The wind is gusting, trying to spin the boat around, torquing the foil to its limit. “Get out! Get out of the boat!” he yells back to me. He dives over, desperately trying to lighten the load to float the boat free. I follow. It’s pure chaos in 3 feet of water until we’re able to wrest the foil out of the sand. Cranking the foil up to the surface on its hinge, Larsen sees that the leading edge of his precious foil is pitted and chipped. But when he glances at the GPS readout in my cockpit, despair turns to celebration. “Fifty-four!” he laughs, incredulous. “You’ve now gone as fast as I have ever gone in a boat.” The crash was nothing in light of the fact that his wedge-shaped foil worked its magic.

Data downloaded from the GPS show that we hit a top speed of 54.4 knots and raced at a sustained average of nearly 50 knots. And all that with a load of human ballast on a shortened course. SailRocket didn’t even have its full aero-optimization package on—and when we were near our maximum, I was hanging out the back, mugging for the camera. The sub-50-knot average means we haven’t actually broken the 50-knot barrier. The 54 knots was an instantaneous number, a brief Vmax. But we have darted to the other side and returned unharmed.

“This was the breakthrough run,” Larsen announces later that evening, raising a triple rum and Coke in one of many toasts to the team. “Now I know it can be done.”

After our wild ride together, Larsen’s luck goes cold. Although he makes run after run, he’s not able to repeat the 54-knot performance. More and more, our run begins to looks like an outlier, a black swan. Larsen does everything he can think of to egg SailRocket on. He tweaks the pitch of the foil, the angle of the wing, the ride height of the boat, the vector of the sails. Nothing works, and nobody can explain why getting back to 54 knots seems impossible.

Finally Larsen resorts to desperate measures. He takes out a hacksaw and cuts down his foil. Six inches lighter, SailRocket peaks at 48 knots: not good enough. Another 6-inch sacrifice and the speed is … still 48 knots. Finally, on the beach at Speed Spot, Larsen makes his last cut. With a full foot and a half gone from the business end of the foil, SailRocket looks like an amputee. Larsen’s idea is that at some point the sail will so overpower the foil that he can, in his words, “skull-drag it to a world record.” That doesn’t work either.

After three months of trying, he takes the foil, retreats to his home in Weymouth, England, and thinks—for almost a full year. His A-team of aerodynamicists who had modeled the forces on the boat and designed his foil were clearly wrong. Looking for a second opinion, Larsen calls on a noted hydrodynamicist who agrees to act as a B-team. “Oh God, that’s hopeless,” he says to Larsen after he sees the foil. “There’s no way air from the surface will suck down the back of that at high speeds.” If there is no air sucked from the surface, then there is no teardrop-shaped hole in the water. His diagnosis of the problem: good old cavitation attacking the back of the foil. The fix: a radically thinner V.

For months, Larsen bounces revised foil designs off his new adviser. Carbon fiber? “Thinner!” says the hydrodynamicist. Steel? “Thinner!” High-tensile tool steel? “Thinner still!” Until finally Larsen takes matters into his own hands. “Give me the numbers,” he says.

Working through the math himself, he discovers that according to his hydrodynamicist’s model, his old 54-knot keel couldn’t possibly have gone more than 28 knots. The A-team analysis, on the other hand, predicted that it would go at least 65 knots. Neither team could get it right. Digging deeper, Larsen figures out why. Underlying all their computational fluid-dynamics equations and models were the same handful of scientific papers. But the research led the two sides to vastly different conclusions. “Theorists can read a paper and get one thing out of it,” Larsen says, “and you can read it and get something totally different out of it.”

It dawns on Larsen that if the theorists can’t even agree on their theories, much less make an accurate prediction, than all the fluid-dynamics software in the world can’t design a high-speed foil that will work. “All these big computers, all these pretty pictures,” Larsen says. “But at over 50 knots they haven’t been verified. They haven’t been checked against anything real.” Garbage in, garbage out.

“The foil we used on that 54-knot run was the wrong size; it was the wrong shape,” Larsen tells me. But it did yield valuable data. “We put a point on the graph,” he says, “against which everyone can calibrate their machines, their mathematical equations, their theories.”

Post-calibration, the foil that Larsen’s A-team comes up with is still a thick V-wedge of carbon fiber, albeit subtly different in its angles and about half the size of the one that got stuck at the 50-knot mark. The B-team hydrodynamicist quits, not wanting to be associated with certain failure. And Larsen returns to Namibia to put the new foil to the test—this time in front of the World Sailing Speed Record Council, the official governing body of speed sailing.

Almost a year to the day after our harrowing run, Larsen is again floating at Speed Spot, strapped into SailRocket with the wind blowing a steady 30 knots. This time, however, he’s alone. This will be SailRocket‘s last run, and the judges are watching from the beach. They’re not interested in a fluky push through the 50-knot barrier; a Vmax doesn’t count. They’re looking for sustained speeds. They take an average over 500 meters. That is the definition of the outright speed world record.

The launch is ugly. The wind is whipping the water into such a frenzy that it’s actually swallowing the bottom part of the wing. “It was well underwater, and the whole boat was rolling to leeward like it was trying to capsize,” Larsen says afterward. “And then I heard something go pop.” Something had broken. It was time to abort. But when Larsen starts to turn back, the boat suddenly catches the wind and starts moving. With the acceleration, the leeward float pops out of the water, starts to plane, and pulls the bottom of the wing out of the drink. There is the first crucial transition.

To go or not to go? That is the question. “Was that carbon fiber?” Larsen asks himself. No, he decides. When something structural breaks, you can usually feel the boat shudder as the load gives way. He figures that what he heard was the wing’s delicate skin starting to rip. “Nothing looks too bad,” Larsen thinks after swiveling his head around to assess the damage. He decides to go for it. “No more dicking around. We’re out here to do 60.”

Even with the rip in the wing, the boat accelerates through one transition after another. The rear pod starts to fly at 26 knots. The leeward pod flies at 45. Because of the abortive start, he’s forced to take a line very close to the beach. It’s a nearly identical repeat of the run on which I was along for the ride. Again, the leeward pod starts to fly too high. “Is something really broken out there?” Larsen wonders, second-guessing himself. The steering feels loose, as if there is going to be a liftoff soon. As the boat keeps on accelerating through the 50-knot barrier, Larsen clenches: “Just hold it! Just hold it! Hold it! Hold! Holdholdholdhold it!”

At 58 knots, the leeward pod under the wing, which is supposed to just skim the water’s surface, is flying 4 or 5 feet in the air. “The only thing that would make it go that high is speed,” Larsen realizes. Stopping is going to be a problem. As in the 54-knot run, the late start forces him to bail out in shallow water. “I ran out of runway,” Larsen says, “but I was willing to risk it to bag that big number.” Luckily, he avoids grounding the boat on the shoal this time, thanks to the shorter foil and a higher tide. “I couldn’t have done it 20 minutes later,” he says.

Back on the beach, after checking the GPS, Larsen writes the digits in the sand, from right to left, in reverse order, in front of the rest of the team. Seven. Three. Dot. Five. Six. A record-smashing 65.37-knot run! Even Larsen can’t believe it, at first mistaking that 500-meter average for the peak speed. But the peak was 67.74 knots—about 75 miles an hour. (Later, after checking its own more precise GPS, the Speed Record Council set the official average for the run at 65.45 knots.)

“All the way home I’m just thinking of all the people who didn’t see this coming,” Larsen says. His revolutionary new keel shape means all the sailing records are up for grabs again: the round-the-world, the trans-Atlantic, the Transpac, the 24-hour. Larsen’s run makes the record-breaking monohulls, maxi-cats, and trimarans of the richest of the rich obsolete. “They’re old. Just old,” Larsen marvels. “Things are coming now that will put those boat designs with the square riggers.”

When I traveled to Namibia a year before, I had met a daredevil yachtsman with his eyes on a big prize. But now that he has it, all Larsen can talk about is the hydrodynamic frontier. “What we are discovering is that things are not as black-and-white going through this barrier as we thought they were,” he says. The mixture of air, vapor, and very high speed water wrapping itself around a boat at the limit is dynamic and extremely hard to model by computer or even simulate in a high-speed flow tank. To understand it, you must explore it directly, through experimentation. You need SailRocket.

Larsen is too modest to call himself a scientist. But when he talks he no longer refers to SailRocket as a boat—it’s a “laboratory.” Speed Spot is no longer a racecourse—it’s a “wind tunnel.” And he has been dreaming of a whole host of new foil designs. There are tiny ones that could slide along in a giant underwater bubble of vaporized seawater like the Russian torpedo does. There are ultrathin V’s that can plow a trough in such a way that one whole side of the foil would be dry at 50 knots. “We’ve had the test-pilot phase,” he says. And sailing has had its sonic boom.

Adam fisher (adamcfisher@gmail.com) wrote about the high tech scheme to televise the America’s Cup in issue 20.08.