WEBVTT 1 00:00:01.480 --> 00:00:05.599 Welcome to the quart Side. Quantum Physics Podcast, an exploration 2 00:00:05.679 --> 00:00:10.919 of the fundamental structure of reality, where quantum laws govern matter, energy, 3 00:00:11.199 --> 00:00:16.239 and information. Here, uncertainty is a feature, not a flaw, 4 00:00:16.719 --> 00:00:19.239 and understanding begins at the smallest scales. 5 00:00:27.559 --> 00:00:30.719 You know that old joke that goes around the physics departments, 6 00:00:31.079 --> 00:00:33.240 the one where they say nuclear fusion is the energy 7 00:00:33.280 --> 00:00:35.880 of the future, and well it always will be. 8 00:00:36.000 --> 00:00:39.359 Oh yeah, it's basically been the ultimate technological punchline for 9 00:00:39.439 --> 00:00:40.320 decades now. 10 00:00:40.200 --> 00:00:42.920 Right, the promise was always, you know, thirty years away. 11 00:00:43.240 --> 00:00:46.039 It really felt like humanity was stuck on this grueling, 12 00:00:46.159 --> 00:00:49.320 exhausting run on a treadmill. We're sweating, we're putting in 13 00:00:49.359 --> 00:00:52.000 the work, but we're just we're never actually stepping forward. 14 00:00:52.240 --> 00:00:55.119 But something fundamental has shifted there. We are looking at 15 00:00:55.119 --> 00:00:58.399 a reality that was honestly once a joke. The era 16 00:00:58.520 --> 00:01:02.000 of limitless energy isn't thirty years away anymore. I mean 17 00:01:02.079 --> 00:01:04.439 it is happening right now, in twenty twenty five and 18 00:01:04.439 --> 00:01:05.879 twenty twenty six exactly. 19 00:01:06.239 --> 00:01:08.840 So we really need to look at how human engineering 20 00:01:09.120 --> 00:01:12.480 finally caught up to the stars, right, Like, why we 21 00:01:12.560 --> 00:01:15.480 have suddenly stepped off that treadmill and just tread sprinting forward. 22 00:01:15.560 --> 00:01:21.159 Well, the transition from these theoretical physics experiments into actual 23 00:01:21.159 --> 00:01:25.120 commercial power plant development is just moving at a blistering pace. 24 00:01:25.239 --> 00:01:27.079 Right now, the mat has finally flipped. 25 00:01:27.280 --> 00:01:30.000 Okay, but before we get into the wild milestones that 26 00:01:30.120 --> 00:01:32.560 just happened, we should probably set a quick baseline for 27 00:01:32.599 --> 00:01:34.799 you listening, just to make sure we are all looking 28 00:01:34.840 --> 00:01:36.159 at this through the right lens. 29 00:01:36.359 --> 00:01:38.959 Yeah, that's a good idea, because you know, you hear 30 00:01:39.000 --> 00:01:41.879 the word nuclear and a lot of people immediately picture 31 00:01:42.359 --> 00:01:45.159 massive concrete cooling towers. 32 00:01:44.879 --> 00:01:50.200 Right, glowing green rods, centuries of radioactive waste. But that's fission. 33 00:01:50.640 --> 00:01:54.079 Fusion is a completely different beast. 34 00:01:54.040 --> 00:01:57.120 It really is. Fission is the process of taking heavy, 35 00:01:57.280 --> 00:02:00.840 unstable atoms like uranium or plutonium and forcing them to 36 00:02:00.879 --> 00:02:04.560 split apart. It releases a tremendous amount of energy, but 37 00:02:04.640 --> 00:02:08.680 the leftovers are highly radioactive, long lived waste. 38 00:02:08.360 --> 00:02:10.879 And it has that chain reaction element right exactly. 39 00:02:11.280 --> 00:02:14.080 Because fission relies on a chain reaction, there's this inherent 40 00:02:14.120 --> 00:02:16.400 physical risk of it running out of control and causing 41 00:02:16.439 --> 00:02:20.520 a meltdown. But fusion goes in the exact opposite direction. 42 00:02:20.400 --> 00:02:23.120 So instead of splitting heavy atoms, you go the other way. 43 00:02:23.000 --> 00:02:25.080 Right, You take the lightest atoms in the universe and 44 00:02:25.159 --> 00:02:29.159 smash them together. We use different isotopes or variations of hydrogen, 45 00:02:29.439 --> 00:02:32.159 specifically deuterium and tritium. 46 00:02:31.639 --> 00:02:33.800 And when you force them to combine. 47 00:02:33.360 --> 00:02:36.520 They fuse into helium and in the process they release 48 00:02:36.599 --> 00:02:40.159 a massive amount of clean energy along with a stray neutron. 49 00:02:40.560 --> 00:02:42.800 So we're essentially trying to build a tiny star right 50 00:02:42.879 --> 00:02:43.439 here on Earth. 51 00:02:43.599 --> 00:02:47.199 It's the exact same physical process. Yeah, inside the Sun, 52 00:02:47.240 --> 00:02:50.879 immense gravity is the crushing force that pushes hydrogen atoms 53 00:02:50.919 --> 00:02:54.159 together until they fuse. But well, Earth doesn't have the 54 00:02:54.159 --> 00:02:58.439 gravity at the Sun, thankfully, right, thankfully. So to compensate 55 00:02:58.479 --> 00:03:01.240 for that lack of pressure, have to heat the hydrogen 56 00:03:01.280 --> 00:03:04.599 isotopes to temperatures much hotter than the Sun's core. We're 57 00:03:04.639 --> 00:03:08.039 talking over one hundred million degrees celsius. 58 00:03:08.159 --> 00:03:10.000 That is just it's hard to even wrap your head 59 00:03:10.000 --> 00:03:11.159 around that number. It is. 60 00:03:11.319 --> 00:03:13.719 At that temperature, the gas turns into a soup of 61 00:03:13.759 --> 00:03:17.400 electrically charged particles called plasma, and if you can hold 62 00:03:17.400 --> 00:03:21.439 that plasma together long enough and tight enough, fusion happens. 63 00:03:21.039 --> 00:03:23.680 And the byproduct is just helium right exactly. 64 00:03:23.800 --> 00:03:27.080 There are no carbon emissions, no long lived radioactive waste, 65 00:03:27.080 --> 00:03:30.560 and the fuel can literally be extracted from seawater and lithium. 66 00:03:31.000 --> 00:03:36.439 It's the holy grail of physics, clean, abundant, virtually limitless energy. 67 00:03:36.759 --> 00:03:39.319 And the whole reason it was thirty years away for 68 00:03:39.360 --> 00:03:42.719 so long is because of this brutal mathematical reality called 69 00:03:42.800 --> 00:03:43.840 net energy gain. 70 00:03:44.479 --> 00:03:46.000 Right, the energy sink problem. 71 00:03:46.120 --> 00:03:48.479 Yeah, because it took vastly more energy to run the 72 00:03:48.520 --> 00:03:51.319 machines that heat and squeeze the plasma than the plasma 73 00:03:51.360 --> 00:03:54.719 ever gave back. But these recent milestones have actually crossed 74 00:03:54.759 --> 00:03:55.840 that threshold. They have. 75 00:03:56.240 --> 00:03:58.960 The turning point for net gain really arrived via a 76 00:03:59.000 --> 00:04:02.960 few distinct and engineering approaches. One of the primary methods 77 00:04:03.000 --> 00:04:05.599 is called inertial confinement fusion. 78 00:04:05.360 --> 00:04:06.680 And that's the one with the lasers, right. 79 00:04:06.840 --> 00:04:10.080 Yes. The flagship facility for this is the National Ignition 80 00:04:10.159 --> 00:04:14.560 Facility or NIFF out in California. Their approach relies entirely 81 00:04:14.599 --> 00:04:17.839 on lasers. In April twenty twenty five, they fired their 82 00:04:17.879 --> 00:04:20.120 laser ray at a tiny pellet of fuel. 83 00:04:20.399 --> 00:04:23.040 Okay, so dropping a pellet, not a steady cloud, right. 84 00:04:22.920 --> 00:04:26.360 A tiny pellet and the resulting fusion reaction produced eight 85 00:04:26.399 --> 00:04:29.879 point six megajewls of energy output in some of their shots. 86 00:04:29.879 --> 00:04:31.600 The energy that came out of the pellet was actually 87 00:04:31.680 --> 00:04:34.000 more than four times the energy of the lasers that 88 00:04:34.040 --> 00:04:34.639 went into it. 89 00:04:34.879 --> 00:04:37.000 I need to visualize this, So it's like trying to 90 00:04:37.000 --> 00:04:39.879 perfectly crush a basketball down to the size of a 91 00:04:39.879 --> 00:04:42.439 pee by hitting it with I don't know, one hundred 92 00:04:42.439 --> 00:04:45.680 and ninety two baseball bats at the exact same microsecond 93 00:04:45.680 --> 00:04:47.000 from every possible angle. 94 00:04:47.120 --> 00:04:51.600 That's a shockingly accurate analogy. Honestly, the precision required is 95 00:04:51.680 --> 00:04:54.680 just staggering a fraction of a nano er second off 96 00:04:54.920 --> 00:04:57.360 or slight imbalance in the laser energy, and the pellet 97 00:04:57.439 --> 00:05:01.600 just blows apart asymmetrically instead of imploding. Wow, but getting 98 00:05:01.600 --> 00:05:03.959 four times the energy out of that pellet proves that 99 00:05:04.000 --> 00:05:07.040 the fundamental physics of ignition are sound. 100 00:05:07.360 --> 00:05:09.560 Okay, let me pause you right there because I have 101 00:05:09.600 --> 00:05:11.480 an obvious question and I'm sure you listening, or what 102 00:05:11.519 --> 00:05:14.000 are the exact same thing? Okay, go ahead. If NIF 103 00:05:14.040 --> 00:05:17.040 is putting in one unit of energy and getting four 104 00:05:17.160 --> 00:05:20.319 units back, why aren't our houses powered by giant lasers 105 00:05:20.399 --> 00:05:22.519 right now? Like? Why is the grid still struggling. 106 00:05:22.639 --> 00:05:25.600 Ah, right, it comes down to the critical difference between 107 00:05:25.639 --> 00:05:31.079 plasmaet gain and overall system efficiency. NIF achieved plasma. 108 00:05:30.680 --> 00:05:33.600 Net gain, meaning just the pellet itself exactly. 109 00:05:33.879 --> 00:05:36.800 The fuel pellet itself produced four times more energy than 110 00:05:36.800 --> 00:05:39.480 the laser light that directly hit it. But to generate 111 00:05:39.519 --> 00:05:42.560 that laser light, NUFF relies on one hundred and ninety 112 00:05:42.560 --> 00:05:46.639 two stadium sized lasers built with nineteen nineties technology. 113 00:05:46.680 --> 00:05:48.839 Oh, so the power dramas be insane. 114 00:05:48.920 --> 00:05:52.680 It's enomous. The electricity required from the California grid just 115 00:05:52.720 --> 00:05:55.800 to charge up the capacitor banks and fire those massive 116 00:05:55.839 --> 00:05:58.759 lasers is staggering. When you zoom out and look at 117 00:05:58.759 --> 00:06:01.319 the whole building, the system is still operating at a 118 00:06:01.399 --> 00:06:02.560 massive energy loss. 119 00:06:03.079 --> 00:06:07.399 So NIFF is this incredible scientific coving ground. But it's 120 00:06:07.439 --> 00:06:11.839 not a blueprint for a practical, plug and play commercial power. 121 00:06:11.639 --> 00:06:13.600 Plant, right, it's not a power plant design. 122 00:06:13.759 --> 00:06:16.639 So the physics inside the tiny pellet works beautifully, but 123 00:06:16.680 --> 00:06:19.839 the heavy machinery surrounding it is terribly inefficient, which explains 124 00:06:19.879 --> 00:06:22.319 why there is a completely different approach to building a star, 125 00:06:22.639 --> 00:06:24.160 one that doesn't use lasers at all. 126 00:06:24.639 --> 00:06:28.480 Yes, magnetic confinement Instead of blasting Appeller, you use magnets 127 00:06:28.480 --> 00:06:32.079 to trap a super hot, continuous cloud of plasma inside 128 00:06:32.120 --> 00:06:35.519 a giant doughnut shaped chamber. These machines are called dokomax, 129 00:06:35.839 --> 00:06:36.160 and the. 130 00:06:36.120 --> 00:06:38.639 Performance leaps there in late twenty twenty five have been 131 00:06:38.680 --> 00:06:39.839 pretty phenomenal. 132 00:06:39.439 --> 00:06:43.399 Too, right, unbelievable really. China operates a tokamac called the 133 00:06:43.439 --> 00:06:48.160 East Reactor. They refer to it as their artificial Sun. Recently, 134 00:06:48.319 --> 00:06:52.439 East managed to maintain a statle plasma at extreme densities 135 00:06:52.480 --> 00:06:54.720 that pushed right past the Greenwald limit. 136 00:06:54.879 --> 00:06:58.000 The Greenwald limit that is essentially the theoretical traffic jam 137 00:06:58.040 --> 00:07:00.439 limit for plasma. Right. Basically, yeah, if you try to 138 00:07:00.480 --> 00:07:02.800 pack too many particles into the magnetic donut, they just 139 00:07:02.879 --> 00:07:06.360 crash into each other, destabilize, and the whole reaction fizzles out. 140 00:07:06.560 --> 00:07:09.680 That was the operating assumption for decades. It was viewed 141 00:07:09.720 --> 00:07:12.560 as this hard ceiling for how dense and powerful a 142 00:07:12.639 --> 00:07:15.680 tokemac could get. But East didn't just break that ceiling. 143 00:07:15.839 --> 00:07:19.079 They sustain those dense, high temperature pulses for over one 144 00:07:19.120 --> 00:07:20.120 thousand seconds. 145 00:07:20.160 --> 00:07:22.680 Wow, over one thousand seconds. 146 00:07:22.480 --> 00:07:25.160 Yeah, holding a violently hot star steady for over a 147 00:07:25.279 --> 00:07:27.360 quarter of an hour without it touching the walls or 148 00:07:27.360 --> 00:07:31.000 collapsing is just a monumental feet of control. France's West 149 00:07:31.040 --> 00:07:34.079 Reactor in South Korea's k Star are seeing similar breakthroughs. 150 00:07:34.160 --> 00:07:37.199 But even with these records, the traditional magnetic donut has 151 00:07:37.240 --> 00:07:41.319 a fundamental flaw which brings us to honestly the weirdest 152 00:07:41.360 --> 00:07:44.160 machine in the physics world, the Stellar Rator. 153 00:07:44.360 --> 00:07:45.480 It's a crazy design. 154 00:07:45.680 --> 00:07:49.199 It looks like someone took a normal tokamac and violently 155 00:07:49.240 --> 00:07:53.040 twisted it like a pretzel. It genuinely looks like modern art. 156 00:07:53.279 --> 00:07:55.439 Wait before we talk about the twist, why does a 157 00:07:55.480 --> 00:07:57.199 perfect donut leak in the first place. 158 00:07:57.519 --> 00:08:01.839 Well, in a perfectly symmetrical doughnut, the magnetic field isn't uniform. 159 00:08:02.439 --> 00:08:05.759 The magnetic coils are bunched closer together on the inner 160 00:08:05.800 --> 00:08:07.920 hole of the doughnut and spread further apart on. 161 00:08:07.879 --> 00:08:10.120 The outer edge, so it's stronger on the inside track. 162 00:08:10.240 --> 00:08:13.480 Exactly, the magnetic field is much stronger on the inside 163 00:08:13.560 --> 00:08:16.600 track than the outside track. Because of that imbalance, the 164 00:08:16.720 --> 00:08:20.120 charged plasma particles naturally want to drift outward, like. 165 00:08:20.079 --> 00:08:22.920 A car taking a corner way too fast and sliding. 166 00:08:22.519 --> 00:08:25.680 Toward the guardrail, precisely to stop the plasma from hitting 167 00:08:25.680 --> 00:08:28.639 the outer wall. A traditional tokmac has to induce a 168 00:08:28.759 --> 00:08:32.440 massive electrical current directly through the plasm itself to hold 169 00:08:32.440 --> 00:08:36.480 it together, and sustaining that internal current for long periods 170 00:08:36.679 --> 00:08:38.279 is incredibly difficult. 171 00:08:38.320 --> 00:08:40.159 So how does the twisted pretzel fix that? 172 00:08:40.519 --> 00:08:44.679 A stellarator like Germany's Vandelstein seven X or W seven 173 00:08:44.879 --> 00:08:49.519 X solves this outward drift with geometry by physically twisting 174 00:08:49.559 --> 00:08:53.960 the magnetic field coils into a highly complex, non axisymmetric shape. 175 00:08:54.240 --> 00:08:58.279 The magnetic track itself constantly alternates Oh icy yeah. A 176 00:08:58.320 --> 00:09:01.960 particle drifting outward on one curve is immediately twisted to 177 00:09:01.960 --> 00:09:05.679 the inside on the next curve. The geometry literally cancels 178 00:09:05.720 --> 00:09:07.080 out the drift, so. 179 00:09:07.000 --> 00:09:10.399 The plasma essentially holds itself together just by following the 180 00:09:10.440 --> 00:09:12.320 twisted track. You don't need to pump it full of 181 00:09:12.559 --> 00:09:14.399 internal electrical currents. 182 00:09:14.080 --> 00:09:17.799 Right, It makes the plasma inherently stable. Calculating the exact 183 00:09:17.840 --> 00:09:21.039 shape of those twisted magnets took massive supercomputers, but the 184 00:09:21.080 --> 00:09:22.440 results just speak for themselves. 185 00:09:22.519 --> 00:09:24.559 What did W seven X actually achieve? 186 00:09:24.799 --> 00:09:27.440 In twenty twenty five, W seven X achieved a record 187 00:09:27.480 --> 00:09:31.240 one point eight gigajewels of energy turnover, holding the plasma 188 00:09:31.240 --> 00:09:35.360 steady for extended periods with zero internal current required. It 189 00:09:35.480 --> 00:09:38.759 strongly suggests the stellar rator might be the ultimate architecture 190 00:09:38.799 --> 00:09:40.399 for a commercial power plant. 191 00:09:40.120 --> 00:09:41.960 Because it needs to run twenty four hours a day 192 00:09:42.000 --> 00:09:44.039 three hundred and sixty five days a year without interruption. 193 00:09:44.360 --> 00:09:47.039 But whether you are building a symmetrical doughnut or a 194 00:09:47.080 --> 00:09:52.240 twisted pretzel, there was this massive bottleneck. Historically, these magnetic 195 00:09:52.279 --> 00:09:55.919 reactors had to be colossal, literally the size of sports 196 00:09:55.919 --> 00:09:58.799 stadiums just to generate a strong enough field to squeeze 197 00:09:58.799 --> 00:10:02.679 the plasma. Yeah, was a huge issue. But recently a 198 00:10:02.759 --> 00:10:06.639 breakthrough in an entirely different field radically altered the physics 199 00:10:06.639 --> 00:10:10.919 of fusion high temperature superconducting magnets, or HTS. 200 00:10:11.200 --> 00:10:15.600 The arrival of HTS magnets completely rewrote the timeline in 201 00:10:15.639 --> 00:10:19.120 a magnetic confinement reactor. The strength of your magnetic field 202 00:10:19.159 --> 00:10:22.559 dictates how tightly you can squeeze the plasma. The tighter 203 00:10:22.559 --> 00:10:24.879 the squeeze, the hotter it gets, and the more fusion 204 00:10:24.919 --> 00:10:25.919 reactions occur. 205 00:10:25.879 --> 00:10:27.200 So it scales exponentially. 206 00:10:27.320 --> 00:10:30.759 Right, the scaling is exponential. For decades, we relied on 207 00:10:30.840 --> 00:10:34.159 low temperature superconductors. They worked, but they hit a hard 208 00:10:34.200 --> 00:10:36.519 physical limit on how strong a magnetic field. 209 00:10:36.320 --> 00:10:38.320 They could generate. And HTS changes that. 210 00:10:38.559 --> 00:10:43.480 High temperature superconductors can generate dramatically stronger magnetic fields without 211 00:10:43.519 --> 00:10:48.120 losing their superconducting properties. Because a stronger magnet squeezes the 212 00:10:48.159 --> 00:10:51.399 plasma so much more efficiently. You can radically shrink the 213 00:10:51.399 --> 00:10:52.840 physical size. 214 00:10:52.480 --> 00:10:54.679 Of the reactor, so you don't need a stadium. 215 00:10:54.919 --> 00:10:57.000 No, you can build machine the size of a high 216 00:10:57.000 --> 00:10:57.879 school gymnasium. 217 00:10:58.080 --> 00:11:01.039 And the moment that happened, the economics just changed overnight. 218 00:11:01.679 --> 00:11:03.960 If you can shrink the machine, you shrink the volume 219 00:11:04.000 --> 00:11:08.200 of concrete, the steel, the construction time, The price tag 220 00:11:08.360 --> 00:11:11.799 drops from the international lega project scale down to the 221 00:11:11.879 --> 00:11:15.600 venture capital scale. Absolutely, Wall Street realized this wasn't just 222 00:11:15.600 --> 00:11:17.840 a science experiment anymore. It was a commercial race. 223 00:11:18.039 --> 00:11:22.639 For a long time, fusion was synonymous with massive, slow, 224 00:11:22.799 --> 00:11:27.240 bureaucratic public efforts. The prime example is Eider, currently being 225 00:11:27.240 --> 00:11:27.679 built in. 226 00:11:27.639 --> 00:11:29.480 France, the big international one. Right. 227 00:11:29.600 --> 00:11:32.679 Eider is a marvel of human cooperation. You know, thirty 228 00:11:32.720 --> 00:11:36.039 three nations working together to build a towering, low temperature 229 00:11:36.080 --> 00:11:41.440 superconducting tokomac. They recently got their final central solenoid modules, 230 00:11:41.480 --> 00:11:45.320 the massive magnetic pillars, but it's slow. It is fundamentally 231 00:11:45.360 --> 00:11:48.519 a research facility. Its goal is to produce five hundred 232 00:11:48.559 --> 00:11:51.519 megawatts of power, but not until the late twenty thirties. 233 00:11:51.840 --> 00:11:53.279 It's methodical and slow. 234 00:11:53.759 --> 00:11:56.960 But because of HTS magnets, private start is realized. They 235 00:11:56.960 --> 00:12:01.399 could just lap the massive public projects. The agility is insane. 236 00:12:01.480 --> 00:12:04.759 Right now, we're seeing over ten billion dollars invested globally 237 00:12:05.000 --> 00:12:07.799 across more than one hundred and sixty different private facilities. 238 00:12:07.879 --> 00:12:11.240 The speed of the private sector is breathtaking. Take Commonwealth 239 00:12:11.240 --> 00:12:14.480 Fusion Systems or CFS, a company that's spun out of MIT. 240 00:12:15.080 --> 00:12:19.039 They are aggressively leveraging HTS magnets to assemble a compact 241 00:12:19.080 --> 00:12:22.720 to COMAC called Spark near Boston, and their timeline is aggressive, 242 00:12:23.000 --> 00:12:25.919 very They are aiming for first plasma by twenty twenty seven. 243 00:12:26.159 --> 00:12:28.360 Their goal isn't just to turn the machine on. They 244 00:12:28.399 --> 00:12:31.200 intend to demonstrate true net energy gain and a machine 245 00:12:31.240 --> 00:12:32.440 a fraction of the size of. 246 00:12:32.360 --> 00:12:34.919 Iter, and Spark isn't even the end goal, right, it's 247 00:12:34.960 --> 00:12:38.279 just the prototype. They are already designing the actual commercial 248 00:12:38.279 --> 00:12:41.679 power plant called IRC targeted for the early twenty thirties 249 00:12:41.720 --> 00:12:42.639 near Richmond, Virginia. 250 00:12:42.759 --> 00:12:45.759 Yeah, IRC would be a four hundred megawatt facility capable 251 00:12:45.759 --> 00:12:47.759 of powering hundreds of thousands of homes. 252 00:12:48.120 --> 00:12:51.279 And to move that fast, they aren't relying on trial 253 00:12:51.279 --> 00:12:54.399 and error. They're partnering with Siemens in Nvidia to build 254 00:12:54.440 --> 00:12:59.279 AI digital twins, simulating and optimizing every single mellimeter the 255 00:12:59.320 --> 00:13:02.799 reactor in a virtual space before they ever bolt the 256 00:13:02.799 --> 00:13:03.480 steel together. 257 00:13:03.720 --> 00:13:07.159 The integration of the tech industry goes even deeper than simulation. 258 00:13:07.639 --> 00:13:10.799 Look at companies like Helium Energy out in Washington State. 259 00:13:11.279 --> 00:13:15.240 They are taking a completely different magneto in inertial approach 260 00:13:15.320 --> 00:13:15.919 to fusion. 261 00:13:16.200 --> 00:13:17.799 But the backing is the interesting part. 262 00:13:17.919 --> 00:13:21.120 Right, The most telling detail is their backing. Helium is 263 00:13:21.159 --> 00:13:24.240 heavily funded by Sam Altman, the CEO of Open AI. 264 00:13:24.759 --> 00:13:27.799 They already have a signed power purchase agreement to supply 265 00:13:27.919 --> 00:13:31.919 Microsoft with fifty megawatts of fusion electricity by twenty twenty eight. 266 00:13:32.000 --> 00:13:35.240 Which perfectly sets up the massive reality driving this entire 267 00:13:35.279 --> 00:13:38.799 gold rush. Why is big tech suddenly pouring billions into 268 00:13:38.799 --> 00:13:41.759 building stars. I look at the current state of artificial 269 00:13:41.759 --> 00:13:46.240 intelligence like a brilliant but endlessly hungry teenager who has 270 00:13:46.279 --> 00:13:48.279 just moved into the global power grid's house. 271 00:13:48.360 --> 00:13:49.320 That's a good way to put it. 272 00:13:49.360 --> 00:13:51.799 And this teenager is eating absolutely everything in the fridge. 273 00:13:52.320 --> 00:13:54.440 The grids simply cannot handle the appetite. 274 00:13:54.679 --> 00:13:57.799 No, it can't. The energy demands of AI data centers 275 00:13:57.799 --> 00:14:01.159 are excluding at a rate that traditional utili cannot accommodate. 276 00:14:01.720 --> 00:14:04.480 A standard server rack in a normal data center might 277 00:14:04.559 --> 00:14:06.519 draw five to ten kilowatts of power. 278 00:14:06.600 --> 00:14:07.840 Pretty manageable, right. 279 00:14:07.759 --> 00:14:11.159 But an AI optimized rack packed with high end GPUs 280 00:14:11.240 --> 00:14:14.759 processing neural networks draws a staggering fifty to one hundred 281 00:14:14.759 --> 00:14:15.679 and fifty kilowatts. 282 00:14:15.720 --> 00:14:18.320 It's an order of magnitude hire and hyperscale. 283 00:14:18.360 --> 00:14:21.519 AI facilities containing thousands of these racks are being built 284 00:14:21.639 --> 00:14:24.799 all over the globe. Projections show that US data centers 285 00:14:24.799 --> 00:14:28.480 alone could consume nine percent of the nation's total electricity 286 00:14:28.519 --> 00:14:29.960 by twenty thirty nine percent. 287 00:14:30.039 --> 00:14:33.559 That places immense strain on existing power grids. It does. 288 00:14:34.080 --> 00:14:37.080 Wind and solar are crucial, but they are intermittent. They 289 00:14:37.120 --> 00:14:40.320 rely on weather and daylight. An AI data center requires 290 00:14:40.360 --> 00:14:44.039 two hundred and forty seven unwavering base load power. If 291 00:14:44.039 --> 00:14:47.639 a power fluctuation causes a massive AI training run to fail. 292 00:14:47.799 --> 00:14:49.799 It costs millions of dollars. 293 00:14:49.480 --> 00:14:51.799 So fusion becomes the only grocery store. They can keep 294 00:14:51.840 --> 00:14:54.559 up with that hungry teenager without boiling the planet in 295 00:14:54.559 --> 00:14:58.639 carbon emissions. It's compact, it's dispatchable, and it's completely zero 296 00:14:58.720 --> 00:15:03.279 carbon d theoretic drop a four hundred megawatt fusion plant 297 00:15:03.840 --> 00:15:06.360 right next to a hyperscale data center in the middle 298 00:15:06.399 --> 00:15:09.639 of a desert, completely independent of the weather or local 299 00:15:09.679 --> 00:15:13.519 coal plants. Tech giants are funding fusion because they realized 300 00:15:13.559 --> 00:15:16.639 it's a critical infrastructure requirement for their own survival. 301 00:15:17.039 --> 00:15:21.240 The terrestrial application is driving the funding absolutely, but the 302 00:15:21.399 --> 00:15:25.080 extreme physics required to boil water for AI servers is 303 00:15:25.200 --> 00:15:29.559 inadvertently unlocking capabilities that extend far beyond the electrical grid. 304 00:15:29.720 --> 00:15:33.360 This is where things transition from highly practical to something 305 00:15:33.399 --> 00:15:35.960 straight out of a science fiction novel. Let's talk about 306 00:15:35.960 --> 00:15:40.120 the deep space implications first. Because chemical rockets are hitting 307 00:15:40.120 --> 00:15:40.960 a wall. 308 00:15:40.759 --> 00:15:44.000 They really are. In rocketry, we talk about specific impulse. 309 00:15:44.159 --> 00:15:45.720 You can think of it as the miles per gallon 310 00:15:45.759 --> 00:15:48.759 rating of a space ship engine. Chemical rockets have terrible 311 00:15:48.799 --> 00:15:50.159 specific impulse. 312 00:15:49.759 --> 00:15:51.399 So they burn too fast right. 313 00:15:51.320 --> 00:15:53.200 To go fast, you have to burn a massive amount 314 00:15:53.240 --> 00:15:55.360 of heavy fuel, which means your ship has to be 315 00:15:55.440 --> 00:15:58.080 huge just to carry the fuel required to move the fuel. 316 00:15:58.240 --> 00:16:01.519 It's a vicious cycle, and solar power becomes practically useless 317 00:16:01.519 --> 00:16:04.120 for propulsion or life support once you get past the 318 00:16:04.240 --> 00:16:05.000 orbit of Jupiter. 319 00:16:05.320 --> 00:16:07.320 It's like trying to cross the Pacific Ocean in a 320 00:16:07.399 --> 00:16:10.840 rowboat and bringing all your own drinking water. The weight 321 00:16:11.000 --> 00:16:11.679 just kills you. 322 00:16:12.000 --> 00:16:15.360 A fusion reactor solves this entirely. Concepts like the direct 323 00:16:15.360 --> 00:16:19.039 fusion drive being developed by companies like Pulsar Fusion change 324 00:16:19.039 --> 00:16:20.480 the math of the solar system. 325 00:16:20.600 --> 00:16:21.360 How does it work? 326 00:16:21.639 --> 00:16:26.080 A fusion rocket provides incredibly high specific impulse, meaning it 327 00:16:26.200 --> 00:16:30.159 uses a tiny mass of fuel to generate massive sustained thrust. 328 00:16:30.840 --> 00:16:34.559 But crucially, because it's a power plant, it also provides 329 00:16:34.600 --> 00:16:39.200 abundant onboard electricity for heavy scientific instruments and crew life support. 330 00:16:39.480 --> 00:16:42.159 So if we swap chemical explosions for a fusion drive, 331 00:16:42.600 --> 00:16:44.600 how much faster does the solar system become. 332 00:16:44.799 --> 00:16:47.960 A fusion drive could potentially cut travel times to Mars 333 00:16:48.120 --> 00:16:50.639 from the standards seven or eight months down to mere 334 00:16:50.679 --> 00:16:53.639 weeks or a couple of months. It completely opens up 335 00:16:53.639 --> 00:16:54.679 the Outer Solar System. 336 00:16:54.759 --> 00:16:56.519 That's game changing for NASA. 337 00:16:56.360 --> 00:17:00.440 Highly interesting to them. Yes, because it enables sustained ecpsleration. 338 00:17:00.840 --> 00:17:03.159 You could send crude missions out to the asteroid belt, 339 00:17:03.440 --> 00:17:06.480 or establish operations to mine helium three from the upper 340 00:17:06.519 --> 00:17:08.559 atmospheres of gas giants. 341 00:17:08.200 --> 00:17:10.559 Like Jupiter, which could then be shipped back to fuel 342 00:17:10.640 --> 00:17:14.720 even more advanced fusion reactors. Okay, reducing Mars to a 343 00:17:14.759 --> 00:17:18.640 month long trip is unbelievable, But even interstellar travel isn't 344 00:17:18.680 --> 00:17:21.440 the wildest scientific byproduct emerging right now. 345 00:17:21.519 --> 00:17:22.200 No, it's not. 346 00:17:22.519 --> 00:17:24.480 I want to talk about the concept coming out of 347 00:17:24.480 --> 00:17:27.240 the University of Cincinnati because when I realized what they 348 00:17:27.279 --> 00:17:31.519 were proposing, my jaw actually dropped. They are suggesting that 349 00:17:31.559 --> 00:17:35.359 we can use these commercial fusion power plants to hunt 350 00:17:35.640 --> 00:17:36.400 for dark matter. 351 00:17:36.680 --> 00:17:40.319 This represents a breath taking overlap between energy production and 352 00:17:40.440 --> 00:17:44.799 fundamental particle physics. We have the Standard Model of Physics, 353 00:17:44.839 --> 00:17:48.319 which is basically our current, incredibly successful rulebook for how 354 00:17:48.319 --> 00:17:49.720 the sub atomic universe works. 355 00:17:50.039 --> 00:17:51.079 But it's missing something. 356 00:17:51.160 --> 00:17:53.839 It has a glaring hole. It cannot explain dark matter. 357 00:17:53.920 --> 00:17:56.119 Dark matter makes up about eighty five percent of all 358 00:17:56.200 --> 00:17:58.799 the matter in the universe. We know it exists because 359 00:17:58.839 --> 00:18:01.960 we can observe its heavy grap vitational pole holding galaxies together, 360 00:18:02.400 --> 00:18:05.759 but it's entirely invisible. It rarely, if ever, interacts with 361 00:18:05.880 --> 00:18:07.039 ordinary matter or light. 362 00:18:07.640 --> 00:18:10.319 So how on earth is a machine built to power 363 00:18:10.359 --> 00:18:13.920 Microsoft Data Center help us find the invisible skeleton of 364 00:18:13.960 --> 00:18:14.519 the universe. 365 00:18:14.839 --> 00:18:18.480 It comes down to the extreme environment inside the reactor wall. 366 00:18:19.000 --> 00:18:23.079 In a deuterium tritium fusion reaction, a massive flux of 367 00:18:23.160 --> 00:18:27.319 high energy neutrons is thrown outward from the plasmak. To 368 00:18:27.400 --> 00:18:30.319 capture that energy, the interior walls of the reactor are 369 00:18:30.480 --> 00:18:34.319 lined with what are called lithium breeding blankets. The neutrons 370 00:18:34.400 --> 00:18:38.480 smash into the lithium, generating intense heat to make electricity. 371 00:18:38.680 --> 00:18:39.920 Right, that's the power part. 372 00:18:40.000 --> 00:18:43.759 But theoretical physicists realize that in that chaotic high energy 373 00:18:43.759 --> 00:18:47.440 collision zone, the neutrons hitting the reactor walls could trigger 374 00:18:47.559 --> 00:18:52.519 incredibly rare nuclear processes. These processes might generate a hypothetical 375 00:18:52.559 --> 00:18:54.079 particle called an axion. 376 00:18:54.400 --> 00:18:57.279 Wait, an axion, I've seen that term thrown around in 377 00:18:57.319 --> 00:19:00.000 theoretical physics papers, but I thought it was completely hypothetical. 378 00:19:00.400 --> 00:19:03.119 Are you saying a power plant could just accidentally manufacture them? 379 00:19:03.240 --> 00:19:06.559 Yes, axions are ultra light particles that currently stand as 380 00:19:06.559 --> 00:19:09.079 one of the leading candidates for what dark matter actually is. 381 00:19:09.440 --> 00:19:13.359 And the massive continuous neutron bombardment inside a commercial fusion 382 00:19:13.400 --> 00:19:17.319 plant could inadvertently turn the reactor into a giant axion factory. 383 00:19:17.559 --> 00:19:19.359 So let me make sure I'm wrapping my head around this. 384 00:19:20.039 --> 00:19:24.079 We build a machine to generate electricity for AI algorithms. 385 00:19:24.759 --> 00:19:28.240 Inside that machine, we recreate the core of a star, 386 00:19:29.200 --> 00:19:32.240 and as a completely accidental byproduct of the exhaust hitting 387 00:19:32.279 --> 00:19:35.240 the wall, we might be spewing out the very particles 388 00:19:35.240 --> 00:19:37.880 that make up eighty five percent of the universe. We 389 00:19:37.920 --> 00:19:40.640 could just set up a dedicated dark matter detector right 390 00:19:40.720 --> 00:19:42.480 next door in the parking lot to catch them. 391 00:19:42.599 --> 00:19:46.920 Pretty much. Yeah. It transforms a notoriously difficult theoretical physics 392 00:19:46.960 --> 00:19:52.160 problem into a practical experimental pathway. A commercial plant doubles 393 00:19:52.200 --> 00:19:55.960 as a particle accelerator, exploring physics beyond the standard model. 394 00:19:56.039 --> 00:19:56.680 That's wild. 395 00:19:57.079 --> 00:19:59.799 It echoes how early nuclear fission reactors in the nineteen 396 00:19:59.839 --> 00:20:04.599 five provided profound insights into quantum mechanics simply because scientists 397 00:20:04.640 --> 00:20:08.200 finally had a machine that produced enough localized radiation to study. 398 00:20:08.319 --> 00:20:11.200 That is, awe inspiring. We are hunting for the fundamental 399 00:20:11.279 --> 00:20:14.519 nature of reality as a side hustle to powering server farms, 400 00:20:15.039 --> 00:20:18.160 but as incredible as axion factories in Mars rockets are. 401 00:20:18.240 --> 00:20:20.519