WEBVTT 00:00:00.000 --> 00:00:04.040 Imagine the quiet hum of factories. Okay, but 00:00:04.040 --> 00:00:06.540 not on Earth, orbiting our planet miles above. 00:00:07.019 --> 00:00:10.259 Or think about an AI so advanced, it's not just 00:00:10.259 --> 00:00:12.679 smart, they're calling it mathematical superintelligence. 00:00:12.919 --> 00:00:14.580 Right. This week, you know, these aren't just 00:00:14.580 --> 00:00:17.160 ideas from some sci -fi book, they're real. Real 00:00:17.160 --> 00:00:19.559 investments happening right now. Welcome to the 00:00:19.559 --> 00:00:21.679 deep dive. We try to cut through all the noise, 00:00:21.800 --> 00:00:24.079 you know, the constant stream of info and find 00:00:24.079 --> 00:00:27.000 the signal. Our mission today. To unpack the 00:00:27.000 --> 00:00:29.859 latest wave of AI investments, specifically the 00:00:29.859 --> 00:00:32.979 big deals from July 4th to the 10th, 2025. It 00:00:32.979 --> 00:00:34.679 really does feel like riding a rocket, doesn't 00:00:34.679 --> 00:00:36.700 it? Yeah. The speed of AI funding is just, wow, 00:00:36.899 --> 00:00:40.200 nonstop, a wild ride. Absolutely. The scale, 00:00:40.240 --> 00:00:43.990 the pace of AI innovation. it's kind of staggering 00:00:43.990 --> 00:00:46.030 it's easy to get lost in it all but we're going 00:00:46.030 --> 00:00:48.009 to try and show you exactly where the smart money's 00:00:48.009 --> 00:00:50.890 going what it means for ai's reach you know from 00:00:50.890 --> 00:00:53.689 down here on earth all the way up to orbit we'll 00:00:53.689 --> 00:00:55.990 look at ai in space how it's changing industry 00:00:55.990 --> 00:00:59.890 tackling big global problems and then the really 00:00:59.890 --> 00:01:03.189 key stuff the basic tech making it all work let's 00:01:03.189 --> 00:01:07.209 uh let's dig in Okay, so first up, let's talk 00:01:07.209 --> 00:01:09.829 about AI pushing boundaries, like really pushing 00:01:09.829 --> 00:01:12.230 them beyond the usual stuff. Yeah, into places 00:01:12.230 --> 00:01:14.209 and ideas that were just theory not long ago. 00:01:14.269 --> 00:01:17.109 Exactly. First one, Varda Space Industries. They 00:01:17.109 --> 00:01:21.109 just pulled in a huge $187 million, Siri C funding. 00:01:21.409 --> 00:01:26.120 Wow. Yeah. These guys are pioneers in space manufacturing. 00:01:26.299 --> 00:01:28.280 Think about that. Making valuable stuff like 00:01:28.280 --> 00:01:30.780 advanced medicines or special metals in microgravity. 00:01:30.819 --> 00:01:32.700 Stuff you maybe can't make easily down here. 00:01:32.879 --> 00:01:35.299 Precisely. Either impossible or just way too 00:01:35.299 --> 00:01:37.400 inefficient with Earth's gravity. Now, the sources 00:01:37.400 --> 00:01:39.719 don't spell out their exact AI use, but you've 00:01:39.719 --> 00:01:41.680 got to figure. Oh, yeah. Orbital factories need 00:01:41.680 --> 00:01:44.620 serious automation. Robotics. Absolutely critical. 00:01:44.760 --> 00:01:47.859 AI would optimize everything. Handling materials 00:01:47.859 --> 00:01:52.640 in zero -g, keeping things sterile. It's complex, 00:01:52.939 --> 00:01:55.079 this investment. It's not just about rockets. 00:01:55.120 --> 00:01:58.459 It's about scaling unique off -world production. 00:01:58.959 --> 00:02:01.480 Could really change things back here. What gets 00:02:01.480 --> 00:02:04.879 me is just the scale of that vision. Whoa. Imagine 00:02:04.879 --> 00:02:07.840 your factory floor is literally in orbit. The 00:02:07.840 --> 00:02:10.400 amount of data, managing systems, predicting 00:02:10.400 --> 00:02:13.400 failures, optimizing yields. You're talking billions 00:02:13.400 --> 00:02:15.719 of queries, right? Terabytes of sensor data, 00:02:15.819 --> 00:02:18.300 plus orbital mechanics, communication delays. 00:02:18.560 --> 00:02:20.979 The kind of autonomous AI you need for that? 00:02:21.340 --> 00:02:23.060 It's kind of mind -boggling. Yeah, it really 00:02:23.060 --> 00:02:25.319 takes things to a whole other level. Almost dizzying. 00:02:25.699 --> 00:02:27.919 Okay, then there's Harmonic. Harmonic, right. 00:02:28.080 --> 00:02:30.620 They raised a big chunk, too, $100 million, Series 00:02:30.620 --> 00:02:32.219 B, for something they're calling mathematical 00:02:32.219 --> 00:02:34.560 superintelligence. I mean, when I first saw that 00:02:34.560 --> 00:02:36.860 term, I kind of stopped. What does it even mean 00:02:36.860 --> 00:02:39.219 for AI's future, mathematical superintelligence? 00:02:39.379 --> 00:02:41.360 That's a great question, isn't it? It suggests 00:02:41.360 --> 00:02:43.960 AI that's not just crunching numbers faster. 00:02:44.199 --> 00:02:46.479 It's about systems deeply grounded in, like, 00:02:46.500 --> 00:02:50.000 rigorous logic. Abstract reasoning. So thinking 00:02:50.000 --> 00:02:52.719 more like a pure mathematician. Kind of. The 00:02:52.719 --> 00:02:55.099 goal seems to be reaching new levels of understanding, 00:02:55.400 --> 00:02:58.120 generating complex mathematical proofs, maybe 00:02:58.120 --> 00:03:01.099 even new theories. Imagine an AI discovering 00:03:01.099 --> 00:03:04.259 entirely new math concepts, things humans might 00:03:04.259 --> 00:03:07.259 take centuries to find. Wow. This funding gives 00:03:07.259 --> 00:03:09.879 them a long runway for some seriously deep tech 00:03:09.879 --> 00:03:13.629 research. Pure abstract problem solving. That 00:03:13.629 --> 00:03:16.030 could ripple out everywhere into all sorts of 00:03:16.030 --> 00:03:18.189 science. That's a huge leap, almost philosophical. 00:03:18.490 --> 00:03:21.069 And then also in this sort of... pushing boundary 00:03:21.069 --> 00:03:25.069 space, Lightchain AI. Okay. They got $21 .1 million 00:03:25.069 --> 00:03:27.889 of sale bonus round. They're working on Web3 00:03:27.889 --> 00:03:30.270 AI infrastructure. For anyone maybe not deep 00:03:30.270 --> 00:03:32.449 in the weeds, Web3, that's basically decentralized 00:03:32.449 --> 00:03:34.550 networks, right? Like a more open distributed 00:03:34.550 --> 00:03:37.650 internet using blockchain tech. Right. And Lightchain 00:03:37.650 --> 00:03:40.629 AI is essentially building the foundational tools, 00:03:40.750 --> 00:03:43.870 the plumbing, you could say, for AI apps to actually 00:03:43.870 --> 00:03:46.409 work well on these decentralized networks. So 00:03:46.409 --> 00:03:48.330 it's more than just sticking an AI model on a 00:03:48.330 --> 00:03:50.319 blockchain. Oh, yeah. It's about But using those 00:03:50.319 --> 00:03:53.800 Web3 ideas like immutability, transparency for 00:03:53.800 --> 00:03:57.840 things like data integrity, auditing AI. Think 00:03:57.840 --> 00:04:00.539 decentralized data storage for models. Maybe 00:04:00.539 --> 00:04:03.139 open marketplaces for AI services where trust 00:04:03.139 --> 00:04:04.960 is built into the system. Instead of relying 00:04:04.960 --> 00:04:07.939 on one company. Exactly. They're laying the groundwork 00:04:07.939 --> 00:04:12.639 for a more open, verifiable AI future. So stepping 00:04:12.639 --> 00:04:15.189 back from the specific companies. beyond the 00:04:15.189 --> 00:04:17.709 buzzwords. What's the fundamental shift here 00:04:17.709 --> 00:04:21.910 when AI goes off world like Varda or into pure 00:04:21.910 --> 00:04:24.490 abstract math like harmonic? I think it's AI 00:04:24.490 --> 00:04:27.509 extending its reach into really complex, specialized 00:04:27.509 --> 00:04:29.910 places and also into like foundational human 00:04:29.910 --> 00:04:32.069 thinking. It's redefining the edge of what's 00:04:32.069 --> 00:04:34.240 possible. Okay, so now let's bring it back down 00:04:34.240 --> 00:04:35.560 to earth. Let's talk about the factory floor. 00:04:35.620 --> 00:04:38.339 Production lines, the operational centers of 00:04:38.339 --> 00:04:40.100 businesses. Right, practical impact. Yeah, where 00:04:40.100 --> 00:04:42.339 AI is making a huge difference right now. Streamlining 00:04:42.339 --> 00:04:44.319 things, boosting efficiency, automating the boring 00:04:44.319 --> 00:04:45.920 stuff so people can do more interesting work. 00:04:46.060 --> 00:04:48.500 Makes sense. Take MaintainX. They just raised 00:04:48.500 --> 00:04:51.839 a massive $150 million Series D for industrial 00:04:51.839 --> 00:04:56.540 AI. $150 million? Wow. Yeah. Their platform helps 00:04:56.540 --> 00:04:59.800 factories manage maintenance operations, but 00:04:59.800 --> 00:05:03.149 with incredible precision. Their AI can predict, 00:05:03.230 --> 00:05:05.829 like really accurately, when a machine might 00:05:05.829 --> 00:05:08.290 fail. Before a person would even notice. Often, 00:05:08.389 --> 00:05:11.149 yeah. Way before. This isn't just about smoother 00:05:11.149 --> 00:05:13.209 work orders. It changes how factories operate. 00:05:13.490 --> 00:05:17.009 From reactive fixing to proactive, almost self 00:05:17.009 --> 00:05:20.430 -optimizing systems. Imagine less costly downtime, 00:05:20.730 --> 00:05:23.860 more output. across a whole plant. That's real 00:05:23.860 --> 00:05:25.959 impact. That is a big shift. Okay, who else? 00:05:26.160 --> 00:05:27.459 Then there's Centific. They have $60 million 00:05:27.459 --> 00:05:30.100 in Series A. They run an AI data foundry. A data 00:05:30.100 --> 00:05:32.180 foundry. What's that exactly? It's not just collecting 00:05:32.180 --> 00:05:34.240 raw data. A data foundry is all about the hard 00:05:34.240 --> 00:05:36.939 work of cleaning data, structuring it, labeling 00:05:36.939 --> 00:05:40.000 it really precisely, especially large data sets. 00:05:40.180 --> 00:05:42.740 Ah, the fuel for the AI model. Exactly. It's 00:05:42.740 --> 00:05:44.360 essential, especially for big companies where 00:05:44.360 --> 00:05:46.910 data quality has to be top -notch. Centipede 00:05:46.910 --> 00:05:49.209 provides that crucial infrastructure, prepping 00:05:49.209 --> 00:05:52.110 massive, high -quality data for training custom 00:05:52.110 --> 00:05:55.810 AI. Got it. And LevelPath, $55 million, Series 00:05:55.810 --> 00:05:58.290 B. They've built an AI -native procurement platform. 00:05:58.750 --> 00:06:00.970 Procurement. Yeah. So buying stuff for businesses. 00:06:01.470 --> 00:06:03.649 Yeah, think about the old way. Mountains of paperwork, 00:06:03.870 --> 00:06:05.910 back -and -forth negotiations, checking vendors. 00:06:06.449 --> 00:06:09.550 LevelPath uses AI agents to automate a lot of 00:06:09.550 --> 00:06:12.170 that tedious process. Like an AI helping you 00:06:12.170 --> 00:06:14.310 shop for the company. Sort of. It helps find 00:06:14.310 --> 00:06:16.709 vendors, negotiate contracts, even figures out 00:06:16.709 --> 00:06:19.430 the best time to buy. It's about smarter, data 00:06:19.430 --> 00:06:21.750 -driven spending. It turns a cost center into 00:06:21.750 --> 00:06:23.730 something strategic. Interesting. What about 00:06:23.730 --> 00:06:27.709 voice AI? Good question. AOLA raised $25 million, 00:06:28.050 --> 00:06:31.329 Series A2, for enterprise AI and voice AI. So 00:06:31.329 --> 00:06:34.550 think intelligent assistants. They can transcribe 00:06:34.550 --> 00:06:37.269 meetings accurately, summarize them, pull out 00:06:37.269 --> 00:06:40.250 action items, or provide really advanced call 00:06:40.250 --> 00:06:42.800 center automation. Understanding customer issues 00:06:42.800 --> 00:06:45.560 better, resolving them faster, improves the whole 00:06:45.560 --> 00:06:47.720 customer experience. Okay, who else is in this 00:06:47.720 --> 00:06:50.019 industrial enterprise space? Foundation EGI, 00:06:50.339 --> 00:06:53.319 they got $23 million Series A focused on industrial 00:06:53.319 --> 00:06:56.560 AI and engineering. These guys use AI for really 00:06:56.560 --> 00:06:59.500 complex industrial processes, like chemical manufacturing 00:06:59.500 --> 00:07:02.259 or material science. They use AI modeling to 00:07:02.259 --> 00:07:04.000 improve product design through simulation testing, 00:07:04.160 --> 00:07:06.920 tons of variables virtually. Faster R &D. Right. 00:07:07.199 --> 00:07:08.939 And predictive maintenance on heavy machinery, 00:07:09.060 --> 00:07:10.980 again, cutting down costly failures. Nominal. 00:07:11.240 --> 00:07:16.480 $20 million Series A fintech and enterprise ERP. 00:07:16.839 --> 00:07:19.259 ERP, that's the big software running a company's 00:07:19.259 --> 00:07:21.740 operations, right? Finance, supply chain. Exactly, 00:07:21.879 --> 00:07:24.579 the operational backbone. Nominal is integrating 00:07:24.579 --> 00:07:28.000 AI into those ERP systems, building smarter ones 00:07:28.000 --> 00:07:30.519 that automate financial tasks, give much better 00:07:30.519 --> 00:07:33.220 forecasts, maybe even spot hidden problems across 00:07:33.220 --> 00:07:35.079 the whole company. Making the whole operation 00:07:35.079 --> 00:07:37.879 run smarter. That's the goal. And one more, Knox, 00:07:38.240 --> 00:07:41.819 $6 .5 billion Series A cloud infrastructure and 00:07:41.819 --> 00:07:44.540 GovTech. GovTech, technology for government agencies. 00:07:44.879 --> 00:07:47.740 Yep. Knox builds secure AI -powered cloud solutions, 00:07:47.920 --> 00:07:50.379 helps governments run more efficiently, boosts 00:07:50.379 --> 00:07:52.579 cybersecurity, which is huge these days, and 00:07:52.579 --> 00:07:54.279 ultimately deliver better services to citizens. 00:07:54.620 --> 00:07:57.759 Okay, so looking across all these, factories, 00:07:57.980 --> 00:08:00.470 finance, government. What's the common thread 00:08:00.470 --> 00:08:03.069 AI is weaving through their operations? It seems 00:08:03.069 --> 00:08:05.709 to be fundamentally about using data for smarter 00:08:05.709 --> 00:08:08.550 automation, making complex business stuff faster, 00:08:08.670 --> 00:08:11.579 more intelligent, and way more precise. All right, 00:08:11.579 --> 00:08:13.160 let's shift gears a bit. Let's talk about how 00:08:13.160 --> 00:08:16.920 AI investments are tackling some really big societal 00:08:16.920 --> 00:08:19.060 challenges, things like health, the environment, 00:08:19.319 --> 00:08:21.939 stuff beyond just the bottom line. Yeah, the 00:08:21.939 --> 00:08:23.959 impact can be really profound here. For sure. 00:08:24.079 --> 00:08:26.500 Take Renaissance Bio. They landed a big seed 00:08:26.500 --> 00:08:30.339 round, $54 .5 million, health tech and buyer 00:08:30.339 --> 00:08:33.139 tech. That's a lot for a seed round. It is. Companies 00:08:33.139 --> 00:08:35.879 like Renaissance Bio are using AI, machine learning, 00:08:36.039 --> 00:08:38.100 to speed up drug discovery. You know how slow 00:08:38.100 --> 00:08:40.679 traditional R &D can be. Glacial sometimes. Right. 00:08:40.820 --> 00:08:42.940 AI can analyze massive amounts of biological 00:08:42.940 --> 00:08:45.580 data, find potential drug targets, understand 00:08:45.580 --> 00:08:48.220 diseases better, even predict how patients might 00:08:48.220 --> 00:08:50.519 respond to treatments. This funding lets them 00:08:50.519 --> 00:08:53.139 push forward on really innovative R &D, could 00:08:53.139 --> 00:08:55.419 bring life -saving drugs to market much faster. 00:08:55.600 --> 00:08:57.759 That's huge. And kind of related, predictive 00:08:57.759 --> 00:09:00.799 oncology. They raised $10 million through a purchase 00:09:00.799 --> 00:09:03.299 agreement. They're using AI specifically in the 00:09:03.299 --> 00:09:06.179 fight against cancer, analyzing complex biodata 00:09:06.179 --> 00:09:08.940 genomics, proteomics, that stuff, to predict 00:09:08.940 --> 00:09:11.360 how different tumors will respond to drugs. So 00:09:11.360 --> 00:09:13.340 more personalized cancer treatment. Exactly. 00:09:13.519 --> 00:09:16.139 Guiding individual treatments, but also speeding 00:09:16.139 --> 00:09:18.399 up the discovery of totally new cancer medicines. 00:09:18.659 --> 00:09:21.730 It's AI powered. precision medicine. Very powerful. 00:09:21.929 --> 00:09:24.090 What about outside of health? Well, there's LG 00:09:24.090 --> 00:09:26.909 &DAI. They got $9 million in financing. They 00:09:26.909 --> 00:09:29.990 do geospatial AI and data analytics. Geospatial 00:09:29.990 --> 00:09:33.909 AI, that's analyzing location data, like satellites? 00:09:34.210 --> 00:09:36.450 Yeah, satellite imagery, drone footage, mapping 00:09:36.450 --> 00:09:38.730 data, all kinds of location -based info. It has 00:09:38.730 --> 00:09:41.389 huge applications. Think optimizing city planning, 00:09:41.610 --> 00:09:43.889 making farming more efficient, streamlining global 00:09:43.889 --> 00:09:46.169 logistics. Anywhere location matters. Pretty 00:09:46.169 --> 00:09:48.990 much. And related to that, Clear is geospatial. 00:09:49.419 --> 00:09:52.779 $8 .5 million Series A. They're applying geospatial 00:09:52.779 --> 00:09:55.240 AI directly to climate change. How so? Their 00:09:55.240 --> 00:09:58.600 AI uses high -res satellite data to monitor deforestation 00:09:58.600 --> 00:10:01.759 in real time, track carbon stored in forests 00:10:01.759 --> 00:10:05.139 and soils, analyze land use changes. Wow, like 00:10:05.139 --> 00:10:07.279 a planetary health check. Kind of, yeah. It provides 00:10:07.279 --> 00:10:11.399 crucial, actionable data for conservation, for 00:10:11.399 --> 00:10:13.740 carbon accounting, for making smarter climate 00:10:13.740 --> 00:10:16.299 policies. Really important stuff. So looking 00:10:16.299 --> 00:10:18.860 at these companies, Renison, Predictive Oncology, 00:10:19.039 --> 00:10:22.759 LGND, Cloris, beyond just making money, what 00:10:22.759 --> 00:10:25.299 kind of critical real world problems are they 00:10:25.299 --> 00:10:27.580 really trying to solve with this AI funding? 00:10:27.899 --> 00:10:30.639 I think they're aiming AI's power at scientific 00:10:30.639 --> 00:10:32.840 breakthroughs, speeding up medical innovation 00:10:32.840 --> 00:10:35.519 and giving us vital real time intelligence about 00:10:35.519 --> 00:10:38.059 our planet's health. OK, now let's talk about 00:10:38.059 --> 00:10:40.919 the hidden layer, the fundamental tech, the building 00:10:40.919 --> 00:10:42.500 blocks that make all these amazing applications 00:10:42.500 --> 00:10:45.379 possible, you know, from those space factories 00:10:45.379 --> 00:10:48.100 to the cancer drugs. The infrastructure underneath 00:10:48.100 --> 00:10:50.960 it all. Exactly. The digital nuts and bolts making 00:10:50.960 --> 00:10:53.580 the complex stuff easier, making things possible 00:10:53.580 --> 00:10:55.980 that weren't before. Makes sense. Where do we 00:10:55.980 --> 00:10:59.399 start? How about scale AI? They got a big one. 00:10:59.480 --> 00:11:01.399 Ninety eight point six million dollars in project 00:11:01.399 --> 00:11:03.679 funding. But this was different at Grant Public. 00:11:04.059 --> 00:11:06.559 private partnership. So not typical VC funding 00:11:06.559 --> 00:11:09.340 for company growth. Right. This funding is more 00:11:09.340 --> 00:11:12.720 about boosting the entire AI landscape. Think 00:11:12.720 --> 00:11:15.320 of it like investing in the highways or the power 00:11:15.320 --> 00:11:18.879 grid for AI. It helps push AI adoption in industries, 00:11:19.139 --> 00:11:21.320 strengthens the whole ecosystem for everyone, 00:11:21.480 --> 00:11:23.639 makes it easier for lots of other companies to 00:11:23.639 --> 00:11:25.580 innovate. Investing in the foundations, got it. 00:11:25.700 --> 00:11:29.539 Then there's Vellum, $20 million, Series A. They 00:11:29.539 --> 00:11:31.679 offer an AI development platform. And why are 00:11:31.679 --> 00:11:33.799 these platforms so important? They're crucial 00:11:33.799 --> 00:11:35.779 because they give developers and companies the 00:11:35.779 --> 00:11:38.200 tools they need. Integrated infrastructure, pre 00:11:38.200 --> 00:11:41.299 -built components, smoother workflows. It makes 00:11:41.299 --> 00:11:43.980 it much easier to experiment with AI, build models, 00:11:44.220 --> 00:11:46.700 deploy them, manage them carefully. Lowers the 00:11:46.700 --> 00:11:49.240 barrier to entry. Exactly. Reduces complexity, 00:11:49.639 --> 00:11:51.580 speeds up innovation everywhere. Okay. Who's 00:11:51.580 --> 00:11:56.580 next? Sundial. $16 million Series A. They have 00:11:56.580 --> 00:11:59.539 an AI -powered decision analytics platform. This 00:11:59.539 --> 00:12:01.559 goes beyond just showing you data on a dashboard. 00:12:02.240 --> 00:12:05.059 Sundial's AI analyzes information, and then this 00:12:05.059 --> 00:12:07.879 is key. It recommends specific actions. Ah, so 00:12:07.879 --> 00:12:09.580 it tells you what to do with the data. Yeah. 00:12:09.639 --> 00:12:11.440 It predicts the likely outcomes of different 00:12:11.440 --> 00:12:13.899 choices. Helps businesses make smarter, more 00:12:13.899 --> 00:12:15.820 competent decisions, especially when things are 00:12:15.820 --> 00:12:18.580 complex or the stakes are high. Actionable intelligence. 00:12:18.820 --> 00:12:21.720 That sounds very useful. And Castellum .ai. $8 00:12:21.720 --> 00:12:25.080 .5 million Series A. They're in RegTech and FinTech. 00:12:25.279 --> 00:12:28.139 RegTech. Regulatory technology. Helping companies 00:12:28.139 --> 00:12:31.000 follow the rules. Spot on. Especially in finance, 00:12:31.120 --> 00:12:33.019 where rules are complex and always changing. 00:12:33.679 --> 00:12:36.799 Castellum .ai probably uses AI to automate really 00:12:36.799 --> 00:12:39.879 critical tasks. Things like risk screening, checking 00:12:39.879 --> 00:12:42.000 against sanction lists, monitoring transactions 00:12:42.000 --> 00:12:44.639 in real time. Helps banks stay compliant and 00:12:44.639 --> 00:12:47.419 catch fraud without drowning in paperwork. Precisely. 00:12:47.679 --> 00:12:49.419 Keeps things safer and more efficient. Okay, 00:12:49.480 --> 00:12:51.879 a couple more infrastructure plays. Yeah, Cerebrium, 00:12:52.000 --> 00:12:55.000 $8 .5 million seed funding, AI infrastructure 00:12:55.000 --> 00:12:57.940 and cloud services. Companies like Cerebrium 00:12:57.940 --> 00:13:00.139 are kind of like the backbone for AI developers. 00:13:00.440 --> 00:13:02.500 They handle all the tricky stuff with managing 00:13:02.500 --> 00:13:04.799 cloud infrastructure behind the scenes. So developers 00:13:04.799 --> 00:13:07.440 can just focus on the AI model itself. Exactly. 00:13:08.059 --> 00:13:10.179 Cerebrium takes care of deploying the models, 00:13:10.340 --> 00:13:12.960 scaling them up or down as needed in the cloud, 00:13:13.139 --> 00:13:15.409 making sure they run well. Handles the heavy 00:13:15.409 --> 00:13:17.590 lifting. Makes sense. And the last one. Zero 00:13:17.590 --> 00:13:21.649 entropy. $4 .2 million seed round. Also AI infrastructure, 00:13:22.070 --> 00:13:24.990 but with a focus on RAG. RAG -J. Remind me what 00:13:24.990 --> 00:13:27.389 that is again. Sure. Retrieval augmented generation. 00:13:27.710 --> 00:13:30.190 It's a really powerful technique. Helps AI models 00:13:30.190 --> 00:13:33.350 avoid giving outdated or, well, sometimes just 00:13:33.350 --> 00:13:35.529 wrong information. How does it do that? It lets 00:13:35.529 --> 00:13:38.470 the AI fetch real -time, reliable info from outside 00:13:38.470 --> 00:13:41.190 sources, like trusted databases or the current 00:13:41.190 --> 00:13:43.350 internet, before it generates an answer. makes 00:13:43.350 --> 00:13:46.009 the AI's responses much more accurate, timely, 00:13:46.210 --> 00:13:48.570 and importantly, trustworthy. Okay, that sounds 00:13:48.570 --> 00:13:50.269 incredibly useful. You know, I got to admit, 00:13:50.350 --> 00:13:52.110 I still kind of wrestle with prompt drift myself 00:13:52.110 --> 00:13:54.450 sometimes, you know, when the AI just goes off 00:13:54.450 --> 00:13:56.590 on a tangent or makes stuff up. Yeah, hallucination 00:13:56.590 --> 00:13:58.669 is a real challenge. So seeing infrastructure 00:13:58.669 --> 00:14:02.490 being built specifically for more accurate, reliable 00:14:02.490 --> 00:14:06.809 AI models, using things like RG, that feels really 00:14:06.809 --> 00:14:09.129 promising. For AI, we can actually depend on. 00:14:09.230 --> 00:14:12.370 So looking at all these scale AI, Vellum, Sundial, 00:14:12.429 --> 00:14:14.830 the others, these core building blocks, the infrastructure. 00:14:16.029 --> 00:14:18.250 How do they fundamentally speed up the whole 00:14:18.250 --> 00:14:20.629 AI revolution we're seeing? Well, they're making 00:14:20.629 --> 00:14:23.490 AI fundamentally easier to build, much more efficient 00:14:23.490 --> 00:14:26.429 to deploy, and ultimately far more reliable for 00:14:26.429 --> 00:14:30.789 everyone, from tiny startups to huge global companies. 00:14:31.029 --> 00:14:33.370 What an absolutely incredible week for AI. Seriously, 00:14:33.669 --> 00:14:35.610 just seven days. And look at the sheer scale, 00:14:35.769 --> 00:14:38.149 the amazing diversity of these investments. That's 00:14:38.149 --> 00:14:40.409 pretty breathtaking. From Varda Space Industries 00:14:40.409 --> 00:14:42.649 planning factories in orbit, which could totally 00:14:42.649 --> 00:14:45.450 change manufacturing, to Harmonic going after... 00:14:45.450 --> 00:14:47.669 or mathematical superintelligence that might 00:14:47.669 --> 00:14:50.769 unlock, who knows, new levels of human understanding. 00:14:51.009 --> 00:14:53.029 The money flowing isn't just big. It's a huge 00:14:53.029 --> 00:14:55.490 statement about where we're heading. AI is really 00:14:55.490 --> 00:14:57.549 becoming omnipresent, isn't it? It's weaving 00:14:57.549 --> 00:14:59.929 itself into everything. We've seen it go from 00:14:59.929 --> 00:15:02.789 like the vacuum of space right down to the details 00:15:02.789 --> 00:15:06.250 of personalized medicine, optimizing factory 00:15:06.250 --> 00:15:09.049 floors, and even powering the very tools that 00:15:09.049 --> 00:15:12.049 help build AI itself more easily. Yeah, this 00:15:12.049 --> 00:15:13.850 isn't just about numbers on a spreadsheet. It 00:15:13.850 --> 00:15:16.710 really feels like a rapidly accelerating transformation 00:15:16.710 --> 00:15:20.269 of our world, driven by this constant wave of 00:15:20.269 --> 00:15:22.309 innovation. We really hope, you know, listening 00:15:22.309 --> 00:15:26.230 to this maybe encourages you to think about how 00:15:26.230 --> 00:15:28.909 these advances, these big investments we talked 00:15:28.909 --> 00:15:31.169 about today, might start touching your own life, 00:15:31.230 --> 00:15:33.129 your industry, maybe even your daily routine. 00:15:33.330 --> 00:15:35.230 It's probably closer than you think. Absolutely. 00:15:35.769 --> 00:15:39.639 And so thinking about this incredible pace. and 00:15:39.639 --> 00:15:41.879 the huge amounts of money going into really bold 00:15:41.879 --> 00:15:44.659 ideas like mathematical superintelligence or 00:15:44.659 --> 00:15:47.440 factories orbiting Earth, it leaves us with a 00:15:47.440 --> 00:15:49.440 pretty compelling question to ponder, doesn't 00:15:49.440 --> 00:15:52.159 it? What new challenges for humanity, or maybe 00:15:52.159 --> 00:15:55.279 what completely unforeseen opportunities, might 00:15:55.279 --> 00:15:57.659 pop up when AI lets us build things beyond our 00:15:57.659 --> 00:16:00.179 planet's gravity, or lets us reason at levels 00:16:00.179 --> 00:16:02.360 we can barely even comprehend today? Something 00:16:02.360 --> 00:16:02.940 to think about.