0:00 The Nasdaq just dropped one-point-four percent on Friday.0:03 The reason wasn't a Fed announcement.0:06 It was a single AI model launched out of Beijing.0:09 Last week, we flagged Moonshot AI’s rumored three trillion parameter model.0:14 Well, Kimi K3 is here, it’s 2.8 trillion parameters, and it just sent a shockwave through the U.S.0:20 tech market.0:21 Nvidia shares dropped two-point-two percent.0:24 This is what a real shift feels like.0:27 But that’s not the only thing moving.0:29 Here's what else you need to know.0:31 First, Washington is finally reacting.0:34 The White House just unveiled the Genesis Mission, a major new AI funding initiative.0:39 Chief science adviser Michael Kratsios is calling it a "new era for U.S.0:44 science." It's a strategic pivot designed to accelerate American R&D and counter exactly this kind of international pressure.0:53 The question is whether federal grants can move faster than open-source releases.0:58 Second, the money keeps flowing into specialized AI.1:01 CARPL.ai, a medical imaging platform, just locked down ten million dollars in Series A funding.1:08 This isn't about chasing trillion-parameter models.1:11 It's about the unglamorous, high-impact work of deploying AI in places like hospitals, especially in emerging markets.1:19 This is where AI stops being theoretical and starts saving lives.1:23 And third, another top researcher has left a major lab to build.1:27 Miles Wang is out at OpenAI and is launching a new AI drug discovery startup.1:32 He's reportedly seeking two hundred million dollars at a two billion dollar valuation.1:38 The angle is smart: repurposing existing drugs, not discovering new ones from scratch.1:44 It's a faster path to revenue, and it signals a maturation of the AI-for-science space.1:50 So let's go back to Kimi K3.1:51 Because this is the story that changes the map.1:55 This isn't just another model.1:57 We're talking 2.8 trillion parameters.1:59 It uses a Mixture-of-Experts architecture with sixteen active experts.2:04 For perspective, that makes it the largest open-weight AI system EVER released.2:09 And that 'open-weight' part is the real weapon here.2:12 On July 27th, Moonshot says it will release the full model weights.2:17 That means any researcher, any company, any country can download it, dissect it, and build on top of it.2:24 This isn't a demo behind an API.2:26 It's a direct challenge to the closed-system dominance of the big U.S.2:31 labs.2:31 And the market understands this.2:33 That 2.2 percent dip in Nvidia stock wasn't panic.2:36 It was a calculation.2:38 A world where frontier-scale models are open and free is a world where the chokehold on high-end compute is less absolute.2:46 It threatens the entire value chain.2:48 Moonshot AI's own valuation just jumped to 31.5 billion dollars, up fourteen-fold since 2024.2:54 They're preparing for an IPO that could value them at fifty billion.2:59 The money is following the shift.3:01 Now, here's the catch.3:02 There are always catches.3:04 The performance claims are still just that—claims.3:07 Independent verification won't happen until those weights drop.3:12 There's a report from Tech Times citing a 51 percent hallucination rate on coding tasks.3:17 And in a telling move, Moonshot had to pause new consumer subscriptions right after launch.3:23 Why?3:24 They ran out of GPUs.3:25 They built the engine, but they don't have enough fuel to run it at scale.3:30 This is the new battlefield.3:32 It's not just about who has the smartest researchers.3:35 It's about who has the compute, the data, and the nerve to put a two-point-eight-trillion-parameter model on the internet for anyone to grab.3:45 Here's what is different today.3:47 The frontier of AI is no longer a walled garden controlled by a few American labs.3:52 A nearly three-trillion-parameter model is not a research paper anymore.3:57 It’s a downloadable file.3:59 And that changes EVERYTHING.
