Secretary of War Pete Hegseth delivers the Technological Dominance at the War Department speech. (YouTube Video Transcript)

(00:00:00) Your YouTube transcript will appear here (00:00:01) Today, (00:00:02) we are done running a peacetime science (00:00:06) fair while our potential adversaries are (00:00:10) running a wartime arms race. (00:00:14) Right now, from garage startups to (00:00:16) factory floors to boardrooms across our (00:00:18) nation, the question of how to harness (00:00:20) innovation coming out of America's AI (00:00:23) ecosystem is front and center. and (00:00:25) rightfully so because it has the (00:00:27) potential to disrupt and transform every (00:00:30) area of human endeavor. The same is true (00:00:34) in the Department of War. President (00:00:37) Trump's AI executive order spells out (00:00:39) our approach succinctly. It is the (00:00:42) policy of the United States to sustain (00:00:44) and enhance America's global AI (00:00:46) dominance in defense of human (00:00:48) flourishing, economic competitiveness, (00:00:51) and national security. (00:00:54) That, ladies and gentlemen, is the War (00:00:57) Department's mission. We must ensure (00:00:59) that America's military AI dominance so (00:01:03) that no adversary can exploit that same (00:01:05) technology to hold our national security (00:01:08) interests or our citizens at risk. (00:01:11) America first in every domain. (00:01:16) Last month, I took the first step toward (00:01:18) changing how the department does (00:01:20) business with Frontier AI technologies (00:01:23) when we announced the rollout of Gen AI (00:01:25) with our partners from Google. And I (00:01:27) want to thank the Google team for (00:01:29) leaning forward and making the (00:01:30) investment to get their Gemini app to (00:01:32) about 3 million users in the war (00:01:34) department. (00:01:36) But today, we're excited to announce the (00:01:38) next frontier AI model company to join (00:01:41) Genai. Mill, and that is Grock from X (00:01:46) AI, which will go live later this month. (00:01:49) So, I want to thank you, Elon, and your (00:01:52) incredible team for leaning forward with (00:01:54) us on this as well. Very soon, we will (00:01:57) have the world's leading AI models on (00:01:59) every unclassified and classified (00:02:01) network throughout our department. long (00:02:05) overdue. (00:02:07) To further that, today at my direction, (00:02:10) we're executing an AI acceleration (00:02:12) strategy that will extend our lead in (00:02:14) military AI established during President (00:02:16) Trump's first term. This strategy will (00:02:19) unleash experimentation, eliminate (00:02:21) bureaucratic barriers, focus on (00:02:23) investments, and demonstrate the (00:02:25) execution approach needed to ensure we (00:02:28) lead in military AI and that it grows (00:02:30) more dominant into the future. In short, (00:02:33) we will win this race by becoming an AI (00:02:36) first warf fighting force across all (00:02:39) domains from the back offices of the (00:02:41) Pentagon to the tactical edge on the (00:02:44) front lines. (00:02:46) The catalyst for this acceleration will (00:02:48) be seven pace setting projects focused (00:02:51) on mission threads across war fighting, (00:02:54) intelligence, and enterprise missions. (00:02:57) each with a single accountable leader, (00:03:00) aggressive timelines, and measurable (00:03:01) outcomes that answer a familiar (00:03:04) question. Elon, (00:03:06) what have you accomplished this week? (00:03:10) This is the execution standard for AI (00:03:14) first transformation. (00:03:16) Each of the seven pace setting projects (00:03:18) will use the following model. One owner (00:03:22) who reports monthly on their progress. (00:03:24) These projects will not be run in a (00:03:26) vacuum, but will work directly with war (00:03:27) fighters and transition partners to (00:03:30) ensure we incorporate real time (00:03:32) operational feedback. We expect rapid (00:03:35) iterations with failure accelerating the (00:03:38) learning curve. These are not science (00:03:41) projects. They are not governance (00:03:43) boards. They are the execution standard (00:03:46) for the entire department. (00:03:48) We will not win the future by sprinkling (00:03:51) AI onto old tactics like digital pixie (00:03:54) dust. We will win by discovering (00:03:57) entirely new ways of fighting. That's (00:04:00) why we will run continuous (00:04:02) experimentation campaigns, quarterly for (00:04:05) forceonforce combat labs with AI (00:04:07) coordinated swarms, agent-based cyber (00:04:10) defense, and distributed command and (00:04:14) control. pushing the envelope, learning (00:04:16) from failure at every stop, which is (00:04:20) exactly what this place does. You see, (00:04:22) our department doesn't accept failure in (00:04:24) the past, and so we never fail, which (00:04:25) means we never learn. We're flipping (00:04:27) that dynamic. (00:04:30) Before talking about the new rules of (00:04:32) the game, let me talk about a little bit (00:04:34) more about our new team. Because no game (00:04:35) can be won without the right team. And (00:04:37) we're proud to announce that Mr. Cameron (00:04:39) Stanley has been appointed the new chief (00:04:41) digital and artificial intelligence (00:04:44) officer, CDAO, of our war department. (00:04:47) Cam will be leading a new team, many of (00:04:49) whom have foregone, thank you, or left (00:04:52) lucrative careers at pioneer companies (00:04:54) such as AWS, Data Bricks, Palanteer, and (00:04:57) Meta to join the fight. (00:05:00) This team will not only provide a (00:05:01) catalyst for change in this department, (00:05:04) but will also act, we believe, as a (00:05:07) magnet for other talented members of the (00:05:09) tech community who want to join us in (00:05:11) doing the mission focused work to (00:05:14) protect our great republic. (00:05:17) So, let's talk about the new rules. (00:05:19) First, speed. (00:05:22) Speed wins. Speed dominates. Our (00:05:26) enterprise currently operates on (00:05:28) staffing and committee cycles measured (00:05:31) in months and years, and that's (00:05:33) unacceptable. (00:05:35) Military AI is going to be a race for (00:05:37) the foreseeable future, where the risks (00:05:38) to US national security of moving too (00:05:41) slowly outweigh the impacts of imperfect (00:05:45) alignment. (00:05:47) To do this, Cam and his team at TD uh at (00:05:50) TDAO will define AI deployment velocity (00:05:53) metrics for all the pace setting (00:05:54) projects in the next 30 days and report (00:05:58) at least monthly after that. These will (00:06:01) become the new benchmarks for programs (00:06:04) across the department. (00:06:06) Second, bureaucratic blockers. If you (00:06:10) work with Elon, he you know he finds the (00:06:12) blockers and you remove them. We will (00:06:15) take a wartime approach to people and (00:06:17) policies that block this progress. (00:06:22) You want to block, you can work (00:06:23) somewhere else. Barriers to data (00:06:26) sharing, authority to operate, or at (00:06:28) test and evaluation and contracting are (00:06:31) now treated as operational risks, not (00:06:34) simply bureaucratic inconveniences. We (00:06:37) are blowing up these barriers. (00:06:41) That's why today at my direction, I'm (00:06:43) establishing a barrier removal SWAT team (00:06:46) under R& (00:06:48) authority to wave non-stutory (00:06:50) requirements and escalate to our great (00:06:52) deputy secretary Steve Fineberg anything (00:06:55) that slows down the acceleration of AI (00:06:58) capabilities. (00:07:00) Third, compute resource. We will invest (00:07:03) heavily in expanding our access to AI (00:07:05) compute from data centers to the (00:07:08) tactical edge and will tap into hundreds (00:07:10) of billions of dollars in private (00:07:12) capital flowing into American AI. (00:07:15) President Trump's executive order has (00:07:18) directed us to build data centers on (00:07:20) military land and to work with the (00:07:22) Department of Energy to ensure that we (00:07:24) dramatically increase the number and (00:07:26) breadth of resources needed to power (00:07:28) this computing infrastructure. (00:07:31) We will work together with our partners (00:07:33) at Google and AWS and Oracle and SpaceX, (00:07:36) Microsoft and others on these (00:07:38) initiatives. (00:07:40) Fourth pillar, (00:07:42) especially in this room, you'll (00:07:43) understand it, is talent. We will use (00:07:47) every hiring and pay authority available (00:07:49) to us to bring the best American (00:07:52) technical talent and reward effective AI (00:07:55) transformations by our workforce. We're (00:07:58) going to heavily leverage President (00:07:59) Trump's Tech Force initiative to bring (00:08:01) in the best and brightest from industry (00:08:03) and academia. (00:08:06) With people like Elon, David Saxs, Emil, (00:08:09) Mike, and others from the (00:08:10) entrepreneurial and business world (00:08:11) already in government, (00:08:14) we have shown that we can and that we (00:08:16) must enlist the world's leading talent (00:08:19) in this cause. (00:08:22) Fifth, responsible AI. Today, I want to (00:08:27) clarify what responsible AI means at the (00:08:30) Department of War. Gone are the days of (00:08:34) equitable AI and other DEI and social (00:08:38) justice infusions that constrain and (00:08:41) confuse our employment of this (00:08:42) technology. (00:08:44) Effective immediately responsible AI at (00:08:46) the War Department means objectively (00:08:49) truthful AI capabilities employed (00:08:52) securely and within the laws governing (00:08:54) the activities of the department. We (00:08:56) will not employ AI models that won't (00:08:59) allow you to fight wars. (00:09:04) We will judge AI models on this standard (00:09:06) alone. Factually accurate, mission (00:09:08) relevant, without ideological (00:09:10) constraints that limit lawful military (00:09:13) applications. (00:09:15) Department of War AI will not be woke. (00:09:18) It will work for us. (00:09:21) We're building war ready weapons and (00:09:22) systems, not chat bots for an Ivy League (00:09:26) faculty lounge. (00:09:28) Sixth, and finally, (00:09:31) data. AI is only as good as the data (00:09:33) that feeds it. And the US military has (00:09:36) an asymmetric data advantage from two (00:09:38) decades of military and intelligence (00:09:40) operations that no other military in the (00:09:42) world can replicate. (00:09:44) But right now, we are underutilizing (00:09:46) this advantage. Too much of our data is (00:09:49) stranded. It's stuck in bespoke program (00:09:52) databases, locked behind title 10 or (00:09:55) title 50 stove pipes, invisible to (00:09:58) operators, engineers, and industry who (00:10:00) can help us exploit it with winning (00:10:02) speed and scale. (00:10:04) And that's why today at my direction, (00:10:06) CDAO will exercise its full authority to (00:10:09) enforce the DOW data decrees and make (00:10:13) all appropriate data available across (00:10:16) federated IT systems for AI (00:10:18) exploitation, (00:10:20) including mission systems across every (00:10:22) service and component. (00:10:25) Each service secretary and component (00:10:27) head will submit cataloges of their (00:10:29) current data assets to the CDAO within (00:10:32) 30 days. (00:10:34) Denials of data access requests will be (00:10:36) reported to the CTO within 7 days and (00:10:40) they better have a good justification. (00:10:43) Today, I'm also directing the under (00:10:45) secretary for intelligence and security, (00:10:47) Brad Hansel, to ensure appropriate data (00:10:50) from across our intelligence enterprise (00:10:52) receives the same treatment and can be (00:10:54) fully leveraged to warfighting (00:10:56) capability development and operational (00:10:58) advantage. AI is only as good as the (00:11:00) data that it receives. and we're going (00:11:02) to make sure that it's there. (00:11:05) Persistent barriers to data access will (00:11:07) be escalated to the deputy secretary of (00:11:09) war for resolution with authority to (00:11:12) reassign or terminate personnel or (00:11:14) withhold funding from non-compliant (00:11:16) activities within the statutory limits. (00:11:20) We'll be clear here, as I said, data (00:11:23) hoarding is now a national security risk (00:11:26) and we will treat it that way. (00:11:31) AI is an important part of the future. (00:11:33) But here's another truth that we've (00:11:35) ignored for too long. Beyond AI, we've (00:11:38) treated every other kind of innovation (00:11:40) as if they're the same. As you know, (00:11:43) they're not. We need to break down (00:11:46) unnecessary barriers to rapid (00:11:48) technological development, adoption, and (00:11:51) transition. Some of you will remember (00:11:53) this. A generation ago, one of my (00:11:55) predecessors in a dinner speech to (00:11:58) industry, now infamously known as the (00:12:01) Last Supper, advocated for the (00:12:03) consolidation of our defense industrial (00:12:05) base. (00:12:07) This consol consolidation created a (00:12:09) closed innovation ecosystem dominated by (00:12:12) just a handful of prime contractors. (00:12:16) The results have been characterized by (00:12:18) soaring costs, sluggish delivery, and (00:12:21) stagnant innovation. (00:12:23) That's what President Trump's recent (00:12:25) executive order on the defense (00:12:27) industrial base and defense companies (00:12:29) seeks to address. (00:12:32) It makes crystal clear that the priority (00:12:35) of the legacy prime contractors must be (00:12:37) our nation's national security, (00:12:41) not the next earnings call. That means (00:12:44) less focus on stock buybacks and more (00:12:47) investment on the men and women on the (00:12:50) factory floor. (00:12:52) It means less stockholder dividends and (00:12:55) more investment in infrastructure, plant (00:12:58) and equipment. Today, that old era comes (00:13:03) to an end. (00:13:05) The Department of War is reopening to (00:13:08) the disruptive energy and agile (00:13:10) creativity of our nation's tech startups (00:13:13) funded by our world's leading capital (00:13:15) markets. I'm directing my chief (00:13:18) technology officer to lead this charge (00:13:20) and his wingman as always in this effort (00:13:22) under secretary for acquisitions and (00:13:24) sustainment Mike Duffy. (00:13:27) For too long we organized our ecosystem (00:13:30) around stages and silos, labs over here, (00:13:33) so-called rapid units over there, (00:13:36) commercial outreach in a different (00:13:38) building or on another coast altogether, (00:13:40) and war fighters somewhere at the end (00:13:43) almost an afterthought. (00:13:46) The result is duplication, drift and (00:13:48) confusion. And like the acquisition (00:13:50) process, we are already fixing the (00:13:53) creation of organizations to work around (00:13:56) the problems in the innovation ecosystem (00:13:59) rather than taking the bold steps needed (00:14:02) to transform it. We created an old (00:14:04) ecosystem to get around the actual (00:14:06) system. (00:14:08) No more. Every dollar of innovation, (00:14:11) whether it may be in a lab or a startup (00:14:13) or a classified shop, must exist to (00:14:16) deliver one of three things. (00:14:19) Gamechanging technology, (00:14:21) scalable products, or new ways of (00:14:24) fighting. If it's not doing one of those (00:14:27) three things at speed, it will be (00:14:30) realigned or it will go away. (00:14:33) And that's why I'm entrusting Emil as (00:14:34) the CTO to ensure that this directive (00:14:37) across the department is rapidly carried (00:14:40) out. (00:14:41) Emil, you're going to be busy. (00:14:45) Again, what we're talking about today is (00:14:47) a transformation in the way we think (00:14:49) about innovation. (00:14:51) In requirements reform led by Mike (00:14:53) Duffy, we killed an old model, a (00:14:56) sclerotic model, and rewired the (00:14:59) department so that problems, money, and (00:15:01) experimentation live in one system. In (00:15:04) acquisitions reform, we killed the (00:15:06) defense acquisition system and created (00:15:09) accountable portfolio acquisition (00:15:11) executives, (00:15:12) making speed to delivery, speed to (00:15:15) delivery our organizing principle. We're (00:15:18) going to do the same for technological (00:15:21) innovation. (00:15:22) That's why today at my direction, we are (00:15:24) ending the alphabet soup of councils (00:15:27) that meet and brief and write memos and (00:15:30) schedule meetings but never decide and (00:15:33) rarely if ever accelerate outcomes (00:15:36) effective immediately. The Defense (00:15:38) Innovation Steering Group, the Defense (00:15:40) Innovation Working Group, and the CTO (00:15:43) Council are disestablished and (00:15:46) abolished. (00:15:47) In their place, the CTO will convene a (00:15:50) CTO action group to assist him in making (00:15:53) decisions, clearing bureaucratic (00:15:55) blockers, holding leaders accountable, (00:15:57) and most importantly, quickly delivering (00:16:00) new technologies to our war fighters. (00:16:04) Every organization in this ecosystem (00:16:07) must earn its place by delivering (00:16:10) warfighting advantages faster than our (00:16:13) adversaries can adapt. No sacred cows, (00:16:17) no exceptions. (00:16:20) To back up the CTO, today at my (00:16:22) direction, we are realigning two pillars (00:16:24) of the war department's innovation (00:16:26) ecosystem. (00:16:28) First, the defense innovation unit or (00:16:31) DIU. (00:16:33) Since its establishment 10 years ago, (00:16:35) DIU has lived through shifting reporting (00:16:37) structures and uneven administrative (00:16:40) support, doing great things. But at (00:16:42) times, its portfolio overlap with other (00:16:44) parts of the department. Effective (00:16:47) today, DIU is is designated a Department (00:16:50) of War field activity, (00:16:53) providing exceptional tech scouting, (00:16:55) rapid contracting, and other common (00:16:57) services to the department executed at (00:17:00) commercial tempo. (00:17:03) The CTO will provide support to DIU for (00:17:05) administrative and resource matters and (00:17:08) ensure that the unit's efforts fit into (00:17:10) the department's innovation priorities. (00:17:14) The the director of DIU will also (00:17:16) continue to report directly to me as a (00:17:18) principal staff assistant and carry out (00:17:21) its statutory duties. (00:17:23) And I'm appointing Mr. Owen West, (00:17:26) executive of my drone dominance (00:17:29) initiative. As the director of DIU, (00:17:31) starting in March when the FY27 budget (00:17:34) cycle firms up. As a Marine with lots of (00:17:37) combat experience, Owen will bring a war (00:17:39) fighters mentality to DIU's core mission (00:17:43) of transitioning technology to our (00:17:45) troops. (00:17:47) Owen also has the private capital (00:17:49) experience needed to ensure DIU remains (00:17:52) working handin glove with the venture (00:17:54) and investor communities and continues (00:17:56) to on-ramp new entrance into the war (00:17:59) department. Owen led Doge at DO, (00:18:04) the most effective Doge effort across (00:18:07) the administration, saving tens of (00:18:09) billions of dollars for our department. (00:18:11) And now he will lead the (00:18:14) O.Lading World leading defense tech (00:18:15) startups have attracted billions of (00:18:17) dollars in capital. They're reshaping (00:18:19) warfare through the proliferation of (00:18:21) high-tech, lowcost technologies. DIU's (00:18:24) mission is to accelerate the adoption of (00:18:26) this commercial technology to help (00:18:28) convert entrepreneurial products into (00:18:31) tangible combat power. (00:18:34) Second, the strategic capabilities (00:18:37) office or SCO is also being designated a (00:18:41) department of war field activity aligned (00:18:43) under the CTO. SCO will maintain focus (00:18:46) on its core mission of identifying and (00:18:49) prototyping disruptive applications of (00:18:51) new systems, the unconventional uses of (00:18:54) existing systems, and near-term (00:18:56) technologies to create strategic (00:18:58) effects. (00:19:00) SCO will continue to maintain its (00:19:02) statutory direct reporting relationship (00:19:04) to the deputy secretary, but will be (00:19:06) operationally realigned under the CTO to (00:19:09) eliminate duplication and ensure the (00:19:12) relentless daily focus on delivering (00:19:16) near and medium-term capabilities to our (00:19:20) war fighters. Relentless urgency and (00:19:23) focus is our focus. (00:19:27) Today's defense innovation ecosystem is (00:19:29) too fragmented resulted in insufficient (00:19:32) technology transition to the war (00:19:34) fighter. We address some of this in the (00:19:36) transformation we are making in the (00:19:37) acquisitions ecosystem. But it also (00:19:40) needs to enter the innovation ecosystem (00:19:43) for too long. And I know a lot of you (00:19:45) have experienced this and others we met (00:19:47) with in Los Angeles recently. The (00:19:49) experience of founders and entrepreneurs (00:19:52) has been running endless laps around the (00:19:54) Pentagon looking for the right office, (00:19:58) the right program, and the right (00:20:00) motivated sponsor. And too often, new (00:20:03) entrance are ultimately stymied by the (00:20:05) bureaucracy. We hear it time and time (00:20:07) again. They don't know where to go, and (00:20:08) then when they go to that place, it's (00:20:10) not the right place to go. Then they go (00:20:11) somewhere else that didn't want them in (00:20:13) the first place, and the lap continues. (00:20:16) For example, SpaceX and Palanteer had to (00:20:19) sue the Department of War just to get a (00:20:22) shot at competing for department (00:20:24) contracts. (00:20:26) The bottom line is that new entrance (00:20:28) need both a shot on goal, but also (00:20:32) faster yeses and faster nos from the (00:20:34) department rather than being strung (00:20:36) along with a neverending stream of (00:20:39) rudderless may. (00:20:42) At the Secretary of War level, we will (00:20:44) replace the existing maze into two clear (00:20:48) channels. (00:20:50) The mission engineering and integration (00:20:52) activity MIA will tell industry what (00:20:55) problems we're trying to solve, and DIU (00:20:59) will help program offices adopt what (00:21:01) industry has already built. This will (00:21:04) help get to faster yeses or faster nos. (00:21:08) clear guidance, clear guide rails, clear (00:21:12) demand signal, which is what industry (00:21:14) and capital expects. (00:21:17) Now, this isn't just an office of the (00:21:18) secretary reform. The services need to (00:21:21) transform their innovation ecosystems as (00:21:24) well. The Army, the Navy, the Space (00:21:26) Force, the Marine Corps, and the Air (00:21:28) Force. (00:21:29) That's why today at my direction, our (00:21:31) military services will take the (00:21:33) following actions. Within 90 days, the (00:21:35) secretaries of the military departments (00:21:37) will brief the CTO on service innovation (00:21:40) plans, how they will consolidate, (00:21:42) streamline, and refocus their labs, (00:21:45) research expertise, experimental units, (00:21:47) and rapid capability offices around (00:21:50) three innovation outcomes. (00:21:53) Our 60 plus labs are a national asset, (00:21:57) but it's past time they were organized (00:21:59) to deliver and not just to discover. (00:22:03) And beginning with the fiscal year 2028 (00:22:05) budget, every portfolio acquisition (00:22:07) executive will fund an innovation (00:22:10) insertion increment, triple I, dedicated (00:22:14) money for the last mile, integration (00:22:16) test and rapid insertion of validation (00:22:18) solutions into fielded systems. (00:22:22) Innovation cannot be centralized and it (00:22:25) should not be. (00:22:27) In fact, hope hope it's okay I name him. (00:22:31) But just this past weekend, I received (00:22:33) an email, a proposal from an army (00:22:35) captain named Drenan Green, who I've (00:22:36) known for a while. He had a detailed (00:22:39) plan about how he and his unit wants to (00:22:41) deploy AI. (00:22:44) Innovation can and should come from (00:22:46) anywhere and anyone wherever those best (00:22:50) ideas reside. We're going to bring that (00:22:52) good captain in and hear from him. How (00:22:54) can we apply those tools in his unit and (00:22:57) other units bottom up, not just top (00:23:01) down? (00:23:03) The military services and program (00:23:05) offices own the last mile. And to that (00:23:08) end, I'm directing the services to (00:23:10) establish that triple I innovation (00:23:12) insertion increment within the budgets (00:23:15) of all program portfolios. (00:23:18) This will ensure that they have funds (00:23:19) set aside to quickly integrate (00:23:21) innovations during weapon system (00:23:24) development. (00:23:26) This will be a tool to free new (00:23:27) innovative cutting edge weapons and (00:23:29) improvements from the constraints of (00:23:32) yearly budget cycles. Here you iterate (00:23:35) in terms of hours and days, maybe weeks. (00:23:39) In Washington, we talk in terms of (00:23:41) months, years, and oftentimes (00:23:43) multi-year. It's too slow. The CTO (00:23:47) ecosystem will give them the supply. (00:23:49) Services are contributors to this (00:23:51) ecosystem, not bystandards. They must (00:23:54) deliver outcomes and they must deliver (00:23:57) overmatch. (00:23:58) And speaking of budgets and budget (00:24:00) cycles, you may have seen President (00:24:02) Trump's truth post of a few days ago. (00:24:06) He's proposing a $1.5 (00:24:09) trillion budget for the War Department (00:24:12) in fiscal year 27. This is by far the (00:24:16) most ever in our history a historic and (00:24:19) generational investment in American (00:24:22) security. We will not squander this (00:24:25) once-in-a-lifetime opportunity to (00:24:27) rebuild our military. Think of what this (00:24:30) means as our efforts to transform the (00:24:32) department come fully online. You add (00:24:35) resources and you streamline process (00:24:38) speed to deliver those capabilities to (00:24:40) the war fighter. an invitation to new (00:24:42) entrance into the defense industry, (00:24:44) which we need so badly, and an embrace (00:24:47) of AI and other cutting edge defense (00:24:50) tech. All of this will make our forces (00:24:52) more agile, more lethal, and more ready (00:24:56) to deter, and if necessary, win a future (00:25:00) fight. This is what President Trump (00:25:02) demands, (00:25:03) and this is what we will deliver. (00:25:08) This feeds into a next area we need to (00:25:10) fix. You see, reorganizing and a new (00:25:13) attitude are not enough to truly unlock (00:25:15) innovation for the war fighter. Our (00:25:18) department needs to capitalize on a key (00:25:20) advantage America has over her (00:25:22) adversaries, the broadest and deepest (00:25:24) capital markets, and the best (00:25:26) entrepreneurial talent in the world. (00:25:30) You know this capital is the lifeblood (00:25:32) of American innovation. And therefore to (00:25:35) be successful we need to embrace the (00:25:37) role in our department of private (00:25:40) capital which we have not done for far (00:25:42) too long. We need to be a better partner (00:25:44) for private capital so we can help (00:25:46) accelerate capital formation in key (00:25:48) areas and lower the risk for the (00:25:50) department. (00:25:52) Private capital is already helping to (00:25:53) solve a major problem that President (00:25:55) Trump has directed us to confront and (00:25:57) solve across this administration. (00:26:00) ending our reliance on competitors for (00:26:02) access to rare earth and critical (00:26:05) minerals. (00:26:07) The Office of Strategic Capital or OSC (00:26:09) under the leadership of David Lor is (00:26:12) spearheading those efforts and (00:26:14) delivering as we speak. In the past 5 (00:26:18) months alone, the OSC team has deployed (00:26:20) over $4.5 billion in capital commitments (00:26:23) as part of closing six critical mineral (00:26:26) deals, all of which will help free the (00:26:29) United States from market manipulation. (00:26:33) Crucially, OSC's strategy relies on (00:26:36) crowding in private sector investors (00:26:39) from this VC community to our country's (00:26:41) largest financial institutions. (00:26:44) The six OSE deals to date include nearly (00:26:47) 5 billion from our private capital (00:26:48) partners. (00:26:50) OSE has closed most of those deals in (00:26:52) less than 30 days, (00:26:55) which is Trump time for complex (00:26:58) transactions. (00:27:00) That may seem maybe slow for SpaceX (00:27:03) in Washington. That's as fast as it (00:27:05) gets. I mean, that is warp speed in (00:27:07) Washington. (00:27:08) Our all of our investments serve one (00:27:11) purpose. (00:27:13) deliver faster for the war fighter. Now, (00:27:17) some will want to cynically posit, is (00:27:20) this just another innovation (00:27:22) reorganization, (00:27:23) another set of strategies that brief (00:27:25) well but don't actualize? (00:27:28) No, this is the third leg of a single (00:27:31) war plan in fundamentally transforming (00:27:34) our acquisition ecosystem. We killed (00:27:37) JIDS that focused only on process and (00:27:40) turned the J-Rock into a body that ranks (00:27:42) realworld joint operational problems, (00:27:45) not paper requirements. We created the (00:27:47) MIA to run experiment experimentation (00:27:50) campaigns to solve these problems. We (00:27:53) killed the old defense acquisition (00:27:55) system and created the warf fighting (00:27:57) acquisition system to focus on speed, (00:28:00) risk, and accountability. (00:28:03) With the AI strategy and innovation (00:28:05) ecosystem transformation that we have (00:28:07) just outlined, we are welding that third (00:28:10) piece into place. The CTO, DIU, SCO, (00:28:15) DARPA, CDAO, and OSC are no longer a (00:28:18) loose federation. (00:28:20) They are the office of the secretary of (00:28:23) wars innovation operating system. DARPA (00:28:26) delivers gamechanging technology (00:28:28) innovation and strategic surprise. DIU (00:28:32) delivers scalable products. SCO delivers (00:28:34) new ways of fighting. And CDAO and OSC (00:28:38) provide the data, test, and capital to (00:28:39) move at wartime speed. (00:28:43) As of today, they are an ecosystem, (00:28:46) an arsenal of ideas and action for the (00:28:50) arsenal of freedom wired directly into (00:28:54) requirements, portfolios, and (00:28:56) production. Gamechanging technology, (00:28:59) scalable products, new ways of fighting. (00:29:03) All three moving at wartime speed. (00:29:06) Problems drive experimentation. (00:29:08) Experimentation flows to prototypes. (00:29:11) Prototypes flow to our program (00:29:13) executives. Program exe flow these to (00:29:16) production. Production flows to the war (00:29:19) fighter. And the cycle never stops. (00:29:22) Always iterating. One system, one (00:29:25) purpose, speed to the fight. We are (00:29:30) preparing to win the future because we (00:29:32) know and you know, for the free world, (00:29:36) for the west, the stakes could not be (00:29:39) higher. (00:29:41) Before this administration, our (00:29:43) adversaries (00:29:44) may have thought they finally broke (00:29:47) American power. (00:29:50) They're wrong. (00:29:52) They do not have our entrepreneurs. (00:29:54) They do not have our capital markets. (00:29:56) They do not have our combat proven (00:29:58) operational data from two decades of (00:30:01) military and intelligence operations. (00:30:03) They do not have our hard one classified (00:30:06) technologies. They do not have the (00:30:08) ingenuity and tenacity of American war (00:30:11) fighters who refuse to lose. (00:30:14) They don't have a military that can go (00:30:16) 37 hours to downtown Tyrron or downtown (00:30:20) Caracus without being seen in the (00:30:22) process. (00:30:24) You see, but none of that matters, (00:30:26) however, if we suffocate those (00:30:27) advantages under a stifling bureaucracy. (00:30:32) That's why we're unifying the innovation (00:30:34) ecosystem, making this an AI first (00:30:37) department, and holding every lab, every (00:30:41) program accountable, all while pushing (00:30:44) every officer to deliver warfighting (00:30:46) advantages faster than others can adapt. (00:30:51) We will not stop. We will not back down. (00:30:54) We will forge the new arsenal of freedom (00:30:59) with our partners in industry and the (00:31:01) private sector. We believe the future (00:31:04) will be shaped by those who lead in (00:31:06) technology and innovation. We don't (00:31:08) believe we know those who fervently (00:31:11) believe in freedom and the western (00:31:13) tradition like we do must be those (00:31:17) leaders. If not us, if not America, if (00:31:21) not the west, then who? And if not now, (00:31:25) it will be too late to maintain that (00:31:29) advantage. This is not reform for the (00:31:31) sake of reform. It never has been. This (00:31:34) is about whether our warriors fight with (00:31:36) yesterday's tools or they fight (00:31:39) overmatching our adversaries using (00:31:41) tomorrow's technologies. We know the (00:31:43) threat. We know the opportunity. We know (00:31:47) what must be done. We share the urgency. (00:31:51) Now, we will do it and we must do it at (00:31:54) wartime speed. (00:31:56) Thank you. God bless you. God bless this (00:31:59) company that you've built and may God (00:32:01) bless our great republic. Thank you very (00:32:03) much. Appreciate it. Thank you. (00:32:07) Thank you. (00:32:10) Thank you, sir. (00:32:20) Well, what a tour. What an opportunity (00:32:22) to be here at Starbase, Texas with Elon (00:32:26) and the Space X team. There's nothing (00:32:28) like this in America. There's nothing (00:32:30) like this in the world. And what you (00:32:33) have built and what you will build here (00:32:36) is a testament to the strength of (00:32:38) American ingenuity and American (00:32:40) invention. So, I want to thank all of (00:32:42) you, all the folks out here for having (00:32:44) us today. Elon, thank you so much for (00:32:46) hosting us, for what you built, for the (00:32:49) vision you have for this company, the (00:32:51) vision you have for our country, the (00:32:53) vision you have for American innovation. (00:32:56) I could not think of a more fitting (00:32:58) venue to continue our arsenal of freedom (00:33:02) tour and to outline today the future of (00:33:04) technological innovation at the War (00:33:07) Department. (00:33:09) Those of you here at SpaceX will (00:33:11) appreciate this, knowing that as World (00:33:12) War II was ending, the Secretary of War (00:33:15) and Secretary of the Navy wrote to the (00:33:17) National Academy of Sciences and (00:33:19) declared that scientific research was (00:33:22) essential to our national security. (00:33:25) To ensure continued preparedness, they (00:33:28) wrote, the research scientists of the (00:33:30) United States must be called upon to (00:33:32) continue in peace time some substantial (00:33:35) portion that of which they have made so (00:33:38) effectively during the stress of the (00:33:40) present war. The competitive time (00:33:43) element in developing those weapons and (00:33:46) tactics may be decisive in future (00:33:49) conflicts. You see, those secretaries of (00:33:52) war and navy many decades ago recognized (00:33:55) the importance that innovation and (00:33:57) readiness holds for our national (00:33:59) security. They knew what was at stake, (00:34:02) the very freedoms of the country we hold (00:34:05) dear. (00:34:07) All across the United States today, (00:34:09) extraordinary innovation is unlocking (00:34:11) new possibilities for freedom, (00:34:14) prosperity, and security. Now the (00:34:17) question before us is not whether or not (00:34:20) the most powerful technologies of this (00:34:22) century will reinforce free societies. (00:34:24) Is it going to reinforce our free (00:34:25) societies or will that technology be (00:34:27) shaped and twisted by malign regimes (00:34:31) that seek to use those technologies for (00:34:34) control and coercion? You see, over the (00:34:37) past several months, I've talked at (00:34:39) length about the challenges we face in (00:34:41) transforming the War Department to (00:34:44) address current and future missions. all (00:34:46) in service of meeting the needs of the (00:34:48) 21st century warfire. (00:34:51) I'd like to think we've already made (00:34:52) dramatic progress in the Pentagon's (00:34:54) culture by reviving the warrior ethos (00:34:58) and we're moving out quickly in (00:35:00) transforming our acquisition ecosystem (00:35:02) as well. But today is about how we (00:35:05) supercharge innovation at the war (00:35:08) department for the era ahead. (00:35:11) Innovation is happening at a pace we (00:35:13) can't even foresee. And we need the (00:35:15) entire enterprise, our enterprise to (00:35:17) embrace the urgency required for this (00:35:20) moment. (00:35:21) Since the end of the Cold War, the (00:35:23) defense industrial base in our country (00:35:25) has consolidated. (00:35:28) This makes it difficult, if not (00:35:30) impossible, for new creators of (00:35:32) technical innovations to win business at (00:35:35) our department. (00:35:37) The result is a risk averse culture that (00:35:40) prevents us from providing our war (00:35:42) fighters with the best resources that (00:35:44) America has to offer. (00:35:47) That ends today. (00:35:51) Simply put, the United States must win (00:35:53) the strategic competition for 21st (00:35:55) century technological supremacy, (00:35:59) artificial intelligence, autonomous (00:36:01) systems, quantum hypersonics, and long (00:36:04) range drones. If you talk to Elon Musk (00:36:06) long enough, he will tell you how (00:36:08) important hypersonics and long range (00:36:11) drones are. And he's 100% correct. Space (00:36:14) capabilities, directed energy, and (00:36:16) biotechnology are the new areas of (00:36:18) global competition. (00:36:21) The challenge is that our legacy (00:36:23) approach to technological development (00:36:25) assumes that technology moves in a (00:36:28) predictable linear conveyor belt from (00:36:32) lab to design to development to (00:36:34) prototype to test and qualify to program (00:36:36) of record and can only be provided by a (00:36:39) handful of companies that have (00:36:41) consolidated dramatically. (00:36:44) Now, while this system provided us with (00:36:46) the weapons that won the Cold War, it is (00:36:49) archaic and inconsistent with the novel (00:36:52) threat environment that we face today. (00:36:55) At its core, this old approach has the (00:36:58) hubris to assume that you can easily (00:37:00) predict the future, that you can foresee (00:37:03) how an invention becomes a weapon in (00:37:06) eight easy steps three decades from (00:37:09) first discovery. Our system cannot keep (00:37:12) treating innovation as a decadesl long (00:37:15) one-way march that dramatically reduces (00:37:18) who and what is able to run the gauntlet (00:37:21) at our department to get the capability (00:37:23) to the war fighter. (00:37:26) Until this administration, the Trump (00:37:28) administration, the department's process (00:37:30) for fielding new capabilities had become (00:37:33) just one more postcold war peace (00:37:36) dividend relic that has not kept up with (00:37:38) the times. Worse than that, we've done (00:37:42) nothing but add layer upon layer of (00:37:45) committees and councils that coordinated (00:37:48) but never decided. (00:37:50) We created endless projects with no (00:37:52) accountable owners. We have high churn (00:37:55) with little progress and few outputs. (00:37:58) That sounds about like the exact (00:37:59) opposite of SpaceX. (00:38:02) We treated innovation as a box to check, (00:38:05) not an outcome to deliver. In short, (00:38:07) when it comes to our current threat (00:38:09) environment, we are playing a dangerous (00:38:12) game with potentially fatal (00:38:15) consequences. (00:38:17) We need innovation to come from anywhere (00:38:20) and evolve with speed and purpose. (00:38:24) You see, America's open scientific (00:38:25) community is an advantage authoritarian (00:38:28) regimes cannot replicate. They can read (00:38:30) our scientific papers and copy our (00:38:32) invention or distill our ai models, but (00:38:36) they cannot replicate a culture like (00:38:38) this one of opened and distributed (00:38:41) research. (00:38:42) We need to be blunt here. We can no (00:38:44) longer afford to wait a decade for our (00:38:47) legacy prime contractors to deliver the (00:38:50) next perfect system only to find that (00:38:53) it's delivered years behind schedule and (00:38:56) cost 10 times what it should. (00:38:59) Winning requires a new playbook. (00:39:03) Elon wrote it with his algorithm. (00:39:05) Question every requirement, delete the (00:39:08) dumb ones, and accelerate like hell. (00:39:12) That's why I want to make clear across (00:39:14) the war department and for our partners (00:39:16) in the private sector that our under (00:39:18) secretary of war for research and (00:39:21) engineering, Emil Michael, right here in (00:39:23) the front row, is the war department's (00:39:26) single chief technology officer. One CTO (00:39:30) for the entire enterprise. Novel (00:39:33) concept. As the sole CTO, Emil will set (00:39:36) the technical direction, lead the (00:39:38) innovation ecosystem that will welcome (00:39:41) progress from anywhere it resides. And (00:39:44) he'll tell me face to face every day and (00:39:47) frankly whether we are gaining or losing (00:39:50) the technology and innovation (00:39:52) competition. He'll have the decision (00:39:54) authority and will lead through rigorous (00:39:56) evaluation with a focus on real (00:39:59) measurable outcomes. Now, Congress, and (00:40:02) we have member, many members of Congress (00:40:03) with us here today, will be our partner (00:40:05) in this. By affirming the under (00:40:07) secretary's role as the department CTO, (00:40:09) we're carrying out the core mission for (00:40:12) R& that Congress assigned in statute. (00:40:15) The last defense bill expanded the CTO's (00:40:18) authorities, including the power to (00:40:19) direct military departments and other (00:40:22) components to align around clear, (00:40:24) innovative outcomes. We we will use them (00:40:28) to deliver those outcomes with speed and (00:40:31) urgency. (00:40:33) An empowered CTO will inject a (00:40:35) disruptive mindset directly into our (00:40:38) systems, directing the power of (00:40:39) America's worldleading scientists and (00:40:41) entrepreneurs, our cutting edge labs, (00:40:44) the tech ecosystem, and our capital (00:40:46) markets to build what the war fighter (00:40:49) needs, but to do so better, faster, and (00:40:52) cheaper. And Emil is the right man to do (00:40:55) it. You see, this isn't about military (00:40:58) structure. This is about building an (00:40:59) innovation pipeline that cuts through (00:41:01) the overgrown bureaucratic underbrush (00:41:05) and clears away the debris Elon style, (00:41:09) preferably with a chainsaw, (00:41:12) and to do so at speed and urgency that (00:41:15) meets the moment. (00:41:17) As I've said repeatedly to every (00:41:19) audience, the president of the United (00:41:21) States and I have the backs of our war (00:41:23) fighters who have to make split-second (00:41:27) life and death decisions on the (00:41:29) battlefield. (00:41:30) And I want this audience to know that we (00:41:32) also have the backs of innovators who (00:41:36) share that very same urgency. (00:41:39) American taxpayers, especially those (00:41:42) parents whose sons and daughters have (00:41:44) answered the call to serve. I swore in (00:41:46) 10 more today. Demand and I demand and (00:41:50) we demand that we arm our war fighters (00:41:53) with overwhelming and lethal technology (00:41:56) right now, not a decade from now. In (00:42:00) modern warfare, the fastest innovator (00:42:03) and iterator will be the winner. And no (00:42:06) one can out innovate an American (00:42:08) entrepreneur who has been liberated from (00:42:11) the constraints of stifling bureaucracy. (00:42:15) That old era ends today. (00:42:19) We are done running a peacetime science (00:42:23) fair while our potential adversaries are (00:42:26) running a wartime arms race. (00:42:31) Right now, from garage startups to (00:42:33) factory floors to boardrooms across our (00:42:35) nation, the question of how to harness (00:42:37) innovation coming out of America's AI (00:42:40) ecosystem is front and center. and (00:42:41) rightfully so because it has the (00:42:44) potential to disrupt and transform every (00:42:47) area of human endeavor. The same is true (00:42:51) in the Department of War. President (00:42:54) Trump's AI executive order spells out (00:42:56) our approach succinctly. It is the (00:42:59) policy of the United States to sustain (00:43:01) and enhance America's global AI (00:43:03) dominance in defense of human (00:43:05) flourishing, economic competitiveness, (00:43:08) and national security. (00:43:11) That, ladies and gentlemen, is the War (00:43:14) Department's mission. We must ensure (00:43:16) that America's military AI dominance so (00:43:19) that no adversary can exploit that same (00:43:22) technology to hold our national security (00:43:25) interests or our citizens at risk. (00:43:28) America first in every domain. (00:43:33) Last month, I took the first step toward (00:43:35) changing how the department does (00:43:37) business with Frontier AI technologies (00:43:40) when we announced the rollout of Gen AI (00:43:42) with our partners from Google. And I (00:43:44) want to thank the Google team for (00:43:45) leaning forward and making the (00:43:47) investment to get their Gemini app to (00:43:49) about 3 million users in the war (00:43:51) department. (00:43:52) But today, we're excited to announce the (00:43:54) next frontier AI model company to join (00:43:58) Genai.mill (00:44:00) and that is Grock from X AI which will (00:44:04) go live later this month. So I want to (00:44:07) thank you Elon and your incredible team (00:44:10) for leaning forward with us on this as (00:44:12) well. Very soon we will have the world's (00:44:15) leading AI models on every unclassified (00:44:17) and classified network throughout our (00:44:20) department. long overdue. (00:44:23) To further that, today at my direction, (00:44:26) we're executing an AI acceleration (00:44:28) strategy that will extend our lead in (00:44:30) military AI established during President (00:44:33) Trump's first term. This strategy will (00:44:36) unleash experimentation, eliminate (00:44:38) bureaucratic barriers, focus on (00:44:40) investments, and demonstrate the (00:44:42) execution approach needed to ensure we (00:44:44) lead in military AI and that it grows (00:44:47) more dominant into the future. In short, (00:44:50) we will win this race by becoming an AI (00:44:53) first warf fighting force across all (00:44:56) domains from the back offices of the (00:44:58) Pentagon to the tactical edge on the (00:45:01) front lines. (00:45:03) The catalyst for this acceleration will (00:45:05) be seven pace setting projects focused (00:45:08) on mission threads across warfighting, (00:45:11) intelligence, and enterprise missions. (00:45:14) each with a single accountable leader, (00:45:16) aggressive timelines, and measurable (00:45:18) outcomes that answer a familiar (00:45:21) question. Elon, (00:45:23) what have you accomplished this week? (00:45:27) This is the execution standard for AI (00:45:30) first transformation. (00:45:33) Each of the seven pace setting projects (00:45:35) will use the following model. One owner (00:45:38) who reports monthly on their progress. (00:45:41) These projects will not be run in a (00:45:42) vacuum, but will work directly with war (00:45:44) fighters and transition partners to (00:45:46) ensure we incorporate real time (00:45:49) operational feedback. We expect rapid (00:45:52) iterations with failure accelerating the (00:45:55) learning curve. These are not science (00:45:57) projects. They are not governance (00:46:00) boards. They are the execution standard (00:46:02) for the entire department. (00:46:05) We will not win the future by sprinkling (00:46:08) AI onto old tactics like digital pixie (00:46:11) dust. We will win by discovering (00:46:14) entirely new ways of fighting. That's (00:46:17) why we will run continuous (00:46:19) experimentation campaigns, quarterly for (00:46:22) forceonforce combat labs with AI (00:46:24) coordinated swarms, agent-based cyber (00:46:27) defense, and distributed command and (00:46:31) control. pushing the envelope, learning (00:46:33) from failure at every stop, which is (00:46:36) exactly what this place does. You see, (00:46:38) our department doesn't accept failure in (00:46:40) the past, and so we never fail, which (00:46:42) means we never learn. We're flipping (00:46:44) that dynamic. (00:46:47) Before talking about the new rules of (00:46:49) the game, let me talk about a little bit (00:46:50) more about our new team. Because no game (00:46:52) can be won without the right team. And (00:46:54) we're proud to announce that Mr. Cameron (00:46:56) Stanley has been appointed the new chief (00:46:58) digital and artificial intelligence (00:47:00) officer, CDAO, of our war department. (00:47:04) Cam will be leading a new team, many of (00:47:06) whom have foregone, thank you, or left (00:47:09) lucrative careers at pioneer companies (00:47:11) such as AWS, Data Bricks, Palanteer, and (00:47:14) Meta to join the fight. (00:47:16) This team will not only provide a (00:47:18) catalyst for change in this department, (00:47:21) but will also act, we believe, as a (00:47:24) magnet for other talented members of the (00:47:26) tech community who want to join us in (00:47:28) doing the mission focused work to (00:47:30) protect our great republic. (00:47:34) So, let's talk about the new rules. (00:47:36) First, speed. (00:47:39) Speed wins. Speed dominates. Our (00:47:43) enterprise currently operates on (00:47:45) staffing and committee cycles measured (00:47:47) in months and years, and that's (00:47:50) unacceptable. (00:47:52) Military AI is going to be a race for (00:47:54) the foreseeable future, where the risks (00:47:55) to US national security of moving too (00:47:58) slowly outweigh the impacts of imperfect (00:48:02) alignment. (00:48:04) To do this, Cam and his team at TD uh at (00:48:06) TDAO will define AI deployment velocity (00:48:09) metrics for all the pace setting (00:48:11) projects in the next 30 days and report (00:48:14) at least monthly after that. These will (00:48:17) become the new benchmarks for programs (00:48:20) across the department. (00:48:23) Second, bureaucratic blockers. If you (00:48:27) work with Elon, he you know he finds the (00:48:29) blockers and you remove them. We will (00:48:31) take a wartime approach to people and (00:48:34) policies that block this progress. (00:48:38) You want to block, you can work (00:48:40) somewhere else. Barriers to data (00:48:42) sharing, authority to operate, or ATO's, (00:48:45) test and evaluation and contracting are (00:48:48) now treated as operational risks, not (00:48:51) simply bureaucratic inconveniences. We (00:48:54) are blowing up these barriers. (00:48:58) That's why today at my direction, I'm (00:49:00) establishing a barrier removal SWAT team (00:49:03) under R& (00:49:05) authority to wave non-stutory (00:49:07) requirements and escalate to our great (00:49:09) deputy secretary Steve Fineberg anything (00:49:12) that slows down the acceleration of AI (00:49:15) capabilities. (00:49:17) Third, compute resource. We will invest (00:49:20) heavily in expanding our access to AI (00:49:22) compute from data centers to the (00:49:25) tactical edge and will tap into hundreds (00:49:27) of billions of dollars in private (00:49:29) capital flowing into American AI. (00:49:32) President Trump's executive order has (00:49:34) directed us to build data centers on (00:49:36) military land and to work with the (00:49:39) Department of Energy to ensure that we (00:49:41) dramatically increase the number and (00:49:42) breadth of resources needed to power (00:49:45) this computing infrastructure. (00:49:48) We will work together with our partners (00:49:50) at Google and AWS and Oracle and SpaceX, (00:49:53) Microsoft and others on these (00:49:55) initiatives. (00:49:57) Fourth pillar, (00:49:58) especially in this room, you'll (00:50:00) understand it, is talent. We will use (00:50:03) every hiring and pay authority available (00:50:06) to us to bring the best American (00:50:09) technical talent and reward effective AI (00:50:11) transformations by our workforce. We're (00:50:15) going to heavily leverage President (00:50:16) Trump's tech force initiative to bring (00:50:18) in the best and brightest from industry (00:50:20) and academia. (00:50:22) With people like Elon, David Saxs, Emil, (00:50:25) Mike, and others from the (00:50:27) entrepreneurial and business world (00:50:28) already in government, (00:50:30) we have shown that we can and that we (00:50:33) must enlist the world's leading talent (00:50:36) in this cause. (00:50:39) Fifth, responsible AI. Today I want to (00:50:43) clarify what responsible AI means at the (00:50:47) Department of War. Gone are the days of (00:50:51) equitable AI and other DEI and social (00:50:55) justice infusions that constrain and (00:50:57) confuse our employment of this (00:50:59) technology. (00:51:01) Effective immediately responsible AI at (00:51:03) the War Department means objectively (00:51:06) truthful AI capabilities employed (00:51:09) securely and within the laws governing (00:51:11) the activities of the department. We (00:51:13) will not employ AI models that won't (00:51:16) allow you to fight wars. (00:51:21) We will judge AI models on this standard (00:51:23) alone. Factually accurate, mission (00:51:25) relevant, without ideological (00:51:27) constraints that limit lawful military (00:51:30) applications. (00:51:31) Department of War AI will not be woke. (00:51:35) It will work for us. We're building war (00:51:38) ready weapons and systems, not chat bots (00:51:41) for an Ivy League faculty lounge. (00:51:45) Sixth, and finally, (00:51:48) data. AI is only as good as the data (00:51:50) that feeds it. And the US military has (00:51:53) an asymmetric data advantage from two (00:51:55) decades of military and intelligence (00:51:56) operations that no other military in the (00:51:59) world can replicate. (00:52:01) But right now we are underutilizing this (00:52:03) advantage. Too much of our data is (00:52:05) stranded. It's stuck in bespoke program (00:52:09) databases locked behind title 10 or (00:52:12) title 50 stove pipes invisible to (00:52:14) operators, engineers, and industry who (00:52:17) could help us exploit it with winning (00:52:19) speed and scale. And that's why today at (00:52:22) my direction CDAO will exercise its full (00:52:25) authority to enforce the DOW data.