Jack Welde

Is the age of the legacy defense prime over? While most see Anduril as a hardware company, the real story is the complete reinvention of the business model behind defense contracting. Their "Venture-Backed Prime" approach is a fundamental threat to the old guard. We synthesized over 100 sources, and the narrative is clear: 1️⃣ Financial Dominance: Anduril achieves 40-50% gross margins, while traditional primes operate on 8-10%. 2️⃣ Unmatched Speed: They deliver on 12-18 month development cycles, compared to the industry standard of 5-7 years. 3️⃣ Strategic Disruption: By self-funding R&D with private capital, they sell finished products directly to the government, bypassing the traditional cost-plus model. This isn't just a new player; it's a new game. See the full, data-backed narrative in our deep-dive report. Full Report: https://lnkd.in/gQV-DGKq What other hardware-heavy industries are ripe for this kind of software-first disruption? #DefenseTech #VentureCapital #Anduril #Disruption #Strategy #Innovation

16

🚨2025: The Year the Defense Tech Wakeup Call 🚨 For the past several years, defense tech has had tunnel vision, riding a wave of unrealistic optimism, some of it warranted, much of it not. A frenzy of investment, hype, and glossy marketing created the belief that a new generation of unmanned systems would define the battlefield of the future. But 2025 is the year the curtain finally came down. Across U.S. testing ranges and the Ukrainian front lines, a consistent and uncomfortable story has emerged: many U.S. drone manufacturers are failing to meet the realities of modern, EW-heavy combat. Systems that look impressive in PowerPoints, marketing videos, fluffed-up TAM, and tightly-controlled demos are collapsing the moment adversary electronic warfare enters the equation. 📰 Headlines That Changed the Narrative This wasn’t an isolated failure; it was a pattern. -January: Army Blocks Air Force’s AI Program (OpenAI) Over Data Security Concerns (Air and Space Magazine) -May: A Navy experiment off the coast of California saw over 30 drone boats automatically idle as a fail-safe, rendering them “dead in the water.” (WSJ) -June: A DIU test in Alaska reported that the multi-hundred-million-dollar Replicator program was not operationally viable despite years of trials and repeated setbacks. (Defense News) -August: The Director of DIU resigned following a series of unsuccessful unmanned systems experiments. (Defense News) -October: Reports of "very high risk" inside NGC2. (Reuters) 💡The uncomfortable conclusion: The COTS-first approach to building expensive unmanned platforms and fragile comms links doesn’t survive in spectrum-denied environments. Yet inside parts of the DoW ecosystem, the narrative persists: “Drones are the future.” Yet, every operator who has been downrange over the past 12 months would tell you a very different story. The truth emerging from modern combat is clear: drones are useful but only when they are cheap, iterative, resilient, and built for the contested spectrum. ⚔️Ukraine has become the global crucible for modern unmanned warfare. They iterate faster. They field cheaper. They survive EW environments that instantly disable most U.S. systems. 📡The U.S. Is Behind in the Spectrum Fight At the tactical edge, where information dominance decides outcomes. We are fielding capabilities that look strikingly similar to what we used during the GWOT era. Modern warfare assumes a contested spectrum as a default, not an exception. And in that environment, we are behind. 2025 Isn’t a Story of Failure, It’s a Reset This year marks a turning point not because the tech failed, but because the illusions finally did. -To break from the thinking that got us here. -To build for EW-contested environments, not PowerPoint-perfect ones. -This isn't a video game; people's lives are on the line. When we get this right, the next wave of U.S. defense innovation won’t just catch up; it will redefine the fight. 🇺🇸

51

13 Comments

Airspace is rapidly becoming the next major challenge in public safety. Once considered passive infrastructure, the sky is now a contested domain—filled with drones, crewed aircraft, and airborne threats that legacy systems weren’t designed to handle. Ash Alexander-Cooper OBE, part of our team at Dedrone by Axon, captured it well in his recent Financial Times op-ed: “Each delay in adapting to the evolving threat makes this new reality harder to defend.” He’s right. But the good news is, we don’t have to meet this moment with fear—we can meet it with speed. With connected systems. With AI that shortens the distance between detection and response. Axon has the tools to transform the sky from a vulnerability into a vital layer of protection. Ash’s perspective is well worth a read—and the discussion it should inspire is even more important. As we confront this next frontier in safety, we want to hear from others in the field: How can we move faster, together, to solve this challenge before it grows? https://lnkd.in/g2ARip7Y

178

6 Comments

Beyond Subscription for the Service Providers I propose additional revenue models for high-tier service providers. While current leaders like Google, LinkedIn, Meta, and Netflix utilize cutting-edge technology, their revenue models remain primitive, relying almost exclusively on advertising or flat-rate monthly subscriptions. These models are sufficient for average consumers, but they fail to capture the value sought by power users willing to pay for specific, high-value additions. My proposition is to charge based on incremental value. Netflix: Tiered Content and Feature Monetization Flat-fee subscriptions disadvantage premium content producers by averaging the value of high-budget and low-budget media. Premium Access: Specific high-value content should allow for an additional per-view charge, enabling the inclusion of more specialized or high-budget productions. Feature-Based Billing: Services such as dubbing or specialized subtitles could be charged as add-ons. This creates a self-sustaining revenue stream for localization, increasing total content availability without inflating the base subscription cost. LinkedIn: Usage-Based Utility Current subscription models often force users into long-term packages for short-term needs, or conversely, penalize high-intensity interaction by flagging accounts. On-Demand Actions: Advanced searches and high-value interactions should be available via per-action or short-term bursts rather than monthly commitments. Retention Logic: Rigid account blocking for high-intensity use disincentivizes platform interaction. A usage-based model allows for "heavy use" periods without compromising account status, preventing the increase of "ghost accounts." Google Gemini: Compute-Based Pricing Standard monthly fees for AI do not align with the high infrastructure and energy costs of data centers. For productivity-focused users, pricing should scale with computational load. Dynamic Assessment: Upon receiving a query, the system should provide three processing tiers: Fast Thinking: 0.2 cents Moderate Thinking: 1.2 cents Heavy Thinking: 2.0 cents Efficiency: This allows users to pay only for the required "compute" for a specific task, removing the barrier of a high flat-rate monthly fee for intermittent users. Implementation: Pre-Paid Credit System I propose a pre-paid "wallet" system to facilitate these micro-transactions. Privacy and Security: Pre-payment removes the necessity for service providers to store sensitive credit card data. Financial Logic: Service providers benefit from the time value of money by receiving payments upfront. Micro-transactions: Small-scale billing (in cents) is technically feasible through a pre-paid balance where credit card processing fees would otherwise make it impossible. Transferability: This system allows for easy peer-to-peer balance transfers, such as a parent allocating a specific budget to a child’s account for controlled service consumption.

1

Exciting developments in aviation autonomy! The FAA recently released the Notice of Proposed Rulemaking (NPRM) for Part 108, enabling scalable Beyond Visual Line of Sight (BVLOS) drone operations. This will transform industries like delivery, agriculture, #healthcare, public safety, and more—impacting everyone, no matter your profession or location. Watch Secretary Sean P. Duffy's press conference here: https://lnkd.in/gSbYYRhF Autonomy is here! #FAA #Drones #BVLOS #AviationInnovation #AutonomousTech

12

1 Comment

Anthropic’s Special Forces: Custom Agents for the Expert Workforce - Launches Claude Enterprise Plugins for Finance, Engineering, and Design Why Anthropic is Focusing on Expert-Grade Agentic Workflows While the rest of the AI world is busy building "general" assistants that can write poems and summarize emails, Anthropic is going deep. On February 24, 2026, the company announced a massive expansion of its enterprise ecosystem with dedicated plugins for Finance, Engineering, and Design. By allowing Claude to hook directly into professional-grade software stacks (like Bloomberg terminals, GitHub repos, and Figma files), Anthropic is positioning its "Constitutional AI" as the safest and most competent choice for high-stakes, specialized industries. The launch marks a shift from "Chatbot" to "Colleague," focusing on vertical integration rather than broad consumer appeal. Specialized Intelligence: Unlike standard AI agents, these plugins are tuned with domain-specific "expert knowledge." Finance: Includes real-time audit trails and compliance checks. Engineering: Can navigate massive legacy codebases and perform "unit-test" simulations before suggesting changes. Design: Deep integration with visual assets and brand-style guide enforcement. The "Safety First" Moat: Anthropic is leaning into its "Constitutional AI" framework. These agents operate within strict, verifiable guardrails crucial for finance and engineering departments that cannot risk "hallucinated" numbers or insecure code. Infrastructure Partners: The launch is supported by early partnerships with firms like Infosys (to scale deployment) and Slack, where these specialized agents can live as active members of project channels. Data Privacy: Anthropic emphasized that data processed through these enterprise plugins is never used to train its public models, addressing a major "blind spot" for corporate CSOs. Read more here: https://lnkd.in/e5B-aKdc Who Should Care - CFOs & Financial Analysts: Who need AI that can handle complex spreadsheets without the risk of data leakage or calculation errors. - CTOs & Lead Developers: Looking to automate the "drudgery" of code maintenance and documentation using an agent that understands their specific tech stack. - Creative Directors: Managing large-scale brand assets who need an AI "governor" to ensure global design consistency. - Enterprise CISOs: Who have been hesitant to deploy agents until a "safety-first" vertical solution was available.

13

After decades of relative stability, the global security landscape has fundamentally shifted. Modern defense challenges don’t move in straight lines — they’re dynamic, distributed, and increasingly autonomous. Yet most robotic systems today still operate in isolation: siloed, pre-programmed, and unable to adapt together in real time. Gambit AI is building the collaborative intelligence layer for defense robotics — enabling air, ground, and maritime systems to coordinate, adapt, and execute as a unified force. Think adaptive autonomy, cross-platform interoperability, and real-time decision-making at machine speed. This is is about giving operators a decisive advantage in an era where speed, resilience, and coordination define security. It’s a crucial time to be building advanced capabilities in collaborative autonomy - Eclipse is proud to partner with Joshua Giegel, Benjamin Richardson, Andrew Kemendo and their team as they build foundational infrastructure for the next generation of defense. Read more here on Colin Demarest's article in Axios: https://lnkd.in/g63bmiBM Lior Susan Seth Winterroth Kaitlyn Glancy Joe Fath Jiten Behl Charly M. Ryan Gibson Greg Lyon Greg Reichow Michael Laser Allyson Johnson, CAIA Mislav Tolusic Laura Spaventa Lewis Heather Mack L3 Technologies Sierra Nevada Corporation RTX Even Rogers Daniel Jablonsky Diana Zhou

107

7 Comments

Worth the read! Commcrete is breaking SATCOM tradeoffs at a time when governments are prioritizing resilient communications as a foundational layer for autonomy, electronic-warfare resilience, and distributed command & control, while commercial operators in aviation, shipping, energy, and transportation face similar requirements as their operations digitize and move beyond terrestrial coverage. Raz Mangel | Meir Cohen | Greenfield Partners

16