Samuel Gerace

While investing in #AI, many corporations fail to achieve strong ROI as a result. The root of this failure is not technological. It is macroeconomic and organizational. US enterprises are attempting to execute a paradigm-shifting technological transition within a market environment that offers them no structural shock absorber, in the way a formal industrial policy can. These enterprises must navigate the AI transition while addressing a rapidly changing market environment and the AI Stakeholder Squeeze. The post builds expands on some of the ideas presented in the recent podcast with Ben Lorica 罗瑞卡 about remaining employable in the AI era (https://lnkd.in/gAGRYxQN) #AIstakeholdersqueeze #AItransition #industrialpolicy

15

2 Comments

17 ELITE DC-BASED VCS 1) Red Cell Partners: venture studio that builds, launches, and scales tech companies in healthcare, cyber, and national security. Portcos: Zephyr AI, DEFCON AI, Valinor, Epirus. cc Grant Verstandig. 2) NQB8: incubation firm founded by Auren Hoffman 📚 from Flex Capital. Portcos of Flex: Chime, Gusto, Carta, Replit. 3) Construct Capital: early-stage venture investing in foundational industries. Portcos: Hadrian, Amca, Veho. cc Rachel Holt, Dayna Grayson. 4) AE Industrial Partners, LP: multi-stage private investments in aerospace. Portcos: Firefly Aerospace, Bigbear.ai. cc Beckett Jackson. 5) Columbia Capital: multi-stage private investments in digital infrastructure, enterprise technology, and mobility. Portcos: CloudSherpas, Cologix, Mesh-AI. cc 6) Lavrock Ventures: early-stage venture investing in enterprise software, cyber, and national security. Portcos: Castelion, ID.me. cc Alex Poulin. 7) Shorewind Capital: multi-stage private investments in frontier technology. Portcos: Fleetzero, Poolside, Synthesia. cc Jake Poliskin. 8) Washington Harbour Partners LP: private investments in enterprise tech and national security companies. Portcos: Northwood, Raft, Anduril Industries, APEX, UiPath, OpenAI. cc Patrick Foley. 9) Shield Capital: early-stage venture investing in AI, autonomy, cyber, and space. Portcos: HawkEye 360, Vannevar, Armada, Obviant. cc David Rothzeid. 10) Point72 Ventures: multi-stage venture investing in AI, enterprise software, and defense. Portcos: Apex, Adonis, Cape, Overland AI. cc Chris Morales. 11) Leonid Capital Partners: private credit for government contracts. Portcos: Zeno Power, N5 Sensors, Zero-G. cc Samuel MacGregor. 12) New Vista Capital: early-stage venture investing in frontier technologies. Portco: Chariot Defense, Castelion, Regent. cc Kirsten Bartok Touw. 13) Sands Capital: mid and growth-stage investing in AI, cyber, healthcare, defense. Portcos: Anduril, DataRobot, Ramp, Rippling. cc Chris Eng. 14) NextEra Energy, Inc. Ventures: venture and growth in cyber, energy, and data companies. cc Mitch Gainer. 15) Overture Ventures: early-stage venture investing in AI, energy, and industrial technologies for the national interest. Portcos: Blue Energy, Crux, Alta. cc Michael O'Neil. 16) The New Industrial Corporation: private investing and project development at the intersection of venture and infrastructure capital. cc Austin Bishop. 17) StepStone Group: early, growth, and secondary investing in tech and tech-enabled companies. Portcos: Chewy, Anduril, SpaceX, Cloudflare. cc Stephen West. There are a lot more investors locally than this - and I have more lists to share (stay tuned) - but please comment who I missed !

467

54 Comments

Four new domains secured to the portfolio: ​Bioidentityprotocol.com Taskpilotlegal.com Fusionbaseload.com Fusionroadmap.com Doubling down on high-utility niches—specifically biotech, legal tech, and energy infra. ​Direct link to the shop and pricing in the comments below. #AI #Infrastructure #Biotech #Investments

An unintended consequence is perhaps a narrower intake into these professions, which certainly supports the idea that AI will further deepen societal inequality.. https://lnkd.in/esnDX3jw

8

I’ve been working on the theory of why small, shippable factories are the perfect inversion to the centralized prefab factory model. It all collapsed into one line of math (credit to Gilles Retsin): Factory Efficiency = Usage × (Throughput × Product Value ) / ( CapEx + Deployment OpEx) The idea is that the closer you can drive Usage to 100 % while keeping the denominator tiny, the faster the flywheel spins.  A shippable microfactory tackles this by primarily focusing on a high Usage factor. Because the microfactory can be packed up afterwards and reused, its CapEx gets amortized over multiple projects, effectively raising its overall Usage across the year. If one project alone doesn’t fully occupy the factory, it can simply roll over to another job to maintain high Usage. Its CapEx is often an order of magnitude lower than a centralized factory (< $1M vs. $10-50M min), so the breakeven throughput is more achievable on a small pipeline of work. The real breakthrough here is portability. There is near-zero stranded capital and almost no idle time. To flesh this theory out, I wrote 3000 words on the topic for Brad Hargreaves and Thesis Driven. The article breaks down: - why “fixed factories, shipped goods” is being inverted to shipped factories, fixed goods - what a sub‑$1 M robotic cell does to CapEx per home versus a £45 M off‑site plant - early production/cost data on a shippable microfactory from Mollie Claypool, Gilles Retsin, Sam Baker and the Automated Architecture (AUAR) team - the hybrid reality of robots tackling the heavy structural components and human crews assembling faster and cheaper - technical constraints to pull this model off and future considerations on the optimal business model (own vs. rent) Major shoutout to Gilles Retsin for helping me pull this together. https://lnkd.in/eRTnbZPa

54

2 Comments

A recent Financial Times analysis—by John Thornhill—offers a timely reminder that beneath the market noise, real progress in quantum is accelerating, and those focused on practical, impactful applications will shape the future far more than those chasing the hype.

38

2 Comments

Peter H. Diamandis recently shared an interesting analysis on the evolving relationship between governments and large tech companies in the age of AI. His core observation: As we transition to an economy increasingly defined by compute power and energy, the traditional roles of nation-states and corporations are being tested. He outlines two potential paths forward: **Path 1: Evolved Public-Private Partnership** Governments focus on guaranteeing citizen outcomes rather than directly providing services. They leverage their control over land and energy sovereignty while tech companies build and operate the infrastructure. Both parties play complementary roles with clear boundaries and interoperability requirements. **Path 2: Corporate Utilities Become Sovereign-Like** If governments remain bureaucratically constrained, tech companies that control energy production, issue stable digital currencies, and provide superior services could attract "subscription citizenship," creating corporate-run economic zones that rival traditional nation-states. To illustrate the scale: The combined market cap of major tech companies (Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta, Tesla) is roughly $14 trillion—comparable to half the US GDP and larger than any country except the US and China. Diamandis argues this isn't speculative—elements are already emerging. The question is whether we move toward partnership or competition between these institutions. Worth reading his full piece for the detailed economic and structural arguments. https://lnkd.in/gPt46cXT Thoughts on which direction seems more likely? Or whether there are other models we should be considering?

23

This is quite the list for Q-CTRL to be on, and very well-deserved to Michael Biercuk and the team. Q-Ctrl is one of the first companies to commercialize applications of quantum control at scale. Among other advancements, they are providing navigation in GPS-denied environments. If you've ever flown out of Istanbul airport on a commercial flight, the plane likely lost GPS at some point. And GPS denial is a frequent concern in geopolitical conflict zones. Commercial applications in quantum computing and sensing are here, and Q-Ctrl is inventing many of them. https://lnkd.in/gRe3tsDp Fen Zhao Alpha Edison

26

2 Comments

Believe it or not, even this astounding measure of AI offering 4x improved accuracy at 20% of the cost versus human physician’s diagnostics vastly undersells the current AI advantage. This paper only includes the cost of the diagnostic effort. What it does not include, is any consideration of the costs of mistaken diagnoses, both in money and in lives. Incorrect diagnoses lead to massively increased cost in treatment, ongoing disease, and worsening conditions. The faster these capabilities are rolled out across public health, the better.

15

1 Comment

The Operator’s Edge – Part IX From Entry to Exit: Why Operator’s MOIC Is the New Standard for Realization By George Konstantopoulos | PE Operatinghttps://lnkd.in/gaXnizQj

1

Jensen also took time in the prepared remarks and throughout Q&A to specifically address bubble and circularity concerns, reiterating the existence of three platform shifts occurring simultaneously, (1) CPU to GPU, (2) Gen-AI now affecting existing applications (Meta's rollout of GEM highlighted), and (3) Gen-AI creating "the next frontier of computing" with entirely new companies and business models being created by AI. All three vectors are growing and require infrastructure investment to support it. As far as concerns on circularity, Jensen looks at investments in OpenAI and others as strategic investments to enable the Nvidia ecosystem. In the case of OpenAI, Jensen called them "one of the most consequential once in a generation" companies and expects any investment to be accretive for Nvidia shareholders. In the case of Ai-on Lab, it was said that this will be their first time adopting Nvidia - a big revelation we expect to get more color on in the future.

2

Hexcel Adds Independent Director in Deal With Activist Investor "Hexcel added a new independent director to its board in a move that avoids a potential proxy fight for the aerospace materials company with activist investor Vision One Fund. "As part of an agreement between the two sides, Vision One will withdraw their nominees to the board that were to be voted on at the company's annual shareholder meeting. Vision One will also support Hexcel's board nominees, and agree to certain standstill restrictions and mutual non-disparagement provisions, according to a filing with the Securities and Exchange Commission on Wednesday. "The new director, former Kaman Chief Executive Officer Neal Keating, was appointed effective Tuesday." Keating's experience in aerospace may make him an ideal director choice. Yet from a bigger picture perspective, I wonder if this settlement is the right move for investors. More and more activist campaigns are settling rather than running the full process of a proxy fight with typically a greater number of directors nominated by the activist. Hopefully, activists are becoming like the public company boards they battle, i.e. more focused on reaching consensus than big results. #governancearbitrage #proxyfights #corporategovernance #activistinvesting #valuecreation https://lnkd.in/gazBUGFp

4

The Hidden Risk in AI-Driven Service Models That Investors Can’t Ignore There’s a growing risk that many long-term investors in AI-based service companies might be underestimating: the pace and unpredictability of disruption from the AI incumbents themselves. This week’s news about OpenAI launching a dedicated consulting arm is just the latest example. They’re embedding their own engineers directly into client organizations, offering custom-built AI solutions with a $10M minimum ticket size. It’s a bold move that puts them in direct competition with consulting giants like Palantir and Accenture—while also encroaching on the business models of hundreds of AI startups and SaaS providers. The implications go beyond consulting. AI-native companies are no longer just providing tools—they’re integrating deeply into core workflows, displacing other software vendors, and absorbing value higher up the chain. That creates real strategic and valuation risk for investors backing startups or growth-stage companies with “AI-powered” service models. How do you underwrite a business when its category might be swallowed by the platform it depends on? I’m curious—how are you thinking about this risk? Are you building this into your due diligence or scenario planning? And what kind of mitigation strategies have you found effective, if any? Would love to hear how you’re navigating this shift. Context here: https://lnkd.in/e5dSvKR9 #AI #investing #risk #startups #strategy #OpenAI #platformrisk

5

Most companies don’t have an AI problem. They have a clarity problem. They buy tools before defining outcomes. They automate before fixing broken processes. They chase “AI transformation” without a strategy that actually creates leverage. We see this every week. AI doesn’t replace thinking. It amplifies it. When AI fails, it’s usually because: • No clear business objective • No operational foundation • No human decision framework • No accountability for results The winners in the next 24 months won’t be the companies with the most AI tools. They’ll be the ones who use less AI, more intentionally. AI should: → Reduce friction → Increase decision speed → Multiply top performers → Create unfair advantages If AI isn’t doing at least one of those, it’s just expensive noise. The real question leaders should be asking isn’t “Which AI tool should we use?” It’s: “What would we want AI to eliminate, accelerate, or multiply if it actually worked?” Curious how other leaders are approaching AI right now. Are you: A) Experimenting B) Scaling what works C) Still watching from the sidelines Drop A, B, or C in the comments 👇 (And if this resonated, share it with someone still chasing tools instead of outcomes.) — 369aiconsulting.com

4

Over the weekend I was thinking about how GPT-5 and other recent model upgrades will push industrial automation forward, and translate into physical-world actions. Our focus at nvp capital is on founders building in vertical AI, and while much of that is LLM based today, in the future (especially for industrial use cases) we see a world where that becomes driven by Vision–Language–Action models (VLAs). GPT-5 takes a real step forward in multimodal reasoning — not just reading text, but interpreting images in context. Drop in a photo of a landmark and it often nails the location (and yes, it will tell you it’s better than GPT-4!). 📸 Multimodal context is a bridge to actions, where productivity gains can really scale. Today, OpenAI and other foundation model players focus on digital actions, while physical actions -- the domain of VLAs -- are mostly pursued by companies with physical assets and data, like Gemini or Grok. For founders building in industrial and robotic automation 🤖, interest in VLAs is growing — even if they’re still too early for broad commercial use. GPT-5’s integrated handling of text, images, and other inputs could make it an ideal VLA “brain,” capable of interpreting scenes, following nuanced instructions, and planning precise, multi-step actions that lead to highly adaptable agents moving seamlessly from perception to action. An open question is whether OpenAI will step into that physical domain — or leave VLAs to those already playing in the hardware-heavy world ⚙️. Either way, if you’re building vertical software using VLAs, we’d love to talk to you!

10

4 Comments

Congrats OPAQUE Systems and Confidential Computing Consortium on organizing another great Confidential Computing Summit! Thanks for having me on the VC panel - I enjoyed the discussion with Larry, Natalia and Aaron. Here are a few takeaways from our discussion: 💡 Shadow AI: similar to shadow IT, it is turning out to be a headache for enterprise buyers. For teams and individuals, it is easy to start working with AI products but that can quickly lead to data leakage. While incumbents will look at this space, it is an opportunity for founders to build something interesting. 💡 Data: it continues to be a barrier to enterprise AI adoption. Data discovery, lineage, policy, governance become key for leaders interested in gen AI. While founders are moving fast with AI, they might be better served helping customers get their data in a state that helps accelerate AI adoption. 💡 AI attack surface: going back to the k8s example - developers want clusters to be spun up quickly whereas security org wants to ensure things are done in a secure/compliant manner. There is always this tension between moving fast and doing it within enterprise guardrails. This problem becomes larger once agents/llms are introduced. Security for AI and AI for security will both continue to become interesting spaces for innovation. 💡 Future: Agents are coming. The most basic agent is a while () loop with a bunch of API calls including tools/LLMs and some recursive logic. Once rich context is added and maintained, these orchestrated systems can solve problems with high accuracy. Customers can already hire agents to make workforce more productive. This creates challenges with identity (authZ, authN), onboarding, observability, communication, networking. Please reach out if you are thinking of starting something new in these areas! #confidentialcomputing #ai #vc SIERRA Ventures

48

7 Comments