Mark Kvamme

Sustained ~252–291 GiB/s of HBM memory throughput on H100s under decode load — essentially hitting the hardware roofline. This matters because HBM throughput, not FLOPs, is the real bottleneck in LLM inference. By keeping memory nearly fully saturated, I’ve unlocked far higher efficiency and throughput than standard engines. The result: 0.9–1.36M tokens/sec with ~0.1 ms first-token latency  #AI #LLM #GPURouter #H100 #Inference #CostEfficiency #Innovation #DeepLearning #AIInfrastructure #HighPerformanceComputing NVIDIA AMD OpenAI Google Shilpa Kolhatkar Keith Strier a16z speedrun

5

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

AI isn’t just a computing story. It’s becoming an energy story. This piece looks at the proposed ~9 GW gas power project in Ohio tied to AI data center demand. The scale is staggering. Projects like this highlight how quickly electricity demand from AI could reshape power infrastructure planning and economics. One point that stands out: the cost of reliable, large-scale power for AI may be far higher than many expect. As hyperscalers push for dedicated generation and firm capacity, the traditional model of grid expansion and cost sharing starts to break down. It raises a bigger question for the energy transition: Who ultimately pays for the massive power buildout required to support AI? Worth the read. https://lnkd.in/e36NsBd5

21

1 Comment

Seed-stage fundraising requires differentiation. Grind Ventures found it at the intersection of three underserved markets Technology transfer from universities creates early-stage opportunities that most VCs miss. Defense and federal marketplace applications provide revenue visibility. Critical infrastructure sectors offer commercial scale. Focused Thesis - Dual-use technologies serving defense and 16 critical infrastructure sectors Technology Transfer - University research teams1-5 years old at $1-3M valuations Revenue Clarity - Federal marketplace provides structured procurement pathways Commercial Scale - Critical infrastructure sectors offer massive addressable markets   Geographic Focus - US-based companies with clear domestic advantages Stage Discipline - Seed investments with defined entry points and milestones The market rewards problem-obsessed founders with healthy margins and cash flow visibility. Dual-use technology delivers both. Which of these thesis elements resonates most with your investment approach? #VC #GrindVentures #ThoughtLeadership

19

6 Comments

🚨NEW EPISODE: Recorded live at FUTURE TITANS 2026 - Jeff Perry of Carta sat down with the iconic Seth Levine, co-founder of Foundry. Seth has been in venture for 25 years, built Foundry from scratch as an emerging manager himself, and has backed about 50 emerging manager funds through his fund of funds. He has genuinely seen every side of this table. They went deep on building Foundry, why VCs are in the influence business, not the decision business, and why the concentration problem in venture is not only bad for LPs, but also for the innovation ecosystem overall. And why Seth's new book, Capital Evolution, is so important for the future of America. 🎧 Links to listen... Apple: https://lnkd.in/ehQUQ2EM Spotify: https://lnkd.in/eU4FExpg

18

1 Comment

AI isn't just powering software. BrightAI just raised $51 million to automate inspection and other hardware-based AI capabilities. As Alexa von Tobel, co-founder of Inspired Capital put it: “So many people are so focused on the future of digital AI, but we’re excited about this new layer of AI: the physical world AI.” We are in the early innings of a series of AI-fueled innovations! https://lnkd.in/enhmafvH

47

Working with NRF to pursue our patent strategy is an important step in recognizing that proprietary AI systems and workflows are becoming a new form of intellectual property for modern firms. Great to partner with George, John McBride, and the team at Norton Rose Fulbright as we work at the frontier of how AI can transform investment decision-making in venture capital. At Decasonic, we believe the next generation of companies will reinvent with AI workflows and systems how organizations operate, how leaders make decisions, and how teams collaborate with intelligent systems. In going through this streamlined process, it's clear to me that many more startups should clarify their IP strategy. Our internally developed AI platform, RLEN, reflects that philosophy. We are building AI not to automate tasks and flourish outcomes, in collaborative ways for human and AI teammates. This enables our investors to test ideas, explore alternative perspectives, and reach deeper insights faster when navigating uncertainty. More broadly, we are seeing trailblazing work across industries and cities as leaders experiment with how AI can reshape communities, companies, and countries. Leadership with AI forever transforms. We are still in the early innings. Appreciate the partnership and collaboration from the Norton Rose Fulbright team as we continue inventing and innovating. Special shoutout to my colleague Abdul Al Ali who's listed as a co-inventor.

21

1 Comment

Seed Stage Venture Debt? Most early founders think venture debt is something you raise at the Series B or later. It is seen as something you layer on once the business is fully “bankable” and later stage. But what if this paradigm is limiting growth for early stage companies? And what if founders can utilize a flexible product at the early stages to reduce dilution? At Top Corner we see it differently: - We underwrite The Founder and to enterprise value, not formulaic metrics - We believe debt can work much earlier, typically right after product-market fit (Seed/Series A) - Our debt is designed to support founders, not constrain them - Our structures are flexible, with none of the traditional debt mechanics that weigh down early stage founders (financial covenants, MAC defaults or investor support clauses) - Our process is fast (think weeks, not months) - Our capital is significantly less dilutive than equity, but can be used just like a dollar of equity: to hire, build, experiment and scale When used for growth, debt can help teams hit more milestones and empower founders to raise their next round on stronger term, all while preserving meaningful ownership for both founders and early investors. A recent example: Top Corner partnered with NxLite in February 2025, providing Seed stage debt capital that allowed the company to acquire critical equipment and fully prepare its 45,000 squarer foot production facility prior to raising additional equity. NxLite produces advanced coatings for windows and transparent barriers that dramatically reduce energy consumption and support a more sustainable built environment. With this early capital, NxLite achieved production capacity to serve a strong roster of Fortune 500 enterprises, creating a strong foundation for its next phase of growth. NxLite closed a $9.2M Series A in September 2025 at a meaningful valuation increase. We’re excited to support David Mather and the NxLite team as they build on this momentum. https://lnkd.in/g-Fipd9K

37

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

Proud to share that Greenfield Partners has led GoodShip’s $25M Series B, combining forces with our friends at Bessemer Venture Partners, Ironspring Ventures, Chicago Ventures, and FUSE! For anyone who has managed freight, the pain is familiar: scattered data, reactive decisions, and too many spreadsheets. GoodShip is building a modern operating system that brings planning, procurement, and performance into one place so transportation teams can act faster with measurable results. By unifying data across systems, teams get real-time visibility and clearer levers. GoodShip is already trusted by the leading enterprise shippers like Tropicana, WK Kellogg, McCain Foods, and KBX Logistics. “The freight industry is moving toward a new operating standard, with GoodShip at the forefront of this shift” says Itay Inbar. “GoodShip is redefining a trillion-dollar industry with a novel approach and rapid AI innovation - unifying procurement and visibility in a platform that’s already proving its value at enterprise scale across the world’s largest shippers.” The team, led by Ryan Soskin and David Tsai, brings deep freight experience and a builder’s mindset, and it shows in every customer conversation. Already, we are hard at work with the GoodShip team in scaling their go-to-market. Our shared goal is straightforward: help more shippers move from reactive decision-making to confident execution. This is the start of an exciting journey ahead. Congratulations to Ryan, David, and the entire GoodShip team—I’m grateful for the partnership and energized for what comes next! Itay Inbar Olivia Levine

129

6 Comments

https://lnkd.in/gcrq9UNj The New Laws of Corporate Hype About AI for 2025: Mueller's Law of Rebranding 2.0: Any technology becomes “agentic” if you add the word “autonomous” to it. The Extended Gartner Paradox: A technology reaches the peak of the hype cycle exactly when no one can explain how it really works anymore. The Law of Quantum Marketing: A product can exist simultaneously in two states: “completely obsolete” and “revolutionary,” until it is observed by an investor. The Venture Capital Axiom: The amount of funding received is directly proportional to the number of times the words “autonomous,” “agentic,” and “multimodal” appear in the pitch deck.

11

1 Comment

Jazzed for this keynote by Jeff Erhardt at our 12th annual LDV Capital Vision Summit: The Materials Innovation Gap: Why AI Predictions Aren't Enough — And What It Will Take to Transform a $6 Trillion Industry RSVP: https://lnkd.in/eeUHN_m AI is transforming industry after industry, and generating enormous excitement in the world of materials science. Machine learning models can now predict millions of new material candidates with unprecedented speed and accuracy. At the same time, programmable material classes, including those recognized by the 2025 The Nobel Prize in Chemistry, could unlock breakthroughs in some of the most pressing challenges facing civilization, from environmental remediation to the energy transition. But their impact is constrained by something prediction alone cannot resolve: new materials must be physically produced, scaled, integrated, and qualified within the larger systems into which they are deployed, a journey that remains extraordinarily long, expensive, and unpredictable. Jeff will argue that closing this materials innovation gap requires treating AI and physical experimentation as equal partners, and building a new model of collaborative development that connects materials innovators and industrial partners far earlier in the discovery process.

12

Insightful notes from Aditya on the ABB Robotics sale to Softbank: "ABB is essentially selling a data goldmine: its installed base of nearly 500,000 industrial robots worldwide could have served as a massive real-world data collection network, generating the operational insights needed to train successive generations of increasingly intelligent robotic systems. By exiting now, ABB trades long-term optionality in the fastest-growing segment of industrial automation and may look short-sighted if Physical AI follows the explosive adoption curve of generative AI."

17

We've officially entered into manufacturing - and we’re doing it with a powerhouse partner. Lantern Partners AI is proud to announce our partnership with Estee Bedding Company, the nation’s leading truck mattress manufacturer out of Chicago with 12 facilities and 45 distribution centers across the U.S. Estee has deployed LanternBRP™ (our AI-native, modular Business Resource Planning platform) to replace a well known legacy ERP that had become a bottleneck - and they’re projected to drive $4.2M in annual savings by improving: ✅ Bill of materials (BOM) management ✅ Inventory control and imbalance reduction ✅ Production scheduling and downtime prevention ✅ Sales visibility and faster decision-making What makes this meaningful isn’t just the software - it’s the shift: mid-market manufacturers deserve operational intelligence that matches the pace of their business, without the rigid, fragmented tools they’ve been stuck with for years. As Estee’s CEO Tim Enright put it: their former ERP couldn’t keep up. LanternBRP™ gives them one coherent platform so teams can operate as a synchronized unit — and build for what’s next. This is the first step into manufacturing, and we’re just getting started. 🌐 Learn more: LanternPartners.ai #lanternpartnersai #Manufacturing #ERP #SupplyChain #Operations #SaaS #AI #DigitalTransformation #MidMarket #Industry40 #BusinessIntelligence

51

7 Comments

Camber Creek led the $3.7 million Seed round for TwinKnowledge, which is tackling an important problem in the built environment. Construction projects get delayed and rework is required because of discrepancies across construction documents, submittals, requests for information, design, and other plans. TwinKnowledge’s AI Agents absorb structured and unstructured project data, surfacing and preventing conflicts before the work begins. We’re excited about our partnership with CEO Ivan Panushev and his team. #construction #AEC #AI #technology https://ow.ly/mBYX50VtaVG

10

There’s a massive opportunity in automating industrial manufacturing processes. That’s why we’re backing RobCo in their series C as they scale their physical AI driven robotics platform and expand their U.S. presence. As global labor supply continues to tighten, automation in manufacturing is shifting from “nice to have” to essential. RobCo's Robotics-as-a-Service model removes traditional friction, enabling rapid deployment and real production output in weeks – faster than most B2B software implementation timelines. In Roman Hölzl we immediately saw that rare leader – one who pairs a bold, long-term vision for transformative AI with a relentless focus on real-world productivity today. We couldn’t be more excited to back RobCo as they execute on an ambitious roadmap! Press release: https://lnkd.in/euCx6WX2 Roman Hölzl, Dennis Hauser, Arjan van Staveren, Stefan F. Heidenreich Lightspeed, Sequoia Capital, Kindred Capital VC, Greenfield Partners, Leitmotif, The Friedkin Group, Alexander Schmitt, Luciana Lixandru, Leila Rastegar Zegna, Marc Sieberger, Raz Mangel #Lingotto #LingottoInnovation #PhysicalAI cc Clemente Dadoo Lomeli

154

8 Comments

Terra Innovatum and GSR III Acquisition Corp. File Form S-4 After Business Combination Announcement Advancing Public Listing. Transaction To Accelerate Terra Innovatum’s Pathway to Commercialization of Groundbreaking Micro-Modular Nuclear Technology by 2028. https://lnkd.in/eNpGF9bU

36

4 Comments

Today we launched Seraphim’s 2025 Ecosystem Map. The headline: space is maturing with discipline. Capital is spreading to a broader set of high-quality companies; business models are stronger; dual-use demand is anchoring revenues; and public markets are starting to reopen. What’s new this year: • A dedicated Software category • In-space services shifting from demos to repeatable operations • A healthier balance of Build & Launch through to Analyse & Product At Seraphim our focus is simple: back category leaders, connect defence and commercial demand, and keep standards high so this sector compounds for the long term. Credit to the founders and operators setting the pace featured in the map. If you want to learn more about this year's Map, Seraphim's Lucas Bishop discussed it with Via Satellite: https://lnkd.in/eyni5DRH #SpaceTech #VentureCapital #SpaceEconomy #Innovation #SeraphimSpace #DualUse

329

21 Comments