Allocator One

Deeptech Bridges wrapped last week at SF Tech Week, hosted at SAP in San Francisco. Co-hosted by Allocator One and OPEN AUSTRIA, with community partners Hello Tomorrow and EWOR. One question ran through the afternoon: how do we design better incentive bridges between VCs, CVCs and other asset classes, so that global deep tech is stewarded for the long term? The lightning panel took on the new deep tech capital stack: where sovereign, strategic and venture money meet, and where they don’t. Niko Woischnik (Allocator One), Steven Jacobs (Drumbeat Capital), Pae Wu (SOSV), and Heriberto Diarte (Catalyzer Ventures). Then the room split in two. Kyle Petersen (3M Ventures) and Heriberto Diarte (Catalyzer Ventures) asked what actually enables breakout returns in early stage hardware. Lidiya Dervisheva (Chromium Ventures, Walter Ventures Family Office) asked which global deep tech strategies are well positioned, and why. Our view, a week on: in deep tech, the edge belongs to the managers who can bring sovereign, strategic and venture money to the same table, early. Thank you to Stanish Gunasekaran and the Climate Deep Tech Bridges team, to Isabella Tomás and Michael Dobersberger at OPEN AUSTRIA, to the SAP team for the welcome, to Hello Tomorrow and EWOR, to every speaker, and to everyone who stayed on through the networking hour. The conversation continued this week at Norrsken Impact Week in Barcelona. Stay tuned for more. #AllocatorOne #InfraOne #DeepTech #SFTechWeek #VentureCapital #EmergingManagers #OPENAUSTRIA #SanFrancisco #AllocateOrStagnate

Fredrik Debong was four, had dropped ⅓ of his body weight and his parents and doctors didn’t know what was going on. Type 1 diabetes, it turns out; he and around 10 million people live with it. Since then: 170 therapy decisions a day, for four decades. So he co-founded mysugr to make diabetes "suck less". Cleared as a medical device in Europe and the US, now serving >6M people in 70 countries. Bought by Roche. It wasn’t enough. Next he built a fintech in health insurance, hi.health. Removed payment from the patient journey for 1M people. Bought by Pliant. And before that, he cofounded Pioneers, which created the Austrian startup eco system. Bought by startup300 AG. It wasn’t enough, so he doubled down. "The world is big. You can't work on everything yourself." So he started funding the people who could. Nineteen Twenty-One is named after the year insulin was first isolated, the year his disease stopped being a death sentence. It backs ideas that are "digital, biological, data-driven, infrastructure, and gloriously weird." He calls himself a "serial entrepreneur (aka idiot)". By his own count, a life with type 1 has held three children, visits to almost 100 countries, mountain running, a string of companies, free diving… he also knows the other side: "Without insulin and support, you die." He now spends his weeks making sure more people get both. In type 1 diabetes, a yes from Fredrik is a seal of quality and aim. Venture is moving from generalists to specialists. Fredrik is about as specialist as it gets. Why we anchored him: - Patient, founder and investor in one person. Founders in his field don't explain the disease to him. He has lived it for 40 years. - He has already done the impossible part. Regulators on two continents, 6M users, a sale to Roche. He knows where health companies break, and how to get through. - Every company he backs makes the next one sharper. One disease, every angle, one person connecting the dots. - He wants to put himself out of business. "If 1921 ends up being my only fund because we have a cure, that is completely fine." Most GPs raise Fund I hoping for Fund II. Fredrik is hoping for a cure. We anchored 1921 to be in the room when it comes.

What does an anchor-LP fund of funds actually look for in Emerging Managers? I sat down with Michael Ströck and Felix Staeritz for one of my favorite Embracing Emergence formats: LP interviews. I wrote it all down and documented our honest conversation. There are some real gems in there!! We talked about: 🏡 What started Allocator One 💡 What gave them the conviction to start a fund of funds that anchors Emerging Managers 👀 What does A1 look for in the GPs we back 📊 How important is a track record? The newsletter issue is in the comments below ⬇️

Arian Ghashghai was pseudo-homeless in the weeks before his first day at Facebook. Today Marc Andreessen, Chris Dixon, Adam Draper are LPs in his fund. First he taught himself to code, then in 2013 he went all in on a music startup. It failed, but it pulled him into machine learning, and he never looked back. That obsession took him from a graduate without a fixed address to Meta in January 2019, he started on integrity work in the middle of the misinformation crisis and ended up deep inside Meta Reality Labs, building spatial AI for edge hardware as a senior ML engineer. The angel cheques started quietly, on the side. Then he did what most people only talk about: "I left my cushy role as an overpaid senior ML engineer at Meta ... to burn my savings to pursue the opportunity to build this venture firm." His reasoning: "Going all-in on asymmetric or non-consensus convictions (and being right) is the only chance to win big when starting from 0." From inside Reality Labs he watched AI leave the screen and start reshaping the physical world. As an engineer he could build one piece of that shift. As an investor he could back the founders rebuilding entire industries. In 2025 he announced Earthling VC's $5M Fund I, raised "in a historically brutal market for new funds, and as a new GP with no prior professional VC background." His LPs include Marc Andreessen, Chris Dixon, Andrew Chen and Adam Draper. He writes first cheques into physical industries before they're obvious: microagi, Loki Robotics, @Continuum VR, REK, Reactor and Syncere. Durable alpha is formed upstream, before an opportunity becomes consensus. Arian has spent his whole career upstream. Here’s why we wrote a cheque into Earthling Fund I in 2025: • A builder before a backer. He shipped spatial AI at Reality Labs for years, so he can read a founder's technical stack without outsourcing the diligence. • Non-consensus on purpose. "Going all-in on asymmetric or non-consensus convictions (and being right) is the only chance to win big when starting from 0." • Skin in the game. He put his own savings on the line to start the firm, then raised Fund I anyway. • A taste for the strange. Robotic guide dogs, AI agents setting Pokémon world records, a social VR games studio. He goes where the next interface is being built, long before it looks obvious. When you start from zero, conviction is the only edge you get. Arian has never traded it for comfort. Arian, keep going all in.

Since January 2025, the US government has signed 37 equity or quasi-equity deals worth roughly $27.6bn. About 10% of Intel in August 2025. Nine quantum companies in a single day this May. Most investors read this as a policy story. At Allocator One, we read it as a capital stack story. Sovereign money runs on election cycles, corporate money on product roadmaps, venture money on a ten-year fund. Three clocks, one company. Europe sits on the other side of the same equation. Deep tech took a record 32% of European venture in 2025, $20.3bn. The growth rounds still come largely from abroad: in Switzerland, foreign investors supply 88% of late-stage deep tech funding. The science is formed in Europe. The scale-up capital sits in the US, and the US government is now part of it. The early data already points to what works. Jefferies found that the government deals with an offtake or price floor attached, like MP Materials, held up. Equity alone did not. Hardware investors will recognise the pattern from their own portfolios. Capital without a customer is not conviction. That is why Allocator One is co-hosting Bridges at SF Tech Week this Wednesday, together with OPEN AUSTRIA. What to expect: Lightning panel: The new deep tech capital stack, where sovereign, strategic and venture money meet (and don't) Niko Woischnik (Allocator One), Steven Jacobs (Drumbeat Capital), Pae Wu (SOSV), Heriberto Diarte (Catalyzer Ventures), Victoria Slivkoff(Chemonics) Roundtable 1: What actually enables breakout returns in early-stage hardware? Kyle Petersen (3M Ventures), Heriberto Diarte(Catalyzer Ventures) Roundtable 2: Global deep tech strategies: who is well positioned, and why? Petr Smid (Rockaway Ventures), Lidiya Dervisheva (Chromium Ventures, Walter Ventures Family Office) Then an hour of open networking at SAP Labs, with finger food and Austrian drinks from OPEN AUSTRIA. 🗓️Wednesday 7 October, 3:30 PM PT 📍SAP Labs, San Francisco. ✍️RSVP and sources in the comments. Co-hosted by Bridges, OPEN AUSTRIA and Allocator One, with ecosystem partners EWOR, Plug and Play and Hello Tomorrow. Thank you to the Bridges and OPEN AUSTRIA teams for building this with us. If you deploy into deep tech and you are in town for Tech Week, come with the question your own IC cannot answer. Durable alpha is formed upstream, before the opportunity becomes consensus. #AllocatorOne #InfraOne #Bridges #SFTechWeek #DeepTech #CVC #FamilyOffices

A History graduate from Brown University walked onto a New York capital markets desk in 2007, months before the financial system began to collapse. Today Priya Shah backs electric air taxis, EVs that charge in 15 minutes, and the clean-up of India's heaviest industries. Most people who survive their first job in a crisis stay in finance for life. Priya treated Bloomberg as a training camp. "I was excited about building my technical skill set in finance and being able to apply that as a transferable skill later on." Then she went and collected every skill the energy transition would one day demand. Policy at Asia House in London. An MBA at Cambridge Judge Business School. Development finance with Acumen and CDC Group plc. A solar business in India, Simpa Networks. The India fund of Yunus Social Business, the organisation co-founded by Nobel Peace laureate Muhammad Yunus. By the time the energy transition became venture's favourite theme, she had already spent a decade in the rooms where it actually gets decided. From those rooms she saw a gap she could not walk past. India had the engineers, the factories and the urgency. "There's no shortage of talent. What's lacking is visibility, access to early capital, and credible networks." And for her it was never just a market: "it's important to do a service for the next generation and help save whatever we can." So she built Theia Ventures in Bangalore. Its portfolio reads like a preview of a different India: Sarla Aviation, building electric air taxis and now backed by IndiGo (InterGlobe Aviation Ltd); Novyte Materials, using AI to design new materials instead of finding them by trial and error; Lemnisca, an AI companion for scaling up fermentation; and Climitra and DashamLabs, taking on heavy industry and low-carbon materials. Then the plot twist. British International Investment, the UK's development finance institution, anchored her fund. Years earlier, when it was still called CDC, Priya had been consulting for it. The institution she once advised came back to back her. The Cisco Foundation AI and Anand Mahindra followed. We came into Theia Fund I in 2025, because almost no one else has stood on every side of this table. Banker, policy hand, development financier, operator, fund builder. She goes where the energy transition is hardest to underwrite, heavy industry, materials, logistics, and she goes there in person: "the best insights don't always come from the lab. They often come from the field, the farm, the factory." She also knows who is watching. "Being a sole fund manager I'm hopefully setting an example for other young women to believe that this is a role they can do." She started on a trading floor. Now she is rebuilding the factory floor. "The goal is to catalyze, not just capitalize."

We are co-hosting Bridges at SF Tech Week. A deep tech company can be born in a European lab and raise its growth round in San Francisco. Nearly 10% of European scale-ups relocate, and 85% of them go to the US. The reason is not talent. It is the capital stack. Europe deployed about €66 billion in venture in 2025, roughly 22% of the US, for economies of similar size. At seed the gap has nearly closed: top European deep tech rounds now price within 15 to 20% of US equivalents. At scale it opens again: US companies raise $100M+ deep tech rounds about five times more often than European ones. We see roughly 80% of new fund managers, and this is the pattern we watch most closely. Three layers of capital are learning to work together. Venture prices the risk. Strategic capital validates the market and opens the first sites. Sovereign and LP capital funds the scale a venture clock cannot wait for. The misalignment between them is the most important and least discussed problem in deep tech. So together with Bridges we are putting all three in one room. One afternoon on the new deep tech capital stack: where sovereign, strategic and venture money meet, and where they do not. • Lightning panel: Niko Woischnik (Allocator One, LP), Steven Jacobs (Drumbeat Capital , VC), Victoria Slivkoff (Chemonics, CVC) • Roundtable 1: What actually enables breakout returns in early-stage hardware • Roundtable 2: Global deep tech strategies, who is well positioned and why. Moderated by Petr Smid (Rockaway Ventures) Built for emerging GPs, CVCs and strategics, LPs and family offices investing in deep tech and hardware, with a soft spot for very early-stage fund managers. Find them first. 🗓️Wednesday, 7 October, 3:30 to 6:00 PM PT. 📍SAP HQ, San Francisco. Curated, approval required. RSVP link in the comments. Hosted by Bridges, Co- Hosted by Allocator One. Community partners: OPEN AUSTRIA and Hello Tomorrow #AllocatorOne #Bridges #SFTechWeek #DeepTech #EmergingManagers #VentureCapital #FamilyOffices #CVC #SanFrancisco

Venture has never had more money. But the middle rounds have never been in more trouble. US startups raised $412.7B in the first half of 2026, more than in any full year on record. Yet more than 80% of startups that raised a seed round in 2022 still hadn't raised a Series A two years later. Both are true because startups now do one of two things: 𝐫𝐚𝐢𝐬𝐞 𝐨𝐧𝐜𝐞 𝐚𝐧𝐝 𝐧𝐞𝐯𝐞𝐫 𝐧𝐞𝐞𝐝 𝐚𝐧𝐨𝐭𝐡𝐞𝐫 𝐫𝐨𝐮𝐧𝐝, 𝐨𝐫 𝐫𝐚𝐢𝐬𝐞 𝐛𝐢𝐥𝐥𝐢𝐨𝐧𝐬. The rounds in between are disappearing. For thirty years, every venture round paid for the next wave of hires. I call it the 𝐡𝐞𝐚𝐝𝐜𝐨𝐮𝐧𝐭 𝐥𝐚𝐝𝐝𝐞𝐫. AI just knocked out the middle rungs. 𝐓𝐡𝐞 𝐓𝐋;𝐃𝐑 𝐨𝐟 𝐦𝐲 𝐧𝐞𝐰 𝐞𝐬𝐬𝐚𝐲: 01 - Scale no longer needs people. $10M in revenue used to take ~77 employees and several rounds of funding. At the fastest AI-native companies, it takes ~9. That's a seed round. 02 - The money has left the middle. 87.5% of US venture dollars in H1 2026 went into rounds of $100M+. 03 - The Series A stops being a signal of real progress or product-market fit. A profitable company and a zombie now look identical on paper. Underwrite the business, not the round. 04 - Small funds become the smartest money in venture. Exit needed = fund size ÷ ownership. At 10%, a $20M fund needs a $200M exit to return it. A $1B fund needs $10B. That's 50x bigger for the same result. 05 - LPs will be tempted to back the wrong managers. No new rounds means no markups, so the best portfolios may look terrible for years. Everyone is waiting for the Series A to recover. I'm betting it won't. In 2018, roughly 1 in 3 seed-funded startups raised a Series A within two years. Neither the 2025 nor the 2026 seed class will get back there. Full essay in the comments.

Nicola Sinclair spent nearly two decades in the Royal Air Force (RAF). Then she built the RAF its own venture fund. Most of her career happened far from term sheets: international operations, the teams planning them, and a posting at NATO HQ working on how allies share information and plan together. In 2018 she founded the Air Innovation Fund inside the RAF Rapid Capabilities Office. Pre-Series A only, in energy, future compute and hypersonic flight. 20+ of its companies went on to win UK and US defence contracts. Then, about to raise her own defence fund, she did something unusual. She published an essay asking "Is defence tech VC backable?" Her answer: usually no. Today Twin Track Ventures backs the exceptions: deeptech that wins a defence customer early, then scales into commercial markets. The portfolio includes Cassi.ai , Uplift360, Hadean, SE3 Labs and Oshen. We joined Twin Track's first close in 2025. The NATO Innovation Fund (NIF) came in as an LP after us. Here is what we saw: - Early before it was a category. In 2018, she was already running a venture fund inside the RAF. Most VCs wouldn't take a defence meeting for years after that. She knows which programme has budget, which trial matters and how long procurement really takes, because she spent two decades on the buying side. - She found the flaw in her own asset class. Everyone else pitched the spending boom. Nicola asked the question that kills most defence pitches, then answered it herself: "Tech upgrade and obsolescence cycles are now six weeks or less." A product built for a single programme can be outdated before the contract is signed. So she only backs what survives that: deep IP with more than one market. - She turned two reasons to pass into one reason to invest. Defence is slow to buy. Deeptech is slow to find customers. Most investors see two problems. Nicola saw that they bite at different moments, and built a fund where each track covers the other's gap. The name is the thesis. - Inside the Alliance's trusted capital circle. Since then, Twin Track has been admitted to the NATO DIANA Capital Network. It is a selective group of investors, each vetted on track record and on NATO's Trusted Capital standards. Membership gives early access to startups DIANA has tested against Allied needs, introductions to co-investors across NATO, and briefings from defence ministries on what they actually need to buy. Nicola sat on both sides of the procurement table. Few investors can say that. Proud to be on the twin track with you.

It is always a good time for ice cream. If you are an emerging GP or LP please join us in NYC this Thursday. Feel free to share. Link to RSVP in comments.