Picus Capital

Anyone else notice SAP cooking yesterday?👇🏻 New partnerships and investments announced: ► n8n as the orchestration layer. SAP added an investment on top of the existing partnership, valuing them at $5.2B. Also, they are embedded inside Joule Studio, SAP's agent-building environment. [1] ► Parloa landed the service layer via a new partnership bringing agents into SAP Service Cloud, plus a strategic investment building on its Series D. [2] ► Conduct grabbed part of implementation via a new partnership on AI-powered cloud ERP migrations. [3] This comes on the back of two recent acquisitions: Dremio (data platform for agents) and Prior Labs (tabular foundation models). [4] Congrats to everyone involved! 🚀 Anything I missed? :)

Germany's Fachkräftemangel debate is everywhere. In hospital billing, the shortage is not people. It is technology. We have been digging into AI-powered revenue cycle management (RCM) for German hospitals. The pattern is consistent: hospitals have bridged a technology gap with people for decades. German hospital revenue depends on accurate ICD/OPS coding, the standardized diagnosis and procedure codes that determine reimbursement. Most hospitals still do this manually, at roughly 30-50€ per inpatient case. Leading hospital groups have 40+ open coding vacancies, and the talent pipeline is not growing. Complexity makes it worse. Inpatient DRG, outpatient EBM, and private GOÄ each follow different catalogues, rules, and edge cases. One missed or miscombined code triggers an audit or leaves revenue on the table. Hospitals with the weakest billing quality face the highest audit rates plus a per-case surcharge under the KHVVG. The system punishes the understaffed. The bigger insight from our diligence: better code combinations are not just an efficiency play, they actively lift hospital revenue. No one has solved this yet. Traditional HIS vendors (Dedalus, CompuGroup, SAP IS-H) move slowly, with coding buried as a secondary feature. Modern German challengers like Avelios are building HIS 2.0 with embedded coding, but require hospitals to migrate off their core systems. International AI-native RCM players like Nym Health, Abridge, and CodaMetrix have proven the model at scale, but none have localized for German catalogues. We spoke to 16 experts across hospital groups, healthtech vendors, and industry insiders. The winner builds a lightweight AI layer on existing HIS, automates across all German reimbursement catalogues, and captures both cost reduction and revenue upside. Billing is the wedge; the data and trust open a path into adjacent workflows like clinical documentation, surgery planning, and denial management. If you have a perspective on where it's heading, I'd love to hear your take.👇

The Picus Capital team will be attending SuperReturn and SuperVenture next month in Berlin between 8-12 June! If you will be there and would like to connect, we'd love to catch up. Please feel free to schedule a meeting with the team below or drop by the #PicusBus 🚌 which will be parked opposite the venue. We look forward to seeing you in Berlin! Connect with the team ahead of the event: Florian Reichert, Raphael Mukomilow, Anthony Hayhoe

Together with TUM.ai, Aidera and Entire hosted an incredible evening in Munich focused on the future of software engineering in the age of AI. A big thank you to Thomas Dohmke, former CEO of GitHub and founder of Entire, and Sebastian Schaefer, Founder of Aidera (in stealth) and Partner at Picus Capital, for an inspiring discussion on how AI agents are transforming the way software is built. It was fantastic to see such a strong turnout from founders, engineers, students, and AI enthusiasts across the Munich ecosystem. Thanks to everyone who joined us and made the evening such a success.

📢 Spotlight on Marie Ch. Fabiunke: Turning Communications into a Growth Lever for Startups Today, the spotlight is on Marie. As the Co-founder & CEO of the PR firm Cosmyc Partners, she works with startups, scaleups, VCs, and deep-tech companies across Europe, helping them shape clear narratives, build visibility, and turn communications into a strategic growth lever. With a background spanning venture capital, technology, and media, she focuses on one core question: how can companies communicate in a way that builds trust, accelerates credibility, and supports growth? 🧠 Her Advice to Founders It’s key to lay the foundation for a strong reputation early on: It all starts with a clear mission, vision and purpose that shape how others perceive a company. Clear, specific positioning makes a startup more memorable and easier to trust, and a strong narrative helps people understand why your company matters, not just what it does. 🔍 Key Takeaway Reputation is a startup’s trust currency, it influences how investors, customers, and partners perceive and engage with the company. Founders need to define their position in the market, build their brand towards clients, investors and talent - before the market defines it for them.

🚀 We're beyond excited to announce our third investment in Blitzy, meaningfully contributing to their landmark $200M Series A at a $1.4B valuation, led by Northzone, with new investors PSG, and Battery Ventures, and continued support from existing investors Flybridge, Link Ventures, NFX, and Venture Guides. ⚡ Blitzy is the most advanced autonomous enterprise software development platform on the planet, purpose-built for large, complex codebases with infinite code context. Orchestrating thousands of AI agents in parallel to autonomously deliver 80%+ of enterprise-scale projects across millions of lines of code, from legacy modernization to full-stack feature builds. 🤝 After having worked closely with Brian Elliott, Sid Pardeshi and the Blitzy team since meeting them on campus at the Harvard Business School, we are incredibly proud to triple down. Picus meaningfully participated in Blitzy’s $4.4M pre-seed in mid-2024, and has invested multiple times since. From a pro bono project building apps for local SMBs, to the globally leading autonomous software development platform for enterprises. What an extraordinary journey so far, and all in less than two years. 🎥 📖 Check the comments below for the link to our video featuring Blitzy’s CEO, Brian Elliott, and our full 'Why We Continued to Invest' article. Congratulations to Brian Elliott, Sid Pardeshi and the entire Blitzy team. 🚀 CC Julian Roeoes, Robin Godenrath, Florian Reichert, Oliver Heinrich, Sebastian Schaefer, Raphael Mukomilow, Jay Patel, Cosima Paul, Darius Kahlert, Abhijay Thacker #AutonomousSoftwareDevelopment #AI #EnterpriseAI #SeriesA #PicusCapital #Blitzy #VentureCapital

We are thrilled to co-lead XO Market's $6M Seed round alongside our friends from 20VC, with participation from Coinbase Ventures, Ventures Together, Insiders and Foreword Fund. Joining us are 30+ angels, including Pat Cummins, Michele Attisani and Sidney Gakpo. XO Market is building permissionless conviction markets - the next evolution of prediction markets. Our hypothesis is that prediction markets need a fundamentally different architecture than incumbents, whose growth is constrained by manual market creation and weak long-tail liquidity. XO solves this problem first: demand-adaptive liquidity that dynamically deepens as activity spikes with no professional market makers required. Anyone can create Yes/No markets on any belief or event, turn convictions into tradable assets, and benefit from fast, transparent resolution powered by AI + decentralized arbitration. Ali Habbabeh and Walid Al Habboul are relentless builders who ship weekly and obsess over trust and settlement. With $1B+ annualized volume run rate and users across 30+ countries, their execution speaks for itself. Congratulations to the entire XO Market team. We are just getting started! 🚀 #PredictionMarkets #ConvictionEconomy #Web3 #Crypto

Last week, we welcomed an impressive group of future founders from the NextGen Entrepreneurship Network (NGEN) to the Picus Capital office in NYC, hosting them together with Miguel Armaza from Gilgamesh Ventures. We had the chance to hear a range of thoughtful, creative pitches, and it was exciting to see the level of ambition and curiosity in the room. Thank you to everyone who joined and contributed to such an engaging session, and a special thanks to Harsha Ravindran for organizing and bringing this group together. Amazing initiative! It's great to see a broad range of ambitious founders from The Wharton School, Cornell University, Columbia University, New York University, and many more universities. Looking forward to seeing how these ideas evolve!

AI is rewriting the economics of drug discovery, and the white spaces for founders are wider than most people think. Together with Spira Labs, we unpack the third of five opportunity spaces from our cross-industry study (see comments) on how AI is reshaping 13 sectors: drug discovery and development. We map where the economics are shifting and where early-stage founders can build. Full article below.👇 Let's explore these opportunities together.

Running AI pilots is easy. Turning them into value is where 95% of enterprises fail. We have been digging into enterprise AI orchestration, the layer that turns scattered GenAI pilots into business value. Enterprises have poured $30B-40B into GenAI, with 95% delivering zero measurable P&L impact [1]. Meanwhile, the orchestration market is projected to grow from $11B to $30B by 2030 at 22% annually [2], precisely because everyone is aware of the potential AI has yet to unfold at enterprise scale. The reason real impact is lagging behind is not model quality. Instead, enterprises run dozens of pilots across fragmented models with no shared memory, governance, or cost visibility, and only 5% of task-specific AI tools ever reach production. MIT calls it a "learning gap": tools that do not learn, adapt, or integrate into workflows stall out, regardless of the model behind them. More pilots are not solving this. Better orchestration is. The same dynamic shows up in adoption patterns. 90% of knowledge workers already use personal AI tools like ChatGPT for work, while only 40% of their employers officially deploy AI [1]. Bottom-up adoption is running ahead of procurement, security, and governance. That gap, between what employees are already doing and what the enterprise can safely operationalize, is exactly where the orchestration layer has to intervene. The market has tried to solve this in several ways, with mixed results: ► Hyperscalers (AWS Bedrock, Azure AI, ServiceNow) move fast but stay locked in their ecosystems. Microsoft and former AWS leaders told us end-to-end orchestration is still unsolved. ► Point solutions dominate the startup layer. Routing players (Martian, Portkey, OpenRouter), security tools (Lakera, CalypsoAI), and observability platforms (Arize, Langfuse) each solve one slice. None solve the whole. ► In-house builds are the most common path and the worst performing. For example, MIT found that purchased solutions succeed twice as often. Based on expert conversations across Microsoft, AWS, Datadog, IBM Watson, Stripe, and Fortune 500 AI leaders, we believe the winner will need three things at once: ✔ Full-stack and vendor-neutral orchestration across routing, governance, workflow integration, memory, and observability ✔ Learning-capable architecture that adapts to enterprise workflows and improves from user feedback over time ✔ Enterprise-ready from day one, combining trust, clear data boundaries, and on-prem deployment, especially for regulated sectors and European buyers The timing is now. Enterprise AI spend is shifting out of innovation budgets into permanent IT lines over the next 18 months. Orchestration is becoming a standard line item alongside identity and data management. No clear winner exists yet. Do you think this will be solved by an incumbent pivoting in, or by a new pure-play? And will the winner be a regional player or a global leader?👇 Interested to hear how others are thinking about it. #DDwithDarius