Will Wells

Very excited to share that Plural is leading a $60 million Series A in CoMind, the startup building breakthrough optical sensing technology to revolutionise brain monitoring and treatment. For decades, doctors treating critically ill patients have been forced to compromise when monitoring the brain. Today's approaches use risky, expensive, highly invasive procedures that require drilling a hole into a patient's skull, or rely on inaccurate non-invasive monitoring that can compromise treatment decisions. To take just one example of why this matters, in the US alone 3 million patients suffer traumatic brain injuries (TBIs) every year, but only 5% receive an intracranial pressure test, which requires drilling a hole in the skull and carries a 15% complication rate. The rest are treated with limited information, leading to worse data, worse outcomes and higher costs. London-based CoMind is changing that. Its optical neuromonitoring platform can non-invasively measure three core brain health indicators - cerebral blood flow, autoregulation and intracranial pressure - something that has never been done before. It’s a breakthrough that could help tens of millions of patients worldwide and redefine how we understand and treat the brain. The company is led by a true outlier of a young founder, James Dacombe. He started programming when he was 13 and founded CoMind when he was just 17. His work ethic and rapid learning rate have led him to build a working product and take it through two successful clinical research studies, with the help of a stellar leadership team of medical and industry veterans from the likes of Medtronic, Tufts Medical Center and UCL.  This is only the beginning. CoMind plans to expand its product portfolio to measure additional physiological parameters and develop CoVision, an AI platform that transforms sensor data into predictive insights, identifying complications early and personalising treatment. The company aims to gain regulatory approval within two years, unlocking access to the huge ICU market in the US. And while this first sensor can revolutionise TBI monitoring, a second device is already in development to measure more parameters and treat more brain conditions. This technology has the potential to dramatically improve how we monitor our brains, giving doctors better information and more choices for treating patients, while building out a whole new market by providing a far more affordable and safe alternative to existing solutions. Read more about why we invested in CoMind here: https://lnkd.in/e-NPgDVp And Tim Bradshaw’s write up in today’s Financial Times here: https://lnkd.in/eAynvqPF

266

13 Comments

The energy cost of AI has been an open question for a long time but Vrushali Gaud the head of Google's Climate Operations team just dropped some stunning data at the #BNEFSummit. The median Gemini query now have the climate footprint of 9 seconds of TV viewing & 5 drops of water. While this is for a text prompt, it's a 33X improvement in just the past 12 months and is radically lower than any past estimate we've seen in this value chain. It demonstrates the extent of impact the teams and better tools have led to in this space. The release with a link to the technical paper for those curious is linked below. It includes their benchmark against public data from other AI providers as well as the breakdown of each of these components including the active accelerator energy use. Fascinating weekend read - https://lnkd.in/gdXa_qAR

69

News broke last week of our investment in Level Zero Health – a startup developing a continuous hormonal monitoring device using innovative DNA aptamer sensors to detect hormone levels in interstitial fluid (think: CGM but for hormones like progesterone, estrogen, cortisol, and testosterone). Hormonal health is poorly understood, inconsistently treated, and frustratingly administered - especially for those navigating IVF, menopause, or testosterone therapies that require inconvenient, invasive, in-person blood draws. At Black Jays Ventures, our mission is to back Founders building businesses that improve the lives of consumers. Ula Rustamova (former Palantir Technologies Tech Lead) and Irene Jia (former Philips R&D) are a formidable team with deep technical and research experience. We’ve been blown away by both the thoughtfulness with which they are building as well as the scale of their ambition (hint: longitudinal hormonal data sets). We are very excited to be backing them and their vision for delivering personalized hormonal healthcare. More on Why We Invested in Level Zero Health below. Jay H. Kim, Tim Thimmaiah, Brian Muller, Black Jays Ventures Irene Jia, Ula Rustamova, Level Zero Health Daniel Graf, Philip Kneis, Courtney Leimkuhler https://lnkd.in/e8aQyqZv

105

3 Comments

We just launched a new LLM that was fine tuned to reduce false positives in watch-list screening. The headliner results beat all expectations - ◼️ 98% accuracy in name degradation testing (vs ~80% for others) ◼️ True positive to false positive ratio improved from 100:1 to 3:1 ◼️ 100% recall during testing across X PEP/sanctioned individuals ◼️ 20X reduction in false positives per 1000 checks At a high level, we trained the models to understand the context surrounding each potential hit. Then gave them thousands of historical examples of known analysts' resolutions to learn the difference between true positives and false positives. Our own Joshua Linn wrote a deep dive article explaining their process. Link in comments.

149

5 Comments

Drove to the Calculus hacker house with Katherine Jones, investor at Hummingbird Ventures ($500M+ AUM), backing bold founders at the frontier of health x tech Talked about deferring Oxford, going from hospital corridors to neuro labs, to boardrooms, and growing up across Europe without a fixed identity. From the Edge Esmeralda proto-city, to psychedelic therapy, to post-AGI tribal living, body intelligence, and a future where bio-optimism wins.

226

29 Comments

AI's explosive growth is placing unprecedented demands on our infrastructure, with networking and cooling being the bottlenecks. That's where Auradine comes in. I'm excited to share that our portfolio company, Auradine, has secured over $150 million in the Series C funding round, bringing its total funding to over $300 million. This investment will fuel Auradine's mission to deliver scalable, sustainable, innovative blockchain and AI infrastructure solutions. Auradine is not just building for today but for the future. Auradine is tackling critical bottlenecks: - For AI, they are tackling the bandwidth and cooling crisis in data centers with game-changing AuraLink AI solutions based on Ultra Accelerator Link (UAL) and Ultra Ethernet Link. Today, they announced a high-speed, low-latency networking platform that enhances performance, reduces bottlenecks, and maximizes GPU efficiency. - For Blockchain, they are setting new standards for energy efficiency in Bitcoin mining with their Teraflux™ miners. Over 40 leading Bitcoin mining companies deploy their product today. This is a testament to the vision and execution of Co-founder and CEO Rajiv K., Co-founder and COO Barun K., and their incredible team. Their commitment to open standards, sustainability, and rapid innovation is inspiring. As I shared in our discussion, I wanted to invest in Auradine to bring "Silicon" back to "Silicon Valley." They are well on their way to achieving that. Want to hear more about how Auradine is shaping the future of blockchain and AI infrastructure? Watch my in-depth video conversation with Rajiv and Barun about their journey and the exciting road ahead. What are your thoughts on the blockchain and AI infrastructure market? I'd love to hear your insights in the comments! #AIInfrastructure #Blockchain #DataCenter #AI

753

29 Comments

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

HUGE releases from dbt Labs today. So pumped about their new Fusion engine. The Fusion engine is a total rewrite of dbt, in Rust. What's great about it? It's SUPER fast. Up to 30x faster compile times Fusion comes with deep SQL comprehension - baked into the Fusion engine is a SQL compiler, which allows you to validate code before attempting to run it. This also makes column level lineage possible State awareness. dbt will now understand model and data freshness while running. Now models won't re-run if the underlying data hasn't changed What does this all lead to? A MUCH better dbt experience, and meaningfully more efficient compute usage Read more here: https://lnkd.in/g9aFiPeA

113

1 Comment

Everyone loves the 'picks and shovels' thesis—but the real untapped goldmine? Vertical agents and applications. 💡 As the underlying foundation models continue to advance, there won't be much space to build unicorn businesses that sit between the models on one hand and agentic applications on the other. (Don’t be surprised if GPT-5 is natively agentic in its architecture and capabilities.) The real opportunity lies in massive end markets ripe for "agentization": industries centered on structured, repeatable activities. Industries • Data Management • Customer Support • E-commerce Operations • Finance and Accounting • Marketing • Human Resources • Healthcare Administration • Supply Chain and Logistics • IT and Software Support • Legal and Compliance Digital Jobs • Virtual Assistants • Content Moderation • Market Research • Quality Assurance • Online Tutoring • Social Media Management • Transcription Services • Digital Advertising • Graphic Design • Software Development Support At Mucker Capital, we've built over a hundred companies in vertical SaaS. We see vertical agents as it's next generation.

43

6 Comments

We had Tom Charman (Co-founder at Blok) on Ventures, Fresh out of stealth having raised $7.5m He explains why their star-studded team is building Blok "A lot of companies aren't particularly good at predicting how their customers are going to behave, but more importantly, they're not necessarily targeting the right customer." "Most companies will optimise for short-term benefit as opposed to long-term gain - what's actually happening is in the long term they're shooting themselves in the foot." "Nine in ten experiments are failing, it's taking too long to run experiments, teams don't know which experiments to run." "If we're going to move towards a world where it's multimodal interfaces with lots of different UI, knowing what to generate for who and when is going to become significantly more difficult." You can watch the full conversation by searching "Tom Charman, Co-founder @ blok: How We're Predicting User Behaviour Before Writing Code"

57

1 Comment

Does building the application layer feel like playing Jenga with LLM releases? A couple of new model capabilities and the value proposition of your product significantly diminishes? Over enough LLM generations, does the application layer become redundant? Or is there a pattern to which companies survive and thrive? I’m starting to see a few categories of application-layer/agentic startups that may remain more resilient: Multi-persona or cross-function workflows Products that solve for workflows across multiple stakeholders with staggered underlying data require companies to solve for integration and adoption barriers Context-heavy workflows Products that require capturing proprietary user context unavailable in a digital data format. The real innovation here lies in capturing this context without breaking UX. Consultative GTM models Startups solving for ICPs or workflows where a solutioning/consultative sales motion is needed. The org structure in such cases evolves very differently than a model company. Old-school network effects + velocity Firms that build compounding value loops, where each new user or data point strengthens the product are far less vulnerable to model advances LLMs will keep eating into the application layer, but the survivors won’t just be wrappers. They’ll be the ones building workflows, context, GTM motions, or networks that LLMs can’t easily replace. Which other patterns of resilience are you seeing?

38

1 Comment