Participated · Seed · Jun 2026

Pramaana Labs combines conventional large language models with a deterministic verification layer that uses tools of formal verification inspired by the open-source LEAN proof assistant. The company develops LEAN-style, domain-specific formalizations overseen by domain experts to codify rules into executable code. Pramaana is targeting high-stakes verticals—such as tax, law, drug discovery, and cybersecurity—where errors can have serious consequences. Its approach keeps the flexibility of LLMs for natural language reasoning while adding a verification step to ensure outputs check out. The company is collaborating with advisors and academics including former IRS commissioner Danny Werfel and professors from IIT Delhi, IIT Madras, and UC Berkeley. The product roadmap centers on building formal verification systems tailored to each use case and deploying them in regulated or safety-critical settings.

Participated · Series D · Jan 2026

Alpaca operates a brokerage infrastructure API that powers fintechs, banks, broker-dealers, wealth managers, algorithmic trading firms, active traders, and crypto-native financial platforms. The company supports over 10 million brokerage accounts across hundreds of fintechs and institutions in more than 40 countries. Alpaca has expanded its regulated footprint through acquisitions, including an IFSCA-regulated broker-dealer and payment service provider in GIFT City and UK/European entities with passporting across the EEA, and has launched global equities access beginning with European equities. It has onboarded major global crypto exchanges and tokenization platforms, reported doubling revenue year-over-year for three consecutive years, and surpassed $1.5 billion in assets under custody for underlying stocks backing tokenized equities. Alpaca plans to use the new financing to accelerate its agent-first brokerage and API-first prime brokerage infrastructure to support tokenized markets and AI-native financial services.

Participated · Seed · Oct 2024

OpenOrigins builds a provenance-focused blockchain, anchoring solutions, and real-time camera-capture technologies to verify and protect the authenticity of images and videos. The company emphasizes proving content at source rather than retroactive AI detection, offering a decentralized immutable record for media provenance. OpenOrigins has launched a content marketplace with more than 1PB of ready-to-license data to enable media and data companies to responsibly monetize and license archives for AI model training. It has signed partnerships with major media organizations, including ITN, to secure over a million video and audio assets spanning 70 years. The company also pilots point-of-capture technology with newswires and has run global election coverage pilots to validate newly captured media. OpenOrigins began as an academic project at the University of Cambridge while founder Dr. Manny Ahmed completed his PhD, and it is expanding globally from its London base.

Led · Series B · Oct 2024

Bioniq builds AI-driven personalised supplements that are formulated from users' blood test results. The company, launched in 2019 in London, ships globally across five continents and offers products in 70 countries. It has created unique formulas for over 300,000 users and says its total user base will grow more than tenfold by the end of 2024. Bioniq has launched a hyper-personalised product called Build Your Own to allow more precise, goal-based nutrient routines. The company partners with health and sports institutions such as Lanserhof and the UFC, and counts elite athletes — including Soccer World Cup champions and NBA All‑Stars — among its users. Leadership emphasizes measurable before-and-after nutrient snapshots and proven optimisation as core assets for growth and credibility. Bioniq creates 100% customized daily supplements using advanced blood biomarker analysis combined with self-reported health data and a patent-pending AI algorithm. Each user receives a unique formula of up to 120 ingredients delivered in patented prebiotic guar granules designed for improved bioavailability and micro-dosing. Products are Swiss-made, pharmaceutical-grade, and the company offers daily nutrient support, 1:1 nutrition consultations, and educational resources. Bioniq reports sales in 70+ countries and a user base that grew more than fivefold this year, with the U.S. now contributing over 60% of global sales. The company highlights partnerships with leading health and sports institutions such as Lanserhof and the UFC and is used by elite athletes across sports. Management says it will accelerate global expansion with particular focus on the U.S., Europe, and the Middle East, and Cristiano Ronaldo will participate in future brand and product innovation. Bioniq builds AI-driven, patent-pending algorithms that generate personalized supplement formulas from blood biomarker data and individual inputs such as height, weight, age, lifestyle and health goals. The company offers Bioniq PRO, which integrates blood test data (including partner lab testing and 1:1 nutrition consultations), and Bioniq GO, a questionnaire-based, more accessible option. All supplements are Swiss-made, pharmaceutical-grade, micro-dosed and delivered in patented prebiotic guar granules to maximize bioavailability. Bioniq has compiled the largest proprietary biochemical database of its kind with blood test data from customers across five continents and has created unique formulas for over 100,000 users. The U.S. now represents over 50% of its customer base, with some regions growing more than 500% in the last two quarters, and the company partners with institutions such as Lanserhof and the UFC and serves elite athletes. It plans to use new funding to expand global laboratory integrations, accelerate product development, grow market, operational and HR presence in the U.S. and Middle East, and implement a corporate platform for medical, wellness, and athletic institutions. bioniq develops hyper-personalized supplement products, most recently launching bioniq GO, a subscription product priced at £49/month. GO leverages 3 million proprietary biochemical data points drawn from more than 50,000 blood tests across over 30,000 clients and 25 medical publications. The company says its GO formulas can include up to 34 components with as many as 10 million possible combinations and claims personalization 200,000 times higher than similar products. Founders include CEO Vadim Fedotov (a former German national team basketball player) and Dr. Konstantin Karuzin, a neurosurgeon who previously developed performance-improvement systems for professional athletes. Earlier this year bioniq acquired LOEWI, a scientific spin-off of TU München that provides tailor-made nutrients based on blood work and lifestyle. The company positions itself around data aggregation and digital health monitoring to improve personalization over time as it ingests new blood samples.

Participated · Equity · Sep 2022

Deep Instinct is the first company to apply deep learning to cybersecurity, offering a natively architected, prevention-first threat prevention platform. The company says its platform stops >99% of threats before other solutions see them, reducing detection noise and false alert storms. Deep Instinct positions its product to improve SOC team productivity and lower total cost of ownership for customers. The firm intends to use investor support to accelerate growth and drive adoption of a prevention-first model across enterprises. It currently protects customers across North America, Europe, and APAC; enterprise customers tripled in 2022 and it has strategic wins with numerous Global 2000 companies. The article does not disclose revenue or other financial metrics. Deep Instinct offers a prevention-first cybersecurity platform that applies deep learning to stop ransomware and other malware across hybrid environments. Its Deep Instinct Prevention Platform is designed to add multilayered protection to existing security stacks. The company says it will use the new funding to further penetrate the market. Deep Instinct raised $62.5M in the reported financing. As part of the announcement, Lane Bess—who has more than 35 years of experience in cybersecurity and was formerly CEO of Palo Alto Networks and COO of Zscaler—took over as CEO effective immediately. Co-founder Guy Caspi moved from CEO to Board Chair and became Chief Product Officer. Deep Instinct offers an end-to-end deep learning framework and platform designed to predict, prevent, and analyze cyberattacks across endpoints and networks. Its platform is positioned to protect organizations at multiple touchpoints and has expanded enterprise adoption rapidly. The company reports protecting customers across North America, Europe, and APAC, with enterprise customers tripling in the prior year and strategic Global 2000 wins in Q1 2021. Led by CEO Guy Caspi and Chairman Lane Bess, Deep Instinct intends to use the new funds to accelerate its growth plans for 2021 and beyond. The company has raised a total of $200M to date following this round. Deep Instinct applies deep neural network algorithms to cybersecurity, using end-to-end deep learning to identify and stop known and unknown malware, including zero-days and advanced persistent threats. Its system ingests raw data from servers, mobile devices and endpoints and is largely autonomous, with automated processes handled by its "deep learning brain." The company sells primarily to enterprise customers, both directly and via partners such as HP, which bundle or resell its solutions. Deep Instinct emphasizes cross-OS protection and claims higher detection and lower false positives than traditional machine-learning approaches that rely on human feature extraction. Management says the product can prevent first-seen, unknown attacks and is already used to counter rising spyware, ransomware and file-less attacks. The longer-term plan is to build a consumer-facing version of the product. The company is profitable, and highlights strategic relationships with large tech partners. Deep Instinct is a Palo Alto, Calif.-based company that applies deep learning to cybersecurity. Led by CEO and co-founder Guy Caspi, the company offers a deep learning platform that detects malicious behavior across multiple vectors and provides adaptive defenses for endpoints, servers, and mobile devices. The company raised $32M in a Series B financing led by CNTP, with participation from strategic investors including Nvidia, Coatue Management, and existing investors. It plans to use the funds to accelerate deployments in the U.S. and global markets. Mohsen Moazami, founding MD of CNTP and a former Cisco executive, was added to Deep Instinct’s board of directors in connection with the round.