What happened
Ondine Capital secured commitments from institutional and strategic limited partners, including several Taiwanese semiconductor firms, for the first close of its third USD‑denominated venture fund (Fund III). The fund will target investments across the AI stack—from foundation models and next‑generation chips to AI‑native software and physical AI applications. Early portfolio highlights include Recursive Superintelligence (a recursive‑AI startup that raised $650 million at a $4 billion pre‑money valuation), Bot Auto (an autonomous freight company expanding driverless routes in Texas), and Salience Labs (a silicon‑photonics firm that closed a $125 million round). Ondine also appointed Brian Yen as a new partner to strengthen its global sourcing network.
Ondine Capital announced that its third USD‑denominated venture fund has achieved its first close, securing commitments from a mix of top‑tier institutional investors and strategic limited partners, notably several leading semiconductor enterprises based in Taiwan. The exact size of the fund was not disclosed in the source material.
Fund III is positioned to invest across the entire AI stack. The firm explicitly cites AI foundation models, next‑generation semiconductors, AI infrastructure, AI‑native software, and physical AI applications as target areas. This broad mandate reflects Ondine’s stated expertise in deep‑technology and semiconductor markets.
The announcement highlighted three early‑stage portfolio companies. Recursive Superintelligence, founded by former Salesforce chief scientist Richard Socher and staffed by veterans from OpenAI, Google and Meta, recently raised $650 million at a $4 billion pre‑money valuation, with lead investors including Google Ventures, Nvidia and AMD. Bot Auto, a Houston‑based physical AI startup founded by former TuSimple CTO Xiaodi Hou, has partnered with Ryan Transportation to launch driverless freight services between Houston and Dallas. Salience Labs, a silicon‑photonics firm co‑founded by Vaysh Kewada and Johannes Feldmann, closed a $125 million financing round led by Bullhound Capital, with strategic backing from AMD Ventures and Intel Capital.
Coinciding with the fund close, Ondine appointed Brian Yen as a new partner. Yen previously led the Taipei office of Headline Asia and has a track record of investing in consumer technologies, agentic services and physical AI. In his remarks, Yen emphasized the “monumental value creation” expected from the convergence of AI and semiconductors.
Why it matters
The fund’s focus on the full AI value chain signals a growing convergence between artificial‑intelligence research and semiconductor manufacturing, a trend that could accelerate the deployment of more capable and energy‑efficient AI systems. By backing both software‑centric AI startups and hardware‑focused companies, Ondine aims to reduce the latency and cost barriers that have limited large‑scale model training and . The involvement of major semiconductor players as limited partners suggests industry‑wide confidence that AI‑driven demand will shape next‑generation chip design, potentially influencing supply‑chain dynamics and capital allocation in the broader tech sector. Moreover, the sizable financing secured by portfolio companies—particularly Recursive Superintelligence’s $650 million raise—highlights the scale of capital now flowing into advanced AI research, raising questions about competitive advantage, talent concentration, and regulatory oversight.
The convergence of AI and semiconductor technology is a critical driver of future computing performance. By allocating capital to both AI models and the hardware that runs them, Ondine’s Fund III could help bridge the gap between algorithmic advances and practical, scalable deployment, potentially lowering the cost of training large models and enabling new applications in edge computing and autonomous systems.
Large financing rounds for AI‑focused startups, such as Recursive Superintelligence’s $650 million raise, illustrate the appetite of venture capital and strategic investors for high‑risk, high‑reward AI research. This influx of capital may accelerate breakthroughs but also raises concerns about concentration of talent and resources in a few well‑funded entities, which could impact competition and regulatory oversight.
The involvement of Taiwanese semiconductor firms as limited partners suggests a strategic alignment between AI developers and chip manufacturers. Taiwan’s dominant position in advanced node manufacturing could give portfolio companies preferential access to cutting‑edge process technologies, influencing global AI hardware supply chains and potentially reshaping geopolitical dynamics around AI capability.
Brian Yen’s appointment signals an intent to deepen Ondine’s presence in the Asia‑Pacific region, leveraging his network across Japan, Taiwan and Southeast Asia. This could lead to more cross‑border collaborations, joint R&D projects, and a pipeline of early‑stage AI founders from markets that have historically been under‑represented in global VC flows.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Which component of an AI application is the machine-learning model itself?
What to watch next
Investors and industry observers should monitor how Ondine’s Fund III allocations influence the development of AI‑optimized hardware, especially any collaborations that result in new chip architectures or photonic interconnects. The performance and safety of Recursive Superintelligence’s recursive‑self‑improving models will also attract scrutiny from experts. Additionally, the fund’s partnership network in Taiwan may affect regional semiconductor capacity, potentially reshaping global supply chains for AI workloads.
The specific allocation of Fund III capital to hardware‑focused versus software‑focused AI startups will indicate where investors see the greatest near‑term ROI. Tracking the proportion of investments in chip design, photonics, and AI‑optimized silicon versus model‑centric companies will reveal emerging priorities.
Recursive Superintelligence’s roadmap for recursive self‑improving models will be closely watched by researchers. Any public demonstrations or results could trigger discussions about alignment, controllability, and regulatory frameworks.
The partnership dynamics between Salience Labs and major foundries like Tower Semiconductor may produce new silicon‑photonics products tailored for AI data centers. Early product announcements or silicon samples would be a tangible sign of progress.
Regulatory developments in Taiwan and the broader Asia‑Pacific region concerning AI‑enabled hardware, export controls, and data sovereignty could affect the fund’s strategic direction and the ability of portfolio companies to scale internationally.