LLM Investors
CapLink currently tracks 2 verified investors focused on LLM — a small but growing slice of the global funding landscape.
The mix is led by PE/Buy-Out and VC. Deal coverage spans Seed through Series B, with the largest concentration at Seed.
Investor headquarters cluster in Canada, Australia, Belgium, France and Germany, with activity across 14 countries in total. Ticket sizes range from roughly $60K to $240M, covering early angel cheques through to growth-stage rounds.
Use the pre-filtered database below to explore every LLM investor on CapLink, or sign up to unlock contact details, ticket sizes and detailed investment criteria.
LLM investor database
2 investors matched for LLM. Sign up to unlock contact details and full profiles.
| Investor |
|---|
Bpifrance Investissement SAS, Bpifrance-Large cap Bpifrance Investissement SAS is a private equity and venture capital firm specializing in fund of funds and direct investments. For direct investments, the firm specializes in seed, series A, series B, startup, growth capital, emerging growth, early to late venture, middle market to mature stage, industry consolidation and buyout transactions in small and medium-sized companies and mid-caps. For fund of fund investments it seeks to invest in private equity funds, venture capital funds, growth capital funds, mezzanine funds, and turnaround funds, national funds, regional funds of France or in international funds with a link with France. It considers investments in the following countries: France, Germany, Switzerland, Turkey, Morocco, Australia, Japan, South Korea, New Zealand, Mexico, Canada, and the United States of America. For fund of fund investments, the firm can invest in North America, Europe including France and Italy, in North Africa including Algeria, Egypt, Morocco and Tunisia, and in Asia-pacific. It can consider investments in all sectors, some funds prefer to invest in fintech, life sciences, technology (information and communication), industrials, services, energy including nuclear and renewable energy, energetic transition, healthcare, biotechnology, robotics, big data, cloud computing, green chemistry, clothing (accessories, leather goods, shoes, home textiles), tourism, Cinema and audiovisual, video games, music and live, publishing, fashion and luxury, beauty, art and design, gastronomy and leisure sectors including Hotels and all forms of accommodation management (traditional hotels, thalassotherapy or hydrotherapy, campsites, hotel residences, reception structures, tourist residences, etc.); Catering (restaurant chains, take-out, collective catering, fast food, etc.) Leisure and free time activities (sports or leisure facilities, leisure parks, winter sports resorts, performance halls, media in the leisure part); Travel agencies/tourist transport (tour operators, excursionists, river, air transport, car rental), software, media, perfumes, cosmetics, watch making, jewellery, food and financials including private equity and real estate sectors. Generally, for direct investments, the firm can invest between $0.064 million and $240 million in firms reporting revenues between $0.56 million and $1690 million with a maximum enterprise value of $100.31 and minimum EBITDA of $4.43 million. For fund of fund investments, the firm typically invests between $1.11 million and $67.16 million. Within growth capital it seeks to invest greater than €1.5 million ($2.21 million). The firm typically invests between €0.4 million ($0.72 million) and €10 million ($13.21 million) in growth capital and leveraged buyout transactions, in companies headquartered in France with revenues starting at €5 million ($6.61 million). Its venture capital investments range between €1.5 million ($2.11 million) and €10 million ($13.21 million), which it invests in small and medium-sized companies. For fund investments, the firm invests between €3 million ($3.83 million) and €60 million ($76.66 million). It usually takes a minority stake in the companies and the funds in which it invests. Bpifrance Investissement SAS was founded in 1988 and is based in Paris, France, with additional offices located all across France. It operates as a subsidiary of Bpifrance Participations SA., The Large Cap team invests and monitors investments in companies with a strong French presence, whether listed or unlisted ISEs or large caps, with the aim of strengthening them, helping them grow, stabilising their capital or increasing their foothold in France. The large cap team makes equity or quasi-equity investments for amounts of $15 million or more, mainly through the 2020 long-term ISE fund (99 years). The large cap team makes equity or quasi-equity investments for amounts of $15 million or more, mainly through the 2020 long-term ISE fund (99 years). |
 Radical Ventures Radical Ventures is a Toronto-based venture capital firm dedicated to investing in early-stage startups that leverage artificial intelligence (AI) and machine learning (ML) to transform industries. Founded in 2017 by Daniel Farcet and Ajay Agarwal, the firm focuses on North American companies and aims to partner with exceptional entrepreneurs to build enduring AI-driven businesses.
In August 2024, Radical Ventures raised nearly $800 million for its third institutional fund, marking the largest AI-focused fund to date.
This fund is dedicated to growth-stage startups and brings the firm's total assets under management to approximately $1.8 billion.
The firm's portfolio includes notable AI companies such as Cohere, a large language model developer valued at $5.5 billion, and Waabi, an autonomous driving company. |
Understanding LLM investors
What are LLM investors, and what do they look for?
Companies building on language models are judged on what survives the next model release, and investors ask the question directly. Capability that a new model version provides natively erases the value of work done to compensate for the previous version's limitations. Investors want to know which parts of your product get better when models improve and which parts become unnecessary, because founders who have not distinguished the two are exposed to a release schedule they do not control. Serving economics form the second axis. Running language models in production costs money per request, and that cost varies enormously with model choice, context length, caching strategy and whether you serve your own weights or call an interface. Investors examine cost per useful output and how it has moved as the company optimised, because gross margin in this category is an engineering outcome rather than a pricing decision. Third, they assess evaluation discipline. Products built on language models fail in ways that are difficult to detect without systematic measurement, and a company without an evaluation suite cannot know when a model update degraded its own behaviour.
Why LLM is attracting investor interest
Serving economics became the commercial question once building a working prototype stopped being difficult. When any competent team can assemble a demonstration in days, the differentiation moves to whether the thing can be operated profitably at volume, and that turns on caching, routing between models of different cost, context management and knowing when a smaller model suffices. Open weight models changed the calculation substantially. Running models on infrastructure you control removes per-request interface costs, addresses data residency requirements that European buyers raise constantly, and gives predictable performance that does not change when a provider updates something. The trade-off is operational burden, and the choice between the two has become a central architectural decision. European enterprise demand is shaped by that residency question more than by capability. Regulated buyers frequently cannot send data to an interface outside the jurisdiction, which advantages companies designed to run within the customer's own environment. Investors remain alert to how quickly this layer commoditises, since techniques spread fast and the interesting question is what accumulates rather than what works today.
Which funding stages LLM investors are active at
Funding here moves quickly at the early stages and gets harder afterwards. Seed rounds are raised readily on a working product and a credible team, since the category attracts capital and building something functional has become fast. Investors accept limited revenue at this point. Series A is where the durability question bites. Investors want evidence that customers pay for the output rather than the novelty, that retention holds past initial enthusiasm, and that the product would survive a model release providing similar capability natively. Companies whose value was compensating for model limitations struggle here. Series B and later concentrate on margin and accumulation. Investors examine gross margin under real usage, whether inference cost per unit of value is falling, and whether the company has built something that compounds, such as evaluation data, workflow integration or proprietary context that improves output. Later-stage capital treats these as software businesses and applies software benchmarks, which surprises founders who raised earlier rounds on technical narrative.
Types of investors active in LLM
Investors focused on products rather than models, who ask which capability survives the next release and examine evaluation discipline. They are the sharpest audience on serving economics and the most useful on enterprise deployment.
Generalist B2B funds applying conventional metrics, attentive to gross margin under load and to whether retention holds past initial enthusiasm. They benchmark against software rather than against other AI companies.
Backers of the serving, routing, caching and evaluation layers that make production deployment economic. They read adoption among engineering teams and understand where operational cost actually accumulates.
Corporate investors from financial services, healthcare and the public sector who need models running inside their own environment. Their residency requirements are a commercial advantage for companies designed around them.
Working machine learning engineers investing collectively. Their technical validation moves faster than revenue in a category where credibility affects who takes the next meeting, and their judgement on architectural choices is practical.
Ready to reach LLM investors?
Create a free CapLink account to unlock full investor profiles, contact details, ticket sizes and intelligent matching.