How Termina selected and ranked the 2025 Seed 100 and Seed 40 lists
The Seed 100 and Seed 40 lists are now in their fifth year in our partnership with Business Insider to objectively identify and celebrate the success of some of the world's greatest seed investors. In the run of these releases, AI has emerged as a defining characteristic of seed-stage startups today, the results of which will likely drive the lists to come.
To produce the 2025 rankings, we statistically analyzed seed investor performance in 25 areas using Crunchbase and PitchBook data. Since one of our goals is to analyze the prospective success of investors, rather than focus solely on past achievements, we consider only active investors with a minimum of five investments between 2010 and 2025. Our lists include solo venture capitalists and angel investors worldwide, assessed based on their investments in US companies.
The investors that made the cut for the 2025 lists were those with the strongest long-term indicators: exits (initial public offerings or acquisitions). For this year's ranking, investors whose stronger indicators were follow-on fundraising activity have had their positions affected by the recent pullback in venture capital fundraising.
To be named to the list, seed investors must have:
Investments that led to successful exits, including IPOs or acquisitions (exits that were meaningfully above "liquidation preference" or demonstrated increased company value rather than simply raising capital).
Investments that display intermediate signs of future success, with seed investments consistently receiving material sums of follow-on investment.
Active moderate-to-high seed investing over the previous two years.
Exits statistically have the most influence on differentiating investors. We also updated our methodology this year to weight more recent investments for the intermediate milestones, such as follow-on fundraising.
This year, the candidate pool that met the above criteria was 1,974 investors, a 5% increase from last year's. Women accounted for 11% of all seed investors in this year's pool — up from 8% of the first Seed 100's pool in 2021. This final list included 26 investors that increased their rank, 27 with the same or lower rank, and 47 new investors. Twenty-seven investors in this year's ranking were also featured in the original 2021 list.
The generative AI wave of seed investing
It's been two years since OpenAI released its revolutionary large language model GPT-4, which ignited the world's imagination about the future of artificial intelligence. Having AI at the center of large outcomes is not new; many companies, including Alphabet and Meta, have predicated their business models on it and were successful long before ChatGPT.
You can segment AI into discriminative and generative models. The key difference lies in their objectives. Discriminative models focus on predicting labels and values, or making decisions based on input data, while generative models aim to capture the underlying data distribution and generate samples from that distribution.
The first wave of seed investing in AI was from 2010 to 2018, during which companies included discriminative machine learning in their general business models. For example, companies like Uber, Instacart, Airbnb, and DoorDash were built on discriminative matching and recommendation algorithms.
The new wave of AI was kicked off by generative models, which include large language models like those that power OpenAI's ChatGPT, Alphabet's Gemini, and Anthropic's Claude. In a mere two years, the share of companies where AI materially matters to their products has nearly doubled from 13% to 24%. How this evolves will likely greatly influence how the Seed 100 rankings look over the coming years. Since seed investments usually take years to mature, we're still a few years away from understanding how the new wave of AI investments will perform.
In the chart below, we show five seed investment categories to contextualize the growth of AI investment. Note that this is not an exhaustive view of seed investment sectors. In this view, while there has been steady expansion of AI investing at the seed stage, the jump last year stands out.
Termina
Jake Ellowitz is the chief technology officer and cofounder of Termina, which works with venture capital and private equity firms, sovereign wealth funds, and executives to assess investment opportunities.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Forbes
37 minutes ago
- Forbes
Your Mind Can Create Using The Same Process As AI
Woman sketching a business plan on a placard at a creative office Four short months ago, when China launched DeepSeek, its new AI chatbot, the effect on the U.S. stock market was profound. Hardest hit was NVIDIA Corporation, the leading supplier of hardware and software for DeepSeek's entrenched American competitors— OpenAI's ChatGPT, Meta's Llama, and Anthropic's Claude. NVIDIA lost almost $600 billion in market capitalization and its share price plummeted 17%. As this Forbes article reports, DeepSeek also started a price war among its Chinese AI competitors: ByteDance, Tencent, Baidu and Alibaba. What shook up all these constituencies was DeepSeek's revolutionary AI architecture with a secret sauce called Mixture-of-Experts (MoE) a vast simplification from the single giant neural network called Large Language Model (LLM) used by ChatGPT, Llama, and Claude. Both MoE and LLM provide users with the ability to manage access to a huge corpus of digital information. But your mind is capable of doing the same with a huge corpus of human information. With or without AI, whenever you set about to generate, organize, and create information—whether as text or spoken—you must access a huge corpus of information in your brain. So in the spirt of National Creativity Day, this blog will provide you with a very simple four-step process to do that—and to become more effective when you create. Fundamental to all the steps is an understanding of the basic concepts in Nobel Prize winner Daniel Kahneman's bestselling book Thinking, Fast and Slow in which he explains that the human mind functions in two distinct phases. If you start to develop your presentation, report, or email with Slow Thinking factors such as logic, sequence, and word choice (or, for that matter, the color, style, font, and design of your slides) while your Fast Thinking is bubbling up all those random ideas, the whole process descends into disarray. The key then is to allow the Fast Thinking to run its course before attempting to impose a sequence with Slow Thinking. So, control your Fast Thinking with these four steps: Now and only now are you ready to organize and sequence your content. Who needs DeepSeek?
Entrepreneur
an hour ago
- Entrepreneur
Can AI Automate the Job of an Entrepreneur?
Find out how much of your day AI can automate with this bundle. Disclosure: Our goal is to feature products and services that we think you'll find interesting and useful. If you purchase them, Entrepreneur may get a small share of the revenue from the sale from our commerce partners. Staying competitive in today's digital economy means keeping up with rapidly evolving tools and technologies. For professionals and aspiring creators alike, understanding how to use artificial intelligence (AI) in practical, efficient ways can make a real difference. The ChatGPT and Automation E-Degree is a flexible, self-paced learning program designed to help you build those skills from the ground up, and it's on sale for $19.97 (reg. $790). Automate the work you don't want to do This online course bundle includes 12 individual courses with more than 25 hours of video content. Whether you are interested in marketing, automation, business optimization or creative workflows, the lessons are built to show how AI can fit into your professional routine. You'll explore over 20 of today's most widely used AI tools, including ChatGPT, and learn how to apply them in real-world scenarios. The curriculum covers a range of topics, from customizing ChatGPT for different industries to automating business processes and improving communication through AI-driven interactions. You'll also learn how to use data visualization tools to create more compelling presentations and reports. The overall goal is to help you use AI to save time, streamline your workflow, and make more informed decisions. This program is geared toward beginners, so no prior technical experience is required. All you need is a device with internet access. Once enrolled, you'll get lifetime access to the course material, so you can work through the content at your own pace and revisit lessons whenever needed. While software tools are not included, the focus is on how to use them effectively, with hands-on instruction and examples you can apply in your own work. It's only $19.97 to get the ChatGPT Automation and E-Degree, but it won't stay that way. Sale ends soon. ChatGPT & Automation E-Degree See Deal StackSocial prices subject to change
Business Insider
an hour ago
- Business Insider
This YC-backed startup wants to speed up US visa applications with AI. Read the pitch deck it used to raise $2.7 million.
Gale, a startup that uses AI to streamline the US visa application process for employers and workers, has raised $2.7 million in seed funding. Its cofounders, who are all immigrants, launched the startup out of Y Combinator's 2025 winter cohort following their own frustrations navigating the US visa system. The trio — Rahul Gudise, Rishabh Sambare, and Haokun Qin — told Business Insider they found the process confusing, full of delays, and bureaucratic. "If I wanted a question answered about my visa status, it would take the attorneys like four business days to get back to me, minimum," said Gudise, who previously worked at Tesla and Nvidia. He added that the traditional approach of using law firms means cases can drag on for weeks unless clients are willing to pay premium fees. "Simple questions can take multiple steps to resolve because they have to be relayed up and down a chain of staff and attorneys," Gudise said. Gale's web-based platform automates much of the administrative work involved in visa applications. Instead of manually filling out lengthy forms, users can upload their résumés and passports for processing by the Gale platform. A licensed immigration lawyer conducts a final review before submission. Right now, the startup is focused on H-1B visas. Gudise said the goal is to make onboarding "as fast as possible" for applicants, and to free up time for lawyers by automating repetitive tasks. "Lawyers don't have to spend time on busywork or rely on paralegals for document prep," Gudise said. The team aims to answer client questions within hours, rather than days. Compliance is another key focus for Gale. Gudise said most law firms hand off the process once the visa is approved, leaving employers to handle ongoing requirements on their own. Gale is building out its platform to monitor changes like job titles or promotions that might impact a worker's visa status. The team is also developing integrations with company HR systems to help employers stay compliant as requirements change. Gale plans to use the new funding to boost its compliance tools and grow the team. Building new partnerships with employers and immigration lawyers is also a key focus as the company scales, Qin told BI. "At the end of the day, we want to promote legal immigration to the States and help employers stay compliant. This is the way to do it sustainably," Gudise added. The startup's seed round was led by Axiom Partners, with participation from Pioneer Fund, 468 Capital, Elevation Capital, and Y Combinator. Here's an exclusive look at the pitch deck Gale used to raise $2.7 million.
