At a time when the wave of artificial intelligence is sweeping the world, the Thiel Fellowship has successfully bet on multiple unicorn projects such as Figma and Scale AI with its keen insight, becoming a bellwether in the field of technology entrepreneurship. In 2025, the Thiel Fellowship will continue to focus on the field of AI, identifying a group of young entrepreneurs with great potential. These projects not only cover popular directions such as AI infrastructure, digital humans, and human-computer interaction, but also showcase cutting-edge explorations ranging from biological system programmability to extreme climate prediction interventions.
From Airbnb to Figma, from Ethereum to Scale AI, the Thiel Fellowship has been one of the “secret starting points” for “little genius” entrepreneurs.
Founded in 2011 by PayPal founder Peter Thiel, the Thiel Fellowship provides each recipient with an annual $100,000 prize and “freedom to drop out.”
After AI became the main narrative, the Thiel Fellowship began to focus on the paradigm shift brought about by AI. From Walden Yan and Scott Wu, founders of Cognition AI (the company behind Devin), to Scale AI co-founder Lucy Guo, to Mercor founders Adarsh Hiremath, Brendan Foody, and Surya Midha, more and more young founders from the Thiel network have become structural variables in the AI era.
Thiel Fellow in 2025, continuing this trend. This year’s inductees present several notable characteristics:
- Geographically, it has expanded from traditional universities in the Bay Area and Northeast China to a broader global perspective, including emerging technology ecosystems such as Germany, Israel, and Latin America.
- In terms of project direction, AI dominates, especially AI infrastructure, Digital Human, and Human-Computer Interaction are high-frequency themes;
- In the financing stage, many projects have completed seed rounds led by well-known funds such as a16z, Sequoia, Valar, and SV Angel before the announcement of the Fellowship, showing maturity and market-oriented capabilities beyond their age.
In this Cohort, we saw several clear types of “Moonshot Project” paradigms: overcoming the basic modeling capabilities of anthropomorphic AI and trying to create “truly empathetic digital humans”; Challenging the programmability of biological systems in an attempt to unlock the engineering route of organ regeneration; Others have gone into the eye of the storm and used AI to reconstruct the prediction-intervention system for extreme weather.
How can product managers do a good job in B-end digitalization?
All walks of life have taken advantage of the ride-hail of digital transformation and achieved the rapid development of the industry. Since B-end products are products that provide services for enterprises, how should enterprises ride the digital ride?
View details >
This article will sort out and analyze the most representative entrepreneurial projects in the 2025 Thiel Fellow, trying to answer several key questions:
- What kind of “future picture” are this “little geniuses” betting on?
- Under the AI trend, can their Moonshot project be the next unicorn?
The answer is not clear at the moment, but Peter Thiel once said: “We want to support ideas that seem crazy at first and can eventually change the world. ”
01. Group portrait of Thiel Fellow
This Thiel Fellow in 2025 is like a group of builders who “do infrastructure with models and code”. They abstract the underlying propositions of the AI era into concrete paths, from sensory simulation and human-computer interfaces to financial and biological system reconstruction, trying to solve long-term problems with engineering methods.
Three popular directions: AI Infra, Financial Infra, and Programmable Life Systems
The three most representative of this year’s Thiel FellowThe topics are:
1）AI Infra
Companies including Canopy Labs, Intempus, and others revolve around digital human construction, low-latency reasoning, multimodal modeling, etc., redefining interactive interfaces with infra logic. Most of these projects emphasize real-time, openness, and embedded experience, and are aimed at vertical scenarios such as education, psychotherapy, and content creation.
2) New generation financial infrastructure
Represented by Ivy, it tries to become the global standard layer for A2A (account-to-account) payments, breaking through the financial fault line between countries in a technology-neutral way. Unlike traditional payment entrepreneurship, they emphasize “regulatory adaptability” and “asset-light compliance”, and the investors behind them include fintech heavy funds such as Valar and Creandum.
3）Health Tech & Biocomputation
For example, Phase Labs and Orbit focus on biological system modeling, neuromodulation equipment, renewable human organs and other fields, starting from system modeling and bioelectrical signals, and propose more imaginative paths than stem cells and CRISPR. These companies usually have longer R&D cycles, but if they break through, they have the opportunity to reconstruct the first principles of medicine.
In addition, Digital Human and Human-Computer Interaction also show a signal of “moving from toys to platforms”, and more and more entrepreneurs are beginning to build full-stack capabilities (voice + expression + spatial behavior) of “AI + Embodiment”.
Founder Portrait: Young, cross-border, and highly experimental
Almost all founders are concentrated between the ages of 17-21 and generally have the following characteristics:
- Strong interdisciplinary background: Biology + computing, neuroscience + engineering, art + programming combinations frequently appear, and they are “boundary blurred talents”. For example, Orbit’s team is composed of neuroscience self-educators + aviation hardware geeks + Harvard biophysical consultants, which is a closed loop of “vision + technology + validation”.
- Early Project Experience: Many have gained real-world project experience in open source communities, internships, or even starting their own companies since the age of 14. The founder of Innerphases started making products at Chess.com at the age of 16 and is now involved in Neuralink’s BCI research.
- Extremely sensitive to long-term problems: Unlike the general sense of “doing AI applications” or “chasing the wind”, this group of people is generally solving problems that belong to the “structural layer”. They are not satisfied with optimizing one part of LLMs, but want to define “what AI should look like in the future.”
Is Thiel Fellowship still an incubator for high-executive unicorns?
The answer is still yes, but the portrait has changed.
The early Thiel Fellow was more like a “lone traveler”, relying on extreme ingenuity to penetrate the market in a very short period of time (such as Vitalik Buterin, founder of Ethereum, Dylan Field, founder of Figma); After the arrival of AI, Thiel Fellow is more like a “structure builder” with systematic thinking and resource integration capabilities. They are no longer just pursuing “fast”, but are building a technology platform that can sustain output, accepting longer horizon and deep engineering risks.
From the financing trajectory, it can also be seen that many projects have received support from leading funds such as a16z, Sequoia, and Initialized before the Fellowship was announced, which is not only a recognition of the founders, but also reflects the escalation of the synergistic relationship between Thiel Network and mainstream capital.
Three types of projects worth paying attention to
AI projects with infra thinking: such as Canopy Labs, which uses open voice models to cut into the base of digital humans, is not a tool product but an “infrastructure for future human interfaces”.
Deep tech startups with physical grippers: such as Orbit’s non-invasive neuromodulation hardware, Aeolus’ intelligent storm acquisition system, Neros’ tactical FPV drone. They all revolve around a core logic: reconstructing real-world hardware interfaces beyond the AI software narrative.
Biotech teams with life computing philosophies: such as Phase Labs, no longer use a single target to find medicine, but build systematic regeneration models; It embodies the change in direction of the new generation of biotech entrepreneurship: from “chemical experiments” to “systems modeling + biological control”.
02. Specific analysis of highlight projects
AI & Digital Human
Canopy Labs (Elias Fizesan)
Foundation Model Lab for Building Foundation Models for Digital Humans
positioning
Canopy Labs is committed to creating virtual humans that are indistinguishable from real people, becoming the underlying infrastructure provider of “human-as-a-service” to serve future AI-native interaction scenarios, including education, medical care, psychotherapy, etc. Founder Elias hopes to make virtual people appear on Zoom, podcasts, and educational platforms “with fakes”.
products
Canopy Labs’ first open-source model, Orpheus, focuses on human-level speech generation, and will be expanded to AI infra to support full-stack virtual human modeling. Its core values are reflected in the following points:
- Anthropomorphic speech performance: expressing empathy, emotional changes, zero-shot cloning, etc.
- Multi-modal real-time interaction: Control virtual humans to perform real actions such as walking, writing, and expressing.
- Extremely low-latency real-time inference: Latency as low as 200ms, supporting streaming inference in conversational applications; CNN-based non-streaming tokenizer, avoiding token popping, unique engineering implementation.
- Full-stack open source: More open and developer-friendly than Meta or Synthesia.
After expanding to the full stack of Digital Human in the future, Canopy Labs wants to simulate human body structure and behavior (walk, clap, look around), use 3D environment data to train and simulate spatial perception and gestures, and render MetaHumans-like rendering, but go further in terms of naturalness of movements.
Founder background
Elias Fizesan, a Math and Computer Science dropout at Columbia University, is a 2024 Thiel Fellow winner. He has participated in many open source projects and has strong technical hands-on skills and product storytelling capabilities.
Not a Serial entrepreneur, but showing strong vision articulation ability, obtaining top seed investment such as a16z speedrun, SV Angel, etc. The project is inspired by long-term observations of the humanized experience of AI interaction, believing that “long-term simulation of human perception systems should be done, rather than single-point model optimization”.
Financing history
Among them, SV Angel, BoxGroup, and Sunflower are extensions of the Thiel network that can support resources related to future AI infra construction.
Business model and customer base
Orpheus adopts an Infra-as-a-Service business model to provide B2B customers with anthropomorphic interaction modules for a long time as APIs to access various products. Its target customers cover verticals such as language learning platforms, virtual doctors, AI mental health assistants, and content creation tools, with a focus on products with a strong demand for high-fidelity interactive experiences. The initial go-to-market strategy focused on partnering with AI companies with existing LLM products to enhance user engagement and engagement by integrating Orpheus modules. It’s still in its early stages, with the product open sourced and positive feedback on GitHub, with several early partners already testing it internally.
Why it’s worth paying attention to
- Virtual humans are the core entrance to the long-term structural theme (AI+Embodiment), and every LLM that interacts with humans in the future may need a “humanoid shell”.
- Unlike companies like Synthesia, Heygen, etc., which are content generation-oriented, Canopy Labs takes the infra route, targeting open ecological niches.
- Orpheus technology starts from voice, but the vision behind it is to build a digital human platform with personality and empathy across modalities.
Potential risks
- Technical Challenges: Uncanny Valley is a long-standing challenge, with naturalness of movements and eye contact still to be overcome
- Computing power and data bottlenecks: High-quality multimodal training costs are high and rely on large-scale partner support
- Ethical Concerns: Deepfakes are confused with virtual human cognition and require policies and standards
Intempus (Teddy Warner)
“Reinvent the human-computer interaction experience and make robots ’emotional'”
positioning
Intempus aspires to be a game-changer in the global human-computer interaction field, solving the long-standing problem of stiff interaction between robots and humans, lack of emotional understanding, making it difficult for users to trust and collaborate efficiently, and allowing robots to truly integrate into human life and work scenarios.
products
At present, Intempus’ product form is mainly to retrofit existing robots, by adding emotional expression functions to robots, allowing them to express emotions through movement and help users better understand machine behavior. The core values are:
- Improve human-robot collaboration efficiency, such as reducing downtime due to unpredictable robot behavior in industrial environments.
- Use this interactive data to reverse optimize AI models to enhance their trainability and performance.
Founder background
Teddy Warner is a graduate of USC Iovine and Young Academy, where his unique blend of design, engineering, and entrepreneurship education has shaped his interdisciplinary intellectual background, giving him a deep understanding of the potential of technology convergence in different fields. In his early days, Teddy worked as a user research and content creator at Autodesk, a machinist at Fab Lab, and an R&D engineer at Rhythmlink.
Additionally, he is a member of the Midjourney hardware team, where he is involved in building AI models with spatial understanding. During his time at Midjourney, he observed that “most of the data used to train AI models comes from bots that lack spatial reasoning capabilities,” which became a key opportunity for him to start his business. He realized that it was crucial for robots to understand the world like humans, communicate with people naturally, and give them a human-like “physiological state”, and solving this pain point became the original intention of Intempus.
Business model and customer base
The company is still in its early stages, and its business model mainly adopts a B2B model, and provides emotional intelligence transformation solutions for the robots it sells to large enterprises by cooperating with robot manufacturers. The user group is mainly enterprises in the fields of industry, education, customer service, medical companion robots and other fields. Currently, Intempus has signed up for 7 enterprise-level robotics partners.
Why it’s worth paying attention to
- At a time when human-computer interaction (HRI) has become a core pain point in robot commercialization, Intempus focuses on empowering robots with emotional intelligence to accurately address key issues in the industry.
- Intempus has shown great application potential in the fields of education and medical companionship, and if technological breakthroughs can be achieved, it will significantly improve the practicality and user acceptance of robots in these scenarios.
- If Intempus can successfully standardize the emotional and motor expression of robots and form a “protocol” or “operating system” similar to the emotional language level of robots, it is very likely to become a category – defining player in the field of human-computer interaction emotional intelligence.
Potential risks
- The company is still in its very early stages, and its technology and business model are not fully proven.
- It is extremely difficult to standardize the expression of robots and solve the problem of cross-cultural adaptability, and there are differences in the understanding and expression of emotions in different cultures.
Biotech & Healthcare
Phase Labs(Juan Ebrath)
“Bioprogrammability research lab to unlock human organ/limb regeneration. ”
positioning
It has become a “biological system programming platform” in the field of biomedicine, solving medical problems such as organ failure and trauma repair, and promoting the vision of human immortality towards “whole body regeneration”.
products
Phase Lab integrates developmental biology, bioelectrical signaling, and machine learning to construct molecular network models of organ regeneration (e.g., vascular response mechanisms for kidney regeneration) to guide the regeneration process by regulating somatic bioelectrical patterns. Different from traditional technology paths such as stem cell research or organ printing, Phase Labs focuses on the interdisciplinary innovation of “bioelectricity + system modeling” and aims at the underlying logic of “programmable life”. Its core technologies mainly include two major labs:
- The work of the Dry lab is to construct geometric models of gene regulatory networks (GRNs) and protein interaction networks (PPIs) to analyze the molecular pathways of organogenesis and regeneration.
- Wet lab focuses on vascular responses, cell reprogramming trajectories, and bioelectrical signal markers in kidney regeneration.
Founder background
Juan graduated from Columbia University with a bachelor’s degree in computer science, where he applied machine learning to personalized treatment optimization in hospital-level scenarios. Since then, he has actively turned to the field of biology, with a particular focus on interdisciplinary disciplines such as developmental biology, regenerative medicine and bioelectrical signaling. Unlike the traditional PhD academic path, Juan first built his entrepreneurial capabilities and then entered the field by returning to wet lab research. Prior to founding Phase Labs, he founded Macondo, an anti-inflation product designed for Latin America that served more than 20,000 users in more than 20 countries. Macondo’s core innovation is a tradable real estate token, which was later acquired by Bricksave.
Currently, Juan is working in the Bioldes laboratory at Uniandes University, focusing on the modeling of organogenesis and organ regeneration, and has expressed a strong interest in somatic bioelectric patterns. His personal website (jpge.space) shows that he has in-depth and independent thinking on philosophical issues such as artificial life and “life programmability”.
Business model and customer base
At present, Phase labs has established a basic model of organ regeneration, and is verifying the key mechanism of kidney regeneration in wet lab, and plans to launch a prototype of trauma repair-related technology within 5 years.
Why it’s worth paying attention to
- Regenerative medicine is a potential trillion-level market, and bioelectrical signal modulation is a “next-generation technology” that has not yet been fully developed and has the potential to become an industry standard.
- Founder’s super compound ability: Juan is a Serial Entrepreneur with technical implementation capabilities and interdisciplinary scientific research vision, which can effectively connect laboratory breakthroughs and commercialization paths.
Orbit (Steven Pang & Colton El-Habr)
“Pioneer of Non-Invasive ‘Human Brain API’ to Unlock VR Immersion and Medical Applications with Vestibular Nerve Stimulation”
positioning
To become a leader in the field of non-invasive brain-computer interfaces (non-invasive BCI), break down the ethical and regulatory barriers of invasive technology, and unlock the long-term potential of humans in the fields of immersive interaction, mental health, and medical treatment with safe and convenient “neuromodulation tools”.
products
Orbit has developed a small, non-invasive behind-the-ear neuromodulator that can simulate motion perception (such as rotation, flight, etc.) by stimulating the vestibular system, and extend its application scenarios to medical fields such as sleep regulation and metabolic control. Its core values are as follows:
- Solve VR industry pain points: reduce motion sickness with vestibular feedback and enhance immersive experiences.
- Open up a new path for non-invasive brain-computer interfaces: no surgical operation is required, which greatly lowers the threshold for technology application, and at the same time provides the possibility for large-scale popularization with the advantage of low cost, but it needs to avoid ethical risks.
- Cross-field application potential: Expanding from consumer-grade VR to the medical and health fields (such as insomnia treatment, metabolic management), will replace some traditional drugs.
Founder background
- Steven Pang (CEO): Dropped out of Georgetown University majoring in computer science and taught himself neuroscience. He is good at building strategic visions and entrepreneurial narratives, and has led the commercialization exploration of the Neuroch project, using the pain points of VR motion sickness as the starting point to outline the development blueprint of the “Human Brain API”.
- Colton El-Habr (CTO): A dropout of Johns Hopkins University School of Engineering, as a hardware geek, he uses engineering systems thinking in the aerospace field to lead hardware research and development and neural modeling, transforming abstract neuromodulation concepts into technical solutions for wearable devices.
- Mark Khrapko (Scientific Advisor): With a background in biophysics from Harvard University, he brings interdisciplinary research depth to the team. Starting from the underlying mechanism of neuroscience, he verifies the scientific nature of the technical path and ensures that the physiological effects of vestibular stimulation are in line with academic logic.
Steven and Colton have no background in traditional neuroscience or hardware, but they have tackled key technical challenges in non-invasive vestibular neurostimulation through self-study. In the early stages of the project, they created a prototype device using batteries purchased from supermarkets, edible salt from restaurants, and sponges, challenging the stereotype that “academia believes that the technology cannot be commercialized” with practical actions.
Business model and customer base
Orbit adopts a B2C cooperation model in the early stages, mainly through hardware sales and subscription services. Among them, the hardware product is priced at $299, focusing on the market launch of non-invasive neuromodulation devices; The subscription service revolves around sleep optimization algorithms, providing users with personalized sleep improvement solutions, and enhancing user stickiness through continuous algorithm updates and data services.
On the one hand, it will cooperate with VR equipment manufacturers (such as Meta, Apple and other technology giants) to integrate vestibular nerve stimulation technology into VR devices to solve users’ pain points of motion sickness in virtual scenes and improve the device experience. On the other hand, it explores the B2B2C model and cooperates with medical institutions to develop customized neuromodulation programs, such as providing specialized solutions for medical scenarios such as insomnia and metabolic disorders.
Why it’s worth paying attention to
- As the first commercially available non-invasive vestibular stimulation protocol, this technology is technologically disruptive as it circumvents the ethical controversies faced by invasive brain-computer interfaces such as Neuralink.
- The technology can be extended from the consumer VR field to medical and health scenarios, with strong reusability and a high market ceiling.
- VR motion sickness is the core pain point of the 100-billion market, and the annual growth rate of medical applications (such as sleep technology) exceeds 30%, and the market potential is huge.
AUG Therapeutics (Jennifer Lin)
“A platform to accelerate the development and distribution of rare disease drugs, filling clinical gaps through asset acquisition and redevelopment”
positioning
Become a “drug bridge” in the field of rare diseases, solve unmet medical needs (especially glycosylation disorders), and shorten the path from drug development to patient use.
products
Currently, AUG Therapeutics has two main product forms: in-licensing and formulation optimization.
In-licensing focuses on preclinical or early-stage rare disease drug assets, such as rapidly expanding the pipeline layout in the field of rare disease treatment through the acquisition of Avalo’s related assets, filling its own R&D gaps in specific disease areas, and accelerating product iteration and commercialization with external resources.
Formulation optimization is to change the injection to oral/ophthalmic dosage forms to improve the convenience and compliance of patients. Among them, the core pipeline under investigation, AUG-803, is an oral L-fucose formulation that targets rare diseases such as glycosylation disorders (such diseases cause multisystem dysfunction due to abnormal glucose metabolism, and current treatment methods are limited). The drug has successfully entered the phase III clinical stage, and if successfully developed, it will provide more accessible and accessible treatment options for patients with rare diseases around the world, and is expected to become the company’s breakthrough product in the field of rare diseases.
The core values of AUG Therapeutics’ products are as follows:
- Relying on the safety data of existing drugs, through dosage form modification (e.g., injection to oral) and indication expansion, the development risk is reduced and the R&D process is accelerated.
- Accelerate drug time-to-market with FDA orphan drug priority approval lanes to accurately address unmet clinical needs in rare diseases.
- By reusing expired patented drugs, avoid high investment in R&D from scratch and achieve efficient control of R&D costs.
Founder background
Jennifer Lin (CEO) has a non-scientific background and has worked in biomedical funds such as Logos Capital and Newfoundland, and is good at capital operation and resource integration, effectively making up for the shortcomings of traditional scientific research entrepreneurs in terms of commercialization capabilities. After founding AUG Therapeutics in 2020, she quickly pushed the core pipeline into Phase III clinical trials with strong execution, and also completed a number of asset acquisitions (such as Avalo assets), demonstrating her industry layout and project promotion capabilities.
Business model and customer base
Business models include B2B and B2G
- B2B Pharmaceutical Collaboration: Accelerate commercialization by licensing improved drugs (e.g., regional distribution rights) to large pharmaceutical companies with their mature manufacturing and sales networks. For example, rare disease drugs with optimized dosage forms are licensed to multinational pharmaceutical companies and promoted in specific regional markets to achieve resource complementarity and risk sharing.
- B2G government/healthcare negotiations: Focus on orphan drugs and negotiate with governments and healthcare providers using their “high-price, rigid demand” characteristics (e.g., pricing $100,000/year/patient). By proving the clinical value and economy of drugs, strive to be included in the medical insurance catalogue or obtain government procurement support to ensure stable income in the high-net-worth market.
AUG Therapeutics has entered into a strategic partnership with Mayo Clinic, which has topped the ranking of the world’s best hospitals for 6 consecutive years. Mayo Clinic’s participation in its clinical studies and data validation will greatly enhance the credibility of AUG pipeline drugs. In addition, with Mayo Clinic’s extensive experience in FDA priority review, orphan drug designation, and more, AUG can optimize clinical trial design and reduce drug time-to-market.
Why it’s worth paying attention to
- By acquiring existing drug assets instead of independent R&D, R&D risks are reduced, and the safety data of mature drugs is used to shorten the time-to-market cycle and achieve asset-light + high efficiency.
- With only 5% of the world’s approximately 7,000 rare diseases approved for treatment, AUG Therapeutics’ “fast track” model (e.g., acquiring early-stage assets, optimizing dosage forms, and leveraging policies to accelerate approval) accurately responds to more than 95% of unmet clinical needs, creating a huge market space.
- Relying on the advantages of orphan drugs’ exclusive policies, including a 7-year market exclusivity period and a high-price pricing strategy, it provides enterprises with long-term commercial return protection.
Potential risks
- Independent R&D capabilities: Mayo Clinic’s support focuses on clinical research and commercialization, rather than drug discovery. If you rely on the in-licensing model to expand your pipeline for a long time, you may face the problems of weak core R&D capabilities and insufficient sustainability of follow-up pipelines, making it difficult to form technical barriers to independent innovation.
Human Interaction & Talent Strategy
Innerphases (Aidan Smith)
Founder background
Aidan Smith studied computer science and engineering at Georgia Tech and dropped out in 2025 to start his entrepreneurial journey. As early as the age of 16, he joined Chess.com to participate in product development, launching several features and gaining a wealth of practical experience.
Since then, Aidan has joined Neuralink, focusing on machine learning research in the direction of brain-computer interface (BCI), especially speech decoding technology, hoping to help ALS patients restore their ability to communicate with the outside world. He is not concerned with “making healthy people more convenient”, but about “using technology to break the limitations of the body”.
Aidan is also an exploratory “crossover engineer” who has completed the Colorado Trail hike alone, is equally passionate about mycology, history, and other fields, and is good at integrating thinking methods from different disciplines into technological innovation. As early as the fourth grade of elementary school, he began to think about the “practicality of knowledge”. Therefore, he is more concerned with using technology to break through physiological limits (such as helping ALS patients), rather than optimizing the convenience of life for healthy people.
Extraordinary (Sigil Wen)
“A ‘Cross-Border Career Transition Highway’ Connecting the World’s Top Talent and U.S. Tech Companies “
positioning
Become a “leader” in the field of global high-end talent flow, break down the barriers to visas and integration for top talents to enter the U.S. technology ecosystem, and promote the formation of a global innovation network of “talents without borders”.
products
Extraodinary’s core business is the full process of O-1 visa agency services, focusing on the “fast track” model (increasing the O-1 visa pass rate to 92%, far exceeding the industry average of 65%), covering: qualification assessment AI tools, automated material generation systems, and the world’s top talent database. Specific services include:
- Qualification Assessment AI Tool: Users input their resume information (such as GitHub code volume, social media influence data, etc.), and the system can automatically generate an O-1 visa match report with an accuracy rate of 85% through algorithms, assisting users in accurately evaluating the feasibility of their application.
- Automatic material generation system: An NLP model trained on 1000+ successful cases that can automatically fill in visa application forms and intelligently mark gaps in key supporting materials, reducing manual errors and improving application efficiency.
- Global Top Talent Database: Build a labeled talent data network (such as “AI ethics experts” and “blockchain developers” in subdivisions), and provide B2B paid search services for enterprises to help employers quickly locate overseas talents who meet O-1 visa qualifications and achieve efficient matching between talents and corporate needs.
Extraordinary’s derivative services cover accurate matching of talents with enterprises (such as matching with Silicon Valley startups), landing support such as housing and tax consulting, closed-door social networks for top talents, etc., while lowering the threshold for obtaining visa knowledge through the iconic symbol of the “Extraordinary Alien” hoodie and the YouTube content matrix.
Founder background
Sigil Wen dropped out of the University of Pennsylvania’s M&T dual degree program (Engineering and Business) and worked on the MIT Entrepreneurship Program. As a Serial Entrepreneur, he started his first project with a monthly income of $100,000 in high school and later co-developed the social app Airchat with Naval Ravikant.
After eight months of complex O-1 visa applications, Sigil realized the inefficiencies of traditional immigration services and the long-term neglect of the “non-professional genius” group, and decided to reshape the industry with a product-oriented mindset. Sigil has strong resource integration capabilities, and as a Silicon Valley entrepreneur and investor, it has accumulated resources from law firms, enterprises, media and other parties to build a cross-border network of “visa + employment”.
Business model and customer base
Extraodinary’s current business model is B2B and B2C. Among them, the B2B business includes providing enterprise recruitment solutions to technology companies, charging a referral fee of $10k/person or an annual subscription fee of $50k (supporting unlimited talent matching), and cooperating with local governments in the United States to attract overseas talents, and charging a service commission of $3k per person. B2C services include O-1 visa agency services ($5,000 for basic package, $12,000 for expedited service with 72-hour express lane) and talent branding services (such as YouTube personal story packaging, industry influencer certification, $2-$5k fees).
The core users are the top 0.1% of the world’s top talents in technology, art, sports and other fields, such as engineers with an annual salary of more than $500,000 and creators with millions of fans. In terms of customer acquisition, Sigil Wen attracted organic traffic through social content marketing and created a “talent circle fission” with the help of a mechanism where old users could get a $500 discount for referring new customers.
The company has currently cooperated with the world’s top technology companies such as OpenAI and SpaceX to deliver talents, and recommends more than 50 high-quality candidates to cooperative companies every month.
Why it’s worth paying attention to
- The contradiction between the urgent need for global talent from U.S. technology companies and the complex process of the O1 visa is becoming increasingly acute, and Extraordinary fills this market gap with its “productization + community” innovation model.
- Abandoning the “academic qualifications/qualifications theory”, the founder endorsed his own dropout entrepreneurial experience, which naturally fits the characteristics of the “non-professional genius” group and forms an appeal to similar talents.
Potential risks
- Currently, uncertainty about U.S. immigration policies (such as tightening O-1 visa approval standards) can directly lead to fluctuations in business volume.
- Traditional law firms are prone to copying standardized service processes, and without technological innovation or strengthening the influence of brand IP, it will be difficult to build competitive barriers.
FINTECH
Ivy (Ferdinand Dabitz)
“Open Banking Version of Visa”
positioning
Become the standard layer for global A2A payments, solve the problem of cross-border payment fragmentation, and drive the efficiency revolution of “bank account as wallet”.
products
With instant payment by bank as its core competitiveness, Ivy connects to local real-time payment systems in 60+ countries around the world (such as Pix in Brazil and UPI in India) through APIs, allowing enterprises to directly receive funds from customers’ bank accounts, achieving instant arrival in seconds with a handling fee of less than 0.5%, and the cost is significantly lower than traditional card payment.
Its technical architecture has two highlights: first, it has a unified API layer, which is compatible with the ISO 20022 message standard, can map payment protocols in different countries in real time, and adopts a microservice architecture that supports 100,000 transactions per second and a delay of < 100ms; The second is compliance middleware, which has a built-in AML rule engine in 70+ countries around the world, automatically screens cross-border capital risks, and realizes the asset-light model of “local compliance + global service” through the acquisition of European licensed electronic money institutions (EMIs).
The product also covers borderless accounts, which support multi-currency settlement for 5,000+ banks and 400 million + users, as well as an enhanced risk control suite, which integrates real-time transaction monitoring, dynamic KYC, and anti-fraud rule engines to ensure payment security and compliance in an all-round way.
Founder background
Ivy founder and CEO Ferdinand Dabitz holds a Master of Laws degree from the University of Munich with a focus on fintech compliance research, and his early experience includes cross-border practice and entrepreneurial accumulation: he worked as a fintech intern at McKinsey and as a researcher at the Max Planck Institute for Innovation and Competition, where he was exposed to the global payments regulatory framework.
Financing history
Early investors have allowed Ferdinand to connect with technology veterans and banking partners from Thiel’s payments companies such as PayPal; Received a Series A (US$20 million) and Series A + (US$30 million) capital endorsement led by Valar to accelerate competition with traditional financial institutions such as Visa and Mastercard; Creandum’s investment layout focusing on European technology has helped Ivy to radiate globally with Berlin as its center and establish real-time payment system docking channels with local central banks such as the European Central Bank.
Business model and customer base
Ivy’s business model adopts a combination of B2B SaaS and transaction commissions: providing API interface services to financial institutions, charging an annual access fee (e.g., bank access fee of $100,000/year), while providing enterprise customers with flexible payment options – pricing can be tiered by the number of transactions ($0.05-0.5/transaction, the higher the transaction volume, the lower the unit price), or subscription (starting from $2,000/month, including 100,000 free transactions). This diversified charging model not only meets the needs of customers of different sizes, but also achieves continuous revenue growth through the dual drive of technical services and transaction scale.
At present, Ivy has achieved access cooperation with mainstream e-commerce platforms such as Shopify and Magento, serving well-known companies such as Ecosia (with 20 million monthly active users). Currently, the customer retention rate is as high as 92%, and benefiting from the growth in demand for cryptocurrency-compliant payments, the proportion of Web3 customers has increased from 15% to 35%, highlighting market penetration and customer stickiness in emerging fields.
Why it’s worth paying attention to
- Track scarcity: Focusing on the core pain point of global payment fragmentation, traditional giants (such as SWIFT) have been slow to respond, while Ivy’s “API unification” path among emerging players is the most replicable. If successfully unified global A2A payment standards, it is expected to replace Visa/Mastercard as the dominant position in the B2B payment space.
- By acquiring licensed institutions rather than building a self-built model, it avoids the long license application cycle and achieves rapid business implementation in regulatory depressions such as Latin America and Southeast Asia.
- Endorsed by Valar Capital, and joined by industry veterans such as Martin Blessing, former CEO of Commerzbank, as well as a number of banks and investment institutions, providing it with deep industry resources and capital support.
Potential risks
- The trend of financial decoupling in China, the United States, Europe and the United States may lead to the ban on Ivy’s payment channels in some countries, affecting the stability of its global business layout.
- Circumventing stringent regulations (such as US state-level money transmitter license requirements) through the acquisition of EMIs (electronic money institutions) may raise questions about “regulatory loopholes” and face compliance review risks.
- Major banks may jointly launch competitors (such as JPMorgan’s PayDirect) to counter Ivy with its capital and channel advantages and squeeze market share.
Fizz (Carlo Kobe)
“Becoming a pocket financial advisor for Gen Z with credit-building debit cards and smart budgets, Fizz is one of the fastest financial tools and platforms to gain popularity among college students since Venmo.”
positioning
Become the “first financial partner” of the new generation of young people, solve the blind spot of the traditional financial system for people without credit records, and use technology to lower the threshold for financial independence.
products
Fizz’s core products focus on the innovative integration of financial services and financial management, including three sectors:
- Credit debit card: Cooperate with banks to launch a credit limit service with no credit history threshold, the user’s consumption amount is automatically deducted from the associated account in real time, the daily clearing mechanism can avoid debt accumulation, and the transaction data is reported to the credit bureau simultaneously to help users gradually establish a FICO credit score.
- AI Financial Assistant: Generate personalized budget suggestions by analyzing user consumption data, and support setting limits by consumption categories such as dining and shopping, with visual charts and “available balance” real-time reminders to prevent overdrafts.
- Educational gamification module: Provide short video financial courses and interactive quizzes, combined with the reward mechanism of partner merchants such as Uber Eats, combine knowledge learning with consumption scenarios, and improve users’ financial literacy in a fun way.
Fizz breaks the vicious cycle of “no credit→ it’s hard to get loans” and allows college students to accumulate credit through daily consumption. And translate complex financial knowledge into the language of Gen Z, bridging the cognitive gap between young people and the traditional financial system.
Founder background
Carlo Kobe (CEO): Bachelor of Statistics from Harvard University, former product management intern at Merantix, Berlin AI lab, with a background at the intersection of data science and fintech. He personally experienced the pain points of being rejected for credit card applications during college, and promoted him to create a “threshold-free credit tool”.
Scott Smith (COO): With a background in finance from Cornell University, he co-founded ShoulderTap AI, an enterprise-level AI knowledge management platform, with experience in B-end products and venture capital. Specializing in corporate growth and commercialization, he leads Fizz’s campus promotion strategies (such as working with Ivy League schools) and investor relations, complementing Carlo’s capabilities.
Financing history
Fizz received endorsements from YC and established Silicon Valley funds. Kleiner Perkins also brings key resources in the field of consumer finance, such as facilitating cooperation with consumer scenarios such as Uber Eats and Domino’s, to accelerate the penetration of products and the expansion of application scenarios among young user groups.
Business model and customer base
Fizz’s business model includes multiple revenue streams: B2B SaaS + B2C + consumer rebates.
At the B2B level, financial education SaaS modules are sold to universities at a price of $1,000 per year; The B2C side adopts a subscription model of $4.99 per month for free basic functions and premium (including exclusive banking courses, higher credit limits, etc.), and at the same time receives 1.5%-3% consumption rebates from partner merchants such as Uber Eats, of which 0.5%-1% is returned to users.
The core users are college students and young professionals aged 18-25, covering more than 300 universities, including all Ivy League universities, with more than 100,000 users. Gain customers through the “College Financial Challenge” series of content marketing on campus ambassadors, TikTok/Instagram and other platforms.
As of 2024, the number of users on the platform has increased by 400% year-on-year, with an average user using the budgeting tool 23 times a month, a credit building function usage rate of 78%, and reaching rebate cooperation with more than 50 brands, and the “Fizz Friday” event has facilitated up to 12,000 transactions in a single day.
