LINK HORIZONS 2026
Link Ventures' investment theses for early-stage startups. A map of where we believe the next great companies will be built, and an invitation to founders who want to build them with us.
Link Ventures
Link Ventures is Link School of Business's platform for creating, accelerating and investing in startups. Our motto sums up the journey: from idea to impact. Three engines work in sequence (school, accelerator and fund), with a presence in São Paulo and Miami, supporting founders from the classroom all the way to the investment check.
In the video alongside, an introduction to Link Horizons: what this map is, why it exists and how to use it.
The school: Link School of Business
It all starts at Link School of Business, the business school where the platform was born. The program combines business disciplines (finance, marketing, technology, people and management) with behavioral development and real-world projects, in an environment that treats entrepreneurship as a craft you learn by doing.
The creation journey: Startup School
Inside the school, Startup School is the program where the company gets off the ground: students build the team, ship the MVP, validate the idea with real customers and present results. The core KPI isn't academic: it's students with a company formed, generating revenue from a validated product or service. It's the natural gateway to the accelerator and the fund.
The accelerator: Link Vector
Our early-stage accelerator. After gaining the business tools and personal development, founders put them into practice with the Link ecosystem as a partner: mentorship, community, knowledge hubs with entrepreneurs and investors, and Demo Days.
The fund: Link Capital
The platform's early-stage investment fund, dedicated to backing the most promising companies in the ecosystem. It works alongside the accelerator to provide capital and support at the stage when every dollar and every piece of advice matters most.
The three pillars
Every thesis that follows passes through three filters. They aren't sectors; they're lenses. A strong opportunity for Link usually combines all three.
B2B businesses
Selling to companies means larger contracts, higher retention and more predictable revenue. In 2026, the best buyers are paying attention: YC reports startups closing pilots and multimillion-dollar contracts with the world's largest companies within their first year.
Global from day 1
The underlying technology is the same everywhere. What changes is who implements it and at what cost. LATAM founders combine world-class teams with implementation costs that speed up returns. Selling abroad starts as a real option, not a distant dream.
Applied AI and deeptech
We're not looking for generic AI. We're looking for AI that solves a specific industry problem, with proprietary data and deep integration. As Sequoia puts it, AI is going through a platform shift as transformative as the arrival of the internet.
The themes across the three criteria
Before diving into each theme, take a look at the full map. Each bubble is a theme from this document, positioned by our qualitative reading of the three pillars: the further right, the more intense the application of AI and deeptech; the higher up, the greater the potential to be born global; the bigger the bubble, the stronger the B2B fit. The solid navy bubbles are the 7 core themes; the hollow yellow circles are the radar, themes we follow without going deep.
Quick glossary
The startup world speaks half in jargon, half in acronyms. Here are the essentials for reading this document without getting stuck. In the tables, rows highlighted in green mark Brazilian companies.
AI as a service: the vertical application layer
What it is
AI as a service is the evolution of SaaS. Instead of selling a tool so the customer can do the work, the startup does the work and delivers the outcome: the books closed, the contract reviewed, the ticket resolved. The product that delivers this work is almost always a vertical AI agent: built for a specific industry, workflow and persona, loaded with the domain knowledge and integrations a generic tool will never have. This is the principle running through this entire document: specialization wins. The underlying language model may be the same (Claude, GPT, Gemini), but the moat (the hard-to-copy advantage) lies in the proprietary data, the integrations and the rules of each industry.
Why now
- The thesis comes from Sequoia partner Julien Bek, in the article "Services: The New Software": for every US$1 spent on software, US$6 is spent on services. The global professional services market exceeds US$6 trillion, versus about US$700 billion for software. AI-powered service companies run at close to 70% gross margin, versus 90% for pure SaaS, but they capture a market six times larger.
- The easiest entry point is what's already outsourced. If a company already pays a third party to do the work, it has already accepted that the work can be done outside, which makes switching to an AI product structurally simpler.
- The proof that the model works is the fastest revenue curves ever seen in software: Harvey (legal) reached about US$190M in ARR (annual recurring revenue) and an US$11 billion valuation; Bret Taylor's Sierra (customer service) reached about US$200M in ARR and a US$15.8 billion valuation, charging per resolution rather than per seat.
- Gartner projects that 40% of enterprise applications will embed task-specific agents by the end of 2026, up from less than 5% in 2025.
- YC dedicated a request to "AI-Native Service Companies," listing insurance brokerage, accounting, tax and audit, and healthcare administration as preferred targets. a16z's "AI Inside" thesis models roughly 10x revenue expansion per customer for those delivering reliable agents.
- A structural advantage for LATAM: the underlying technology is the same for everyone, and the difference lies in who implements it and at what cost. Teams in the region deliver the same result with a faster return for the customer.
Companies building in this space
| Company | Focus | What it does | Accelerator / fund |
|---|---|---|---|
| Harvey | Legal | Agents that carry out complete legal work for law firms and companies | Sequoia, GIC, a16z |
| Sierra | Customer service | Voice and text agents that resolve customer service, charging per resolution | Benchmark, Sequoia, Greenoaks |
| EvenUp | Legal (personal injury) | Builds the injury claim end to end | Bessemer, Bain Capital Ventures |
| Rillet | Accounting | Automated closing of the books | Sequoia, ICONIQ |
| Decagon | Customer service | AI-native support agents | a16z, Accel |
| Cognition / Devin | Software engineering | Autonomous software engineer | Founders Fund, 8VC |
| Enter (Brazil) | Legal | AI agents that handle mass litigation (large volumes of lawsuits) end to end | US$40.5M Series A |
| Blip (Brazil) | Conversational AI and customer service | Conversational AI that automates customer service for major brands | Warburg Pincus, Microsoft (partnership) |
| Tako (Brazil) | Human resources | AI agents that automate payroll and HR processes | Ribbit, a16z |
| Logcomex (Brazil) | Foreign trade | AI that organizes import and export data for foreign-trade decisions | Riverwood, Softbank |
- What service do companies in your city already pay third parties to do (accounting, legal, collections) that an AI agent could deliver ready-made?
- What industry do you know better than most people? What unwritten rule of that industry would an agent need to learn to work in it?
- If you sold the outcome instead of the tool, how would you charge: per completed task, per result achieved or by monthly subscription?
AI agents in the back office: workflows for corporate B2B and SMBs
What it is
Agents that replace entire back offices in finance, accounting and, increasingly, in compliance, KYC/AML and fraud prevention. Unlike a chatbot that summarizes data, this is a digital worker that investigates the case, takes action and produces an auditable result. These are the functions every company needs and nobody enjoys running.
Why now
- The problem is big and expensive. In traditional AML, 90% to 95% of alerts are false positives, and producing a single suspicious activity report can take four days or more.
- Banks allocate 10% to 15% of headcount to KYC and AML activities and still detect only about 2% of global financial crime flows. That gap between cost and effectiveness is the opportunity.
- Variance raised a US$21.5M Series A for AI agents that automate KYC, KYB, AML, transaction monitoring and fraud detection, processing more than 70 million context signals per day.
- The regtech market for financial crime compliance is expected to grow from about US$4.5 billion in 2025 to well beyond that over the decade, driven by milestones such as the European AMLA.
- Direct validation from YC: the Fall 2026 RFS explicitly asks for "AI-Native Compliance Infrastructure," infrastructure that consolidates fragmented tools, reduces dependence on specialized headcount and provides real-time visibility across multiple regulatory regimes. In YC's view, whoever gets it right becomes essential infrastructure for any business operating globally.
Companies building in this space
| Company | Focus | What it does | Accelerator / fund |
|---|---|---|---|
| EvenUp | Legal back office | Produces case documentation end to end | Bessemer, Bain Capital Ventures |
| Ramp | Finance and back office | Automates spend, accounts payable and closing | Founders Fund, Thrive |
| Variance | Compliance and fraud | Agents that investigate KYC/AML and produce auditable decisions | Ten Eleven Ventures |
| Unit21 | Risk and AML | Agentic system that tunes rules and runs investigations | Tiger Global, ICONIQ |
| Vanta | Security compliance | Continuous compliance automation | Sequoia, Goldman Sachs |
| Idwall (Brazil) | Identity and KYC | Real-time identity verification and fraud prevention for businesses | GGV, Monashees, Canary |
| Neoway (Brazil) | Data and compliance | Data intelligence for risk analysis, compliance and credit (acquired by B3) | B3 (acquisition), Temasek |
- What behind-the-scenes work (checking, typing, cross-referencing spreadsheets) have you seen consume an entire team?
- If an agent resolved 90% of cases on its own, how would you handle the 10% of exceptions?
- How would your solution prove to an auditor that every decision was made the right way?
The physical world: industry and logistics with hardware + software
What it is
AI has become superhuman at code, text and images, but 80% of the global workforce doesn't sit at a desk. This theme brings together three fronts that feed each other: robots and flexible automation for factories and warehouses (hardware); the software that plans, routes and coordinates construction, logistics and field service teams (the digital brain); and the smart cameras, drones and sensors that collect dense data from farms, machines and infrastructure (the eyes). Whoever runs the system records real work as it happens, a dataset not even frontier models will have.
Why now
- YC's Fall 2026 RFS made this thesis a central request: there are now three types of workers to orchestrate (AI agents that quote and schedule, robots in the field and humans with wearables, devices that record the work). These sectors spend 10 to 100 times more on labor than on software.
- According to Crunchbase, robotics startups had already raised US$18.8 billion year-to-date in 2026, versus US$15 billion in all of 2025. Goldman Sachs projects a humanoid robot market of US$38 billion by 2035.
- Sensor costs have plummeted, and multimodal AI can read construction blueprints, photos, invoices and work orders, digitizing processes that used to require a human on site.
- YC cites the reference cases in physical data: Gecko Robotics collects data on critical assets with robots, and Sorcerer operates autonomous weather balloons used by the US government.
- The Brazil angle: construction, agri-logistics, mining and field services are enormous, under-digitized and full of founders who understand the operation better than any foreign vendor.
Companies building in this space
| Company | Subtheme | What it does | Accelerator / fund |
|---|---|---|---|
| Figure | Humanoid robots | General-purpose robots for manufacturing | a16z, Microsoft, NVIDIA |
| Apptronik | Humanoid robots | Apollo robot for logistics and industry | Google, Mercedes-Benz |
| Gecko Robotics | Data and inspection | Robots that collect data on critical assets and build predictive models | USIT, Cox Enterprises, YC |
| Trunk Tools | Operations software | AI agents for coordinating construction sites | Redpoint, Innovation Endeavors |
| Flexport | Logistics | Software to orchestrate global supply chains | a16z, Founders Fund |
| Solinftec (Brazil) | Agricultural robotics | Autonomous robots (such as Solix) that monitor crops and apply inputs with precision | Lightsmith, Unbox Capital |
| Tractian (Brazil) | Industrial maintenance | Sensors and AI for machine reliability and field services | Sapphire, General Catalyst, NEA |
| Cobli (Brazil) | Fleets | Turns telematics data into fleet management and safety | Softbank, Fifth Wall |
| XMobots (Brazil) | Drones | Latin America's largest drone manufacturer; maps crops, mining and infrastructure | Strategic domestic investments |
| Auvo (Brazil) | Field services | Software that organizes field maintenance and service teams | Domestic SaaS investors |
- How does a construction site, trucking company or maintenance crew you know organize its work today: paper, WhatsApp, spreadsheets?
- What valuable data is generated (and lost) every day on a job site, in a fleet or on a farm, and what's the cheapest way to capture it: sensor, drone, camera or robot?
- What factory or warehouse could you visit this month to see the bottleneck with your own eyes?
AI-native companies in traditional sectors
What it is
This is Theme 1 seen from the other side of the counter. Instead of building the AI that firms buy, the founder builds the firm itself on top of AI. From the outside the business looks traditional (it serves legal, accounting, architecture and immigration clients), but inside it runs with a fraction of the people and with margins the industry has never seen. The advantage: the founder doesn't need to convince a conservative buyer to adopt technology; they compete with that buyer in the end market.
Why now
- Crosby, an AI-native law firm backed by Sequoia, reviews contracts in hours instead of days and charges for the outcome. It's the proof of concept that the model attracts top-tier capital.
- Garfield AI, in the UK, was the first AI-driven law firm approved by the industry regulator (SRA), setting the regulatory precedent for the category.
- Professional services are a trillion-dollar market that has run on the same billable-hour pyramid model for a century. Every hour AI eliminates becomes margin for whoever operates AI-native.
- For students and recent graduates, this is the most concrete entry point in this document: it requires understanding a trade and rebuilding it with AI, not inventing new technology.
Companies building in this space
| Company | Sector | What it does | Accelerator / fund |
|---|---|---|---|
| Crosby | Law | AI-native law firm that delivers contract review in hours | Sequoia |
| Garfield AI | Law | First AI-driven law firm approved by the UK regulator | Independent capital |
| Pilot | Accounting | Accounting delivered as a service, with proprietary technology | Sequoia, Index |
| Eve | Law | AI that multiplies the capacity of plaintiff law firms | a16z |
| Contabilizei (Brazil) | Accounting | Digital accounting for small businesses, a pioneer of the model in Brazil | Kaszek, Goldman Sachs |
- What professional service would you hire cheaper and faster if someone ran it with AI: accountant, lawyer, architect, paperwork agent?
- If you opened that business tomorrow with 3 people and AI instead of 30, what would you charge and what would you deliver within 24 hours?
- Which rule from the industry's professional board (OAB, CRC, CAU) would you need to follow from your very first client?
Stablecoins and B2B fintech infrastructure in LATAM
What it is
Stablecoin-based payment rails for remittances, treasury and cross-border B2B payments. They settle in seconds, work 24 hours a day and cross borders without the friction of correspondent banking. For LATAM, it's less about speculative crypto and more about rebuilding the region's financial infrastructure from the ground up.
Why now
- According to McKinsey, based on Artemis data, stablecoin payment volume reached about US$390 billion in 2025, more than double 2024. B2B payments accounted for about US$226 billion, up 733% year over year.
- 71% of LATAM companies already use stablecoins for cross-border payments, according to Fireblocks. Brazil is the region's largest crypto market, with more than 90% of flows going through stablecoins.
- Consolidation and validation: Stripe integrated Bridge (acquired for US$1.1 billion), and ARQ (formerly DolarApp) raised US$70M from Sequoia and Founders Fund.
- YC's contrarian take in the Fall 2026 RFS: with prices down and builders leaving, YC declares itself "more optimistic than ever." Bear markets (periods of falling prices) are when real projects get built, and YC expects every startup in its portfolio to use crypto rails, from fundraising to payments, even without knowing it. It highlights BlindPay and Infinia building the developer interface and the ramps for Latin America, and bets on agentic commerce: agents using crypto networks as financial rails.
- Regulatory warning for Brazilian founders: in 2026 the Central Bank of Brazil, through BCB Resolution No. 561, barred regulated electronic foreign exchange (eFX) providers from using stablecoins and other crypto assets to settle the offshore leg of cross-border payments, effective October 2026. The thesis is still alive, but the regulatory framework demands extra care from anyone building in Brazil.
Companies building in this space
| Company | Focus | What it does | Accelerator / fund |
|---|---|---|---|
| Bridge (Stripe) | Stablecoin infrastructure | Stablecoin issuance and settlement via API | Stripe (acquired for US$1.1B) |
| ARQ (ex-DolarApp) | LATAM dollar neobank | Dollar accounts and cross-border transfers | Sequoia, Founders Fund |
| Bitso Business | LATAM settlement | US-Mexico remittance rails and regional treasury | QED, Kaszek, Tiger Global |
| Conduit | Cross-border payments | Stablecoin payouts in emerging markets | Circle, Dragonfly, Altos |
| Rain | Card issuing | Dollar-denominated cards on stablecoin rails | Norwest, Lightspeed |
| Lumx (Brazil) | Stablecoin infrastructure | APIs for banks and fintechs to integrate stablecoin payments; clients include Elo and BTG | Indicator, CMT Digital, Antler |
| Parfin (Brazil) | Institutional digital assets | Custody, trading and the Rayls blockchain for banks to enter the tokenized economy | ParaFi, Framework |
| Transfero (Brazil) | Brazilian real stablecoin | Issues BRZ, Latin America's largest stablecoin pegged to the Brazilian real | Strategic investors |
- Who in Brazil struggles to send or receive money from abroad, and how much do they pay for it today?
- What would a company gain if it received in 10 seconds what currently takes 2 days?
- How would you build this product while following Central Bank rules from day one?
Healthcare: applied AI in health
What it is
AI applied across the healthcare chain: clinical documentation, evidence-based clinical decision-making, administration and revenue cycle, and personalized medicine. It's the vertical where vertical AI turned into real revenue the fastest, starting with doctors' most universal pain point: documentation bureaucracy, which eats up 30% to 50% of their working time.
Why now
- Abridge raised a US$300M Series E led by a16z, with Khosla, at a US$5.3 billion valuation, with partnerships with more than 150 US health systems, a sign of how fast adoption is moving.
- Ambient scribes (AIs that listen to the appointment and write the clinical note on their own) generated US$600M in revenue in 2025, a 2.4x increase year over year, according to market data.
- AMA research shows the average physician already uses AI for 2.3 distinct use cases, up from 1.1 in 2023.
- An adjacent area highlighted by YC's Fall 2026 RFS: AI for the aging population. By 2030, one in five Americans will be over 65, with millions of unfilled caregiving jobs and 53 million family members doing this work unpaid. Voice interfaces, monitoring for independent living and care coordination are the B2B side of this wave.
- A strong Brazilian signal: Arvo, a platform that uses AI to prevent fraud in healthcare transactions, raised a US$20M Series A led by Kaszek, a B2B case born in Brazil within this thesis.
Companies building in this space
| Company | Focus | What it does | Accelerator / fund |
|---|---|---|---|
| Abridge | Clinical documentation | Transcribes the appointment into a clinical note in real time | a16z, Khosla |
| OpenEvidence | Clinical decision-making | Medical evidence search at the point of care | Sequoia, GV, Kleiner Perkins |
| Hippocratic AI | Healthcare agents | Voice agents for non-diagnostic clinical tasks | a16z, General Catalyst |
| Nabla | Clinical documentation | Ambient scribe with a strong presence in Europe | Cathay, DST |
| Arvo (Brazil) | Health fintech | Protects transactions and reduces fraud in medical billing | Kaszek |
| Isa Saúde (Brazil) | Home care | AI that makes hospital-at-home care more efficient | Early-stage investors |
| Huna (Brazil) | Early diagnosis | AI that detects signs of cancer in routine blood tests, with validation published in Nature | US$1.5M pre-seed |
- How much time does the doctor or clinic you know spend on paperwork instead of on the patient?
- Where does money get lost on the way between the appointment, the health plan and the payment?
- What healthcare task doesn't require a diagnosis (scheduling, reminding, explaining, following up) and could already be done by AI today?
Agtech: AI in the sector where Brazil is a powerhouse
What it is
AI applied to agribusiness, the sector in which Brazil is a global powerhouse. The logic: instead of importing technology built for American or European farms, solve the Brazilian farmer's pain points (expensive credit, climate risk, labor shortages, decisions made by eye) with tropicalized data and context, then export the solution to the world's other major breadbaskets. A farmer planting soybeans in Mato Grosso faces the same problems as one planting in Argentina, India or Africa.
Why now
- Radar Agtech Brasil, in its 6th edition (Embrapa, SP Ventures and Homo Ludens), counted 2,075 agtechs in 2025, up 5% from 2024. Of these, 83% use AI in their processes or products and 35% have AI at the core of their value proposition.
- Agribusiness accounts for about 25% of Brazil's GDP, and production financing is moving toward a field of data: satellite and AI monitoring, digital legal certainty and automated credit pipelines.
- Brazil concentrates about 74% of AI investment in Latin America, which makes São Paulo, Campinas and Belo Horizonte strategic hubs.
- Rural credit is moving to the "field of data": satellite and AI analysis cuts the decision to finance a harvest from weeks to days, and whoever builds that pipeline captures one of the country's largest financial flows.
Companies building in this space
| Company | Focus | What it does | Accelerator / fund |
|---|---|---|---|
| Carbon Robotics | Autonomous weeding | Machines that eliminate weeds with lasers, without pesticides | US venture funds |
| Solinftec (Brazil) | Agricultural robotics | Autonomous robots (such as Solix) that monitor crops and apply inputs with precision | Lightsmith, Unbox Capital |
| Nagro (Brazil) | Rural credit | AI that speeds up credit analysis for farmers | Agri-fintech investors |
| Traive (Brazil) | Agri risk | Calculates credit risk using satellite imagery, weather and economic data | Banco do Brasil (partnership), agri funds |
| Agrosmart (Brazil) | Digital agriculture | Sensors and AI to monitor weather and soil and support decisions in the field | SP Ventures, Inovabra |
| Agrolend (Brazil) | Agri fintech | Digital credit bank for small and mid-sized farmers | Lightrock, Yara, Valor |
- What farmer pain point do you know, or could you discover by talking to one this weekend?
- What data does Brazil have in abundance (harvests, weather, soil) that the whole world would pay to use?
- How would your solution work on a farm without a stable internet connection?
Themes we follow, without going deep in this edition
These themes are real and attract capital from top-tier funds, but in this edition they stay on the radar: they demand a very specific technical profile, a large capital check or are still searching for the winning business model. If you're building in one of them, we still want to talk.
Sectors with a huge digitization gap
Sectors that technology has ignored (laundromats, funeral homes, auto repair shops, salons, small services) run on notebooks and phone calls. AI has slashed the cost of building full-stack software, making it viable to digitize large, analog markets with no quality competition end to end. References such as Cents (laundromats), Shopmonkey (auto repair) and Empathy (funeral) show the hidden size.
Infrastructure for agents
The tools that make AI agents work: connectivity (MCP), memory, identity, monitoring. An extremely strong global theme, with big appetite from YC and a16z, but one that demands a deeply technical founder profile and heavy infrastructure.
Energy and compute for AI
AI's bottleneck has become energy and data centers, and Brazil has a clean energy mix to spare. We follow it through the efficiency software layer, which is lighter on capital than physical infrastructure.
Applied voice AI
Voice agents for collections, scheduling and customer service in Portuguese. Today we treat it as a delivery format for Theme 1, not as a theme of its own. If voice is the heart of your product, Theme 1 is your way in.
Cybersecurity in the age of agents
Every agent in production is a new identity with access to systems, and defending against agent manipulation still has no standard. Resilient B2B budgets, in a market where the early-stage category is still taking shape.
TechBio
AI to discover molecules, proteins and materials, with Brazil's bioeconomy as a unique asset. A long, capital-intensive development cycle, typical of frontier deeptech funded by specialized funds such as Lux Capital.
Proof of personhood
The layer that verifies there's a real human on the other side, against deepfakes and fraud. An explicit YC request in Fall 2026 and an acute pain for LATAM banks; we're watching and waiting for the winning business model.
AI for marketing and the creative industry
Generative AI has driven the cost of producing campaigns, videos and design close to zero. As a result, value shifts from production to strategy, data and distribution: the companies that stand out sell marketing outcomes as a B2B service (more leads, more conversions, more sales), measured and billed as performance. References such as Waymark, which generates complete video ads with AI, show the way. Creative startups with a clear B2B service behind them will find an open door in Theme 1.
Transforming education
AI tutors that adapt to each student's pace, automated assessment and tailored corporate training. YC dedicated a request to the theme in the Fall 2026 RFS ("The Primer"): a personal AI tutor able to expand every child's potential. It's a theme that directly intersects with what Link does every day as a business school, which is why we follow it closely, inside and outside the classroom.
The funds and programs we read
Link Horizons draws inspiration from how the world's best investors communicate their theses. These are the reference points we follow, both global and Brazilian.
Y Combinator (RFS)
YC's Request for Startups is the format that inspires this document. In the Fall 2026 RFS, the theme is AI entering the physical world: education, healthcare, defense, finance, infrastructure and work itself. For the first time, a request came from a sitting US Secretary of the Army. The edition's B2C requests (an AI tutor for children and consumer products for 1 billion people) fall outside our theses by choice, given our B2B pillar, but they confirm the breadth of the platform shift.
Sequoia Capital
Author of the "Services: The New Software" thesis. In launching its seed and venture funds, it describes this moment as a platform shift as transformative as the rise of the internet and seeks founders of any background, from all over the world.
a16z
Raised more than US$15 billion, with US$1.7 billion for AI infrastructure, US$1.176 billion for American Dynamism (defense and industry) and US$700 million for bio and health. In "Big Ideas 2026," it sees AI moving from tool to environment, system and agent.
Gama Fund
A co-investment of up to US$2M from Monashees and the Google AI Futures Fund for Brazilian founders at pre-seed and seed building AI-native companies. Its thesis speaks of going from Brazilian founder to global builder, exactly our global-from-day-1 pillar.
Kaszek and Monashees
The two largest VCs in LATAM. Kaszek says AI is to 2025 what the internet was to 1998-99, and that companies not born with AI in their DNA will struggle. Monashees has been investing since 2005 and backs companies such as Tractian.
Canary, Astella and Valor
Canary closed Canary IV, a US$150M fund, keeping its focus on pre-seed and seed. Astella, a value investor that treats AI as a cycle and preaches consistency. Valor Capital, a pioneering cross-border fund that bridges the US and LATAM.
The seven themes in this edition aren't fixed. They evolve, merge and gain contenders as the market moves. The market projections cited (for example, future figures for robotics and stablecoins) are third-party estimates, not established facts, and some valuation and ARR figures come from market estimates and funding trackers.
Build with Link
If you're a founder building on a B2B thesis, designed to be global from day 1 and with AI applied to a real problem, we want to hear from you. You don't need to fit one of these themes to apply; if one of them excites you, take it as extra validation to dive in.