Oct 26, 2025

Funding Frenemies - Crypto and AI Crash the Party Forget Tiktoks of 2021 – investors’ flirtations have pivoted dramatically. In Q1’25, over half of the largest seed/A deals targeted blockchain and crypto infrastructure startups. (Yes, crypto – not the cute puppy coins – just massive cryptobanks and trading rails.) At the same time, fintech-focused AI companies are gobbling up a record share of funding: 16% of all deals in Q1, and jumping to 23% of funding dollars by Q3. Five of the top 10 deals in Q3 even went to AI-enabled finance platforms like Ramp and AppZen, cementing this trend. Another clear trend: bigger bets on fewer companies. After steady decline, total fintech funding actually perked back up to $10.3B in Q1’25 (its highest since early 2023), thanks largely to one $2B Binance megadeal. By Q3, funding was flat at ~$10.9B but mega-rounds (>$100M) made up a whopping 40% of that pile. In plain English, investors are flush with IPO regret, so they’re pouring more cash into fewer, later-stage deals. Indeed, mid-/late-stage rounds hit their highest deal share since 2022, while early-stage funding dipped below 70% of deals for the first time since 2020. (The median late-stage deal size is up 50% vs. 2024 – all those spreadsheet-crunching CFOs insisting on clear paths to profit.) Regionally, Silicon Valley still commands the lion’s share (US was ~39% of Q3 deals), but “unsexy” sectors have global players. Corporate VCs are all over crypto fintech: Coinbase Ventures led Q1 with 6 deals, OKX Ventures had 5. Meanwhile Asia, Latin America, and Africa are represented by success stories, even if their funding slices are smaller. (CB Insights’ Fintech 100 lists Uzum in Uzbekistan and Affinity in Ghana – so yes, fintech unicorns are popping up in “surprising” places.) The New Finance Workforce If you thought “AI in fintech” meant chatty bots for customer service, think again. Fintech 100 firms show that the real action is automating complex workflows. Eleven winners (across 8 categories!) are building purpose-built AI agents to handle finance tasks – debt collection, AML compliance, payroll approvals, you name it. For example, a startup called Borderless built an agent (on Cohere’s large language models) that automates global payroll and hiring, and Murphy’s agents boost debt recovery by ~40%. Catena Labs is even wiring up “agentic” payment rails so bots can pay invoices on our behalf. In short, specialized AI bots are now colleagues in the finance department, not just assistants. This wave is hitting accounting and operations particularly hard. Seventeen Fintech 100 companies (vs. only 10 last year) use AI to rip out manual chores in accounting, treasury, invoicing and payroll. Their pitch: speedier closes, real-time data, zero human error. (Campfire promises 95%-accurate reconciliations; Xelix’s agents hunt fraud and prevent duplicate payments; Niural’s “EMMA” runs payroll in 150 countries.) Crucially, these workflows are fast becoming “all-in-one” platforms: once you have an AI payroll engine, why not add banking, salaries-on-demand, and even corporate card features? Fintech 100 members Khazna (Egypt/Saudi earned-wage access) and InstaPay (Malaysia migrant payroll remittances) are proof-of-concept that embedded finance around payroll is the new normal. One quick aside: the hype cycle isn’t over – if it ever even started. Enterprises are still testing and integrating (as Manlio Carrelli points out, agent-only success is 1) high-quality data and 2) vendor management). But the scoreboard is clear: in the race between “traditional fintech” and “AI-first fintech,” the AI teams are pulling away. If you’re building fintech today, your odds of success probably involve AI in some way (just look: one in five new unicorns in 2025 is an “agent” startup).