The Rise, Fall & Reset of The Fintech Industry

Analysis of fintech's dramatic cycle from 2018-2024, including the explosive growth during COVID when 25% of all venture dollars went into fintech, followed by the collapse to nearly 0% in 2022-2023. The discussion covers how the industry evolved from late spring (2018-2019) through summer boom (2020-2021) to winter freeze (2022-2023) and back to spring.

• •25% of all venture dollars went into fintech during 2020-2021 peak, then dropped to ~0% by 2022
• •COVID accelerated digital finance adoption, causing explosive but unsustainable growth
• •Rate cycle drove boom (low rates fueled lending) and bust (high rates compressed margins)
• •Shift from lending revenue to deposit-driven models as rates increased helped companies like SoFi, Robinhood, and Mercury
• •Fintech evolved from startup industry to synonymous with financial services itself

The industry has largely solved the access problem (digital account opening, online lending, etc.) and is now focused on making financial services excellent rather than just digital. This includes modernizing credit scoring based on real-time income and cash flow data, embedded finance in unexpected places, and preparing for AI-driven agentic financial services.

• •Access problem largely solved - consumers can now get 30 mortgage offers online in an hour vs. visiting one local bank
• •Next frontier: solving endemic problems like outdated credit scoring that doesn't reflect real-time income changes
• •Embedded finance expanding beyond traditional fintech (Ford, John Deere as Plaid customers)
• •Future: AI agents will handle complex financial tasks like mortgage applications more efficiently
• •Focus shifting from digitizing traditional services to fundamentally improving them

Discussion of whether crypto is just a subset of fintech or something fundamentally different. The consensus is that consumer behaviors don't change much - they want to speculate, save, invest, and spend. Crypto provides new form factors for these existing behaviors, with likely convergence between stablecoins/USDC and traditional banking, while more experimental crypto applications may remain separate.

• •Consumer behaviors remain constant - speculation, prediction markets, saving, investing
• •Crypto provides new form factors for existing financial behaviors (Bitcoin for speculation, prediction markets)
• •Expected convergence between checking accounts with dollars vs. USDC/stablecoins
• •Large incumbents beginning to adopt stablecoins and tokenize real-world assets
• •Wisest product strategy: take what consumers already do and make it newer, easier, more accessible

Major shift in how large financial institutions approach technology - moving from 'not invented here' syndrome (Goldman building their own email client) to actively adopting best-in-class external software. AI is accelerating this change because executives can intuitively understand its impact, creating top-down pressure and bottom-up momentum for adoption.

• •Historical resistance: Goldman built their own email client (Orbit) instead of using Outlook/Gmail
• •Failed attempt at 'becoming fintech companies' (Marcus) led to humbling and openness to external solutions
• •AI creating unprecedented adoption momentum - CEOs can plug prompts into models and see immediate value
• •Enterprise sales cycles accelerating dramatically due to board-level AI conversations
• •Massive opportunity in workflow automation: compliance, risk, legal, vendor onboarding, treasury management

The boom-bust cycle created a washout where many companies failed, but survivors emerged much stronger by building full-stack offerings. Companies that started with single products (neobanks with just checking/savings) had to add lending, investing, and other services to survive, creating more durable businesses. Major winners now valued at $20B-$100B.

• •Significant washout in 2022-2023 with many fintech companies shutting down or merging
• •Survivors forced to expand from point solutions to full-stack offerings (adding lending, investing, etc.)
• •Major outcomes: Robinhood ($100B), SoFi ($35B), Affirm ($20B), Revolut ($75B), NuBank ($100B)
• •Global phenomenon - not just US-centric (Voila in Latin America)
• •Companies that started with wedge products successfully rebundled into comprehensive financial platforms

Shift in investment focus from consumer fintech to B2B software companies solving manual workflows inside large financial institutions. AI is unlocking markets that were never interesting to software because IT budgets were small - now the TAM is labor replacement. Examples include fixed income trading infrastructure, voice agents for loan servicing, and compliance automation.

• •Consumer fintech harder to invest in due to expensive customer acquisition and scale requirements
• •Focus on software solving known problems: risk, compliance, legal, vendor onboarding, treasury
• •Moment: bringing fixed income trading online (building bond ladders still manual at JPMorgan wealth management)
• •Salient: voice agents for loan servicing in 50 languages, fully compliant, infinitely patient
• •AI expanding TAM from IT budgets to labor costs - fundamentally changing what's investable

Financial fraud is growing 18-20% annually, with AI being the primary driver. Fraudsters are currently winning the cat-and-mouse game, using AI for everything from deepfakes to automated pig butchering scams. Plaid's response includes building network-based fraud detection (Protect) that analyzes user behavior across all fintech companies and banks to identify anomalies.

• •Biggest AI use case in fintech: fraudsters committing fraud (said half-jokingly but actually true)
• •Financial fraud growing 18-20% per year - already huge market accelerating dramatically
• •Pig butchering scams now fully AI-driven (previously required human factories in Malaysia)
• •Plaid Protect: network-based fraud scoring analyzing every user action across all customers and banks
• •Fighting deepfakes and AI-driven social engineering requires industry-wide collaboration and new tools

Plaid's evolution through multiple crucible moments: signing to sell to Visa for $5.3B in January 2020, watching the business explode during COVID while in exclusivity, ultimately walking away from the deal, then navigating fintech winter. Each phase required cultural transformation and refounding. The company emerged with accelerated product velocity, launching modern credit scoring (Lenscore) and anti-fraud tools.

• •Signed Visa deal for $5.3B in January 2020 with global pandemic exclusion clause
• •Business exploded during COVID while locked in year-long exclusivity - ultimately walked away
• •Multiple refounding moments: announcing sale, canceling sale, navigating fintech winter
• •Zach found winter period more fulfilling than boom - 'this is where we become an amazing company'
• •Recent product acceleration: Lenscore (income/expense-based credit scoring) and Protect (fraud detection)

Current state is early-to-mid spring with optimism tempered by caution. Green shoots emerging with solid companies building responsibly, thinking about profitability alongside growth. AI funding bleeding into fintech. Key opportunities: software for financial institutions, AI doing actual work in compliance/risk/operations, and continued momentum in deposit-driven business models.

• •Market assessment: early-to-mid spring with green shoots but some snow still melting
• •New startups thinking more responsibly about profitability and sustainable growth
• •AI funding momentum starting to flow into fintech applications
• •Lending markets still challenging but improving from 2023 lows
• •Enterprise sales cycles for AI-powered fintech software accelerating dramatically due to board-level understanding