August 2024 Fintech Newsletter: Learnings from Fintech DevCon 2024, the

Learnings from Fintech DevCon 2024: Debating Open vs. Closed Source Models and Apple’s NFC Ecosystem

James da Costa

The a16z fintech team is still digesting what we learned at Fintech DevCon, payments platform Moov’s annual conference for the fintech developer community. This year’s con highlighted the industry’s growth over the past year, as hundreds of fintech developers descended on Austin, Texas, and listened to talks from Sardine CEO Soups Ranjan, Bain Capital Ventures Partner Matt Harris, Gusto’s GM of Embedded Payroll Yi Liu, and more. The conference surfaced issues new and old, including trade offs around LLM model choices, an increase in fraud risks driven by generative AI, and the growing persona of the developer as a buyer.

On the former, financial service firms, like other enterprises, are debating whether to use open source models (e.g., Meta’s Llama 3.1) or closed source models (e.g., OpenAI’s GPT 4o) when embedding generative AI within their products and services. Per an a16z growth team analysis, nearly 60% of AI leaders noted that they were either interested in increasing open source usage or switching to open source, when fine-tuned open source models begin to roughly match the performance of closed source models. We heard this anecdotally at DevCon too, and it isn’t surprising given the industry’s strong emphasis on privacy and compliance. 

At Fintech DevCon, we also saw rich discussions around new platforms and payment options in fintech, driven by the opening up of Apple’s NFC ecosystem for developers to build on, following regulatory pressure in the EU. Developers will now be able to offer NFC contactless transactions within their apps, a feature previously limited to Apple Pay. This could unlock a wide range of interesting use cases inside and outside the Apple Wallet, including car keys, ID badges, government IDs, and event ticketing.

Ramp reports quarterly spend

We’ve been reading Ramp’s 2024 Summer Spending Benchmarks, which highlight AI as the fastest-growing expense in Q2 for Ramp customers. The most striking detail from the report was the dramatic rise in accounts payable (AP) spending with AI vendors, which has roughly 3xed since the start of the year. This is significant because everyone from SMBs to mid-marketing companies tend to reserve AP for longer-term spending and annual subscriptions. Therefore, this shift likely reflects companies seeing AI tools to be business-critical spend, as…