Three conundrums for Fintech startups
Life is changing continuously. Nothing remains the same. Fintech is no different. Some areas are fairly stagnant, some are progressing and some sadly regressing.
Women remain underfunded, underpaid, and underinvested; VCs haven’t been disrupted or held accountable for the ways they allocate and control capital and for the concentration of power in just a few of them; and we can`t expect much progress unless we solve the Digital Identity piece of the Web3 puzzle once and for all.
Today, I`ve picked two areas to highlight because they are `different` now. `Different` because they are more important than in the previous Fintech cycle.
- The Financially vulnerable are a booming class of people.
In the first Fintech cycle, we took care of the unbanked and the underbanked. But guess what, the `underbanked` class has been morphing because their work, their needs, and the economy have been struck by bad, ugly, and not transitory surprises.
So, the fragile part of the economy has unfortunately grown.
Financial vulnerability can originate from low income (nowadays better said, low purchasing power), lack of savings, excessive debt, lack of retirement planning, unexpected medical expenses, and early retirement. These are plaguing developed markets like the US and the UK.
1 in 4 Americans can’t afford a $400 emergency.
Early Wage access fees are rising in the US and some Fintech EWAs are facing lawsuits (Vox reports Activehours). While EWA costs are undertaken by large US employers like Amazon and Walmart, this is not the case in the broader economy.
Ironically, the FCA reported that in the UK 1 in 4 adults had low financial resilience.
Can and will Fintechs tackle this new set of Economic (not only financial) problems?
- Geopolitics are increasingly impacting the digitalization of Financial services and increasing the cost of innovation.
The first phase of Fintech innovation benefited largely from the declining costs of technology (cloud, compute power).
The next phase is faced with the higher costs for 5G networks and new GPUs for the LLMs. Fintech startups won’t be able to leverage at scale these technologies.
In the current phase, the cost of acquiring the GPUs to run the advanced LLMs is high and the cost of running these LLMs is prohibitive. Only the large BigTech can afford these running costs.
Building and training LLMs is expensive. The cost of these GPUs can amount to millions of dollars.
The electricity alone used to train GPT-3 is estimated to have cost $100 million[1]. And training is getting more expensive as the parameters of these models are growing.
On top of that, Fintech startups won’t have the proprietary data to leverage OpenAI and create new services.
At the same time, the cost of 5G technology is rising due to geopolitics.
By the end of 2022, Germany had close to 60% of its 5G equipment sourced from Huawei. The cost of replacing it to reduce the dependence on this one Chinese vendor would cost them billions.
German Huawei ban to cost €2.5B and take years, no thanks to EU
From Embedded Fintech in the mobility sector, in smart cities, in agriculture, and in anything related to the Internet of Things, 5G has to be everywhere and at a reasonable and declining cost structure.
How are the new generations of Fintechs supposed to innovate at the convergence of the real economy, if 5G isn’t ubiquitous and cheap, LLMs are costly to train and operate and of questionable value unless owning large proprietary data sets?
[1] What Large Models Cost You — There Is No Free AI Lunch
📌 Subscribe to my YouTube Channel with my insights and industry leader interviews. New video every Wednesday: https://www.youtube.com/EfiPylarinou
📌 Apple Podcast 👉 https://buff.ly/3P1Cp1Z
📌 Spotify Podcast 👉 https://buff.ly/3xPyWaV
📌 Twitter: https://twitter.com/efipm
📌 Linkedin: https://www.linkedin.com/in/efipylarinou/
📌 TikTok: https://www.tiktok.com/@efiglobal
📌 Web: https://efipylarinou.com/
