Banks and fintech companies depend on the cloud. Indeed, there would be no digital finance or fintech, at least not at today’s scale, without the benefits of outsourcing data storage and computation to the hardware and networking of the so-called hyperscalers: AWS, Google Cloud, Microsoft Azure, and perhaps Oracle. The only non-U.S. cloud providers are from China.
Depending on cloud is one thing. Depending on cloud provided by operators from just one superpower is another. There are other growing problems with this setup: tech debt, rising costs, and more.
When I interviewed David Gurlé about this topic, we kept the focus on the infrastructure and less on the software or the A.I.
A.I. dependencies are, however, back in the headlines, with the way the U.S. government and A.I. lab Anthropic are fighting over its latest release, Fable, and the Trump administration’s decision to ban foreign persons from accessing the model. This is going to have enormous ramifications for A.I. but also for the companies we rely on for infrastructure. Especially because Google, Microsoft (via OpenAI) and the other hyperscalers are also behind most of the leading A.I. labs.
One obvious alternative is open source. David’s company operates this way, as do others. However he also relies on Chinese open source model LLMs. If China mimics the U.S. approach on A.I. nationalism, we are indeed in unchartered territory.
I interviewed David because he’s a serial entrepreneur whom I first met when he was running Symphony, the Bloomberg challenger for trade-floor comms. Today he’s operating twin businesses, one in localized data centers and the other in distributed cloud services.
The hardware side is rarely discussed in the fintech world, which is all about software, but I think we are all going to be spending more time thinking about hardware, resilience, compliance, and sovereignty. I’d love to hear what you think about this topic and this interview.
Timecodes:
0:00 - Jame’s introduction of David Gurlé and his journey from Symphony to data centers and cloud
04:09 - Understanding why costs of cloud usage have risen, both financially and operationally
8:35 - Benefits of using hyperscaler cloud and the tradeoffs of bringing cloud back in-house
11:25 - Compliance and sovereignty
14:09 - Regulatory comfort with new architecture
16:05 - What will it take for meaningful change in how enterprises use cloud
18:59 - Financial services and crypto use cases likely to migrate from, or stick with, hyperscaler cloud
21:49 - Is co-location of servers going to make a comeback?
26:23 - The ethos of decentralization
27:48 - Building anew for agentic A.I.
30:47 - David’s update on his companies
