NVIDIA’s enterprise AI ambitions face deployment hurdles as Bank of America struggles to operationalise its AI Factory. Apart from tech giants, NVIDIA also sells its AI solutions to organisations and banks like Bank of America. However, recent internal emails hint that the massive company is struggling to use the technology. According to internal emails viewed by Business Insider between Bank of America and NVIDIA sales representatives, it looks like the bank is having trouble deploying NVIDIA’s solutions.
The publication says NVIDIA sells “AI Factory”, which includes both chips and software required to build, train and run large-scale AI systems for corporations like Bank of America.
Citing an internal email thread from November last year, the exchange between Bank of America and NVIDIA reveals that despite companies rushing to buy AI infrastructure, operational and regulatory hurdles make it hard for the bank to deploy the technology.
In the thread, NVIDIA executives also reportedly discussed how customers can make use of their AI products. “You sold us a Formula 1 race car, ” the bank told NVIDIA executives, “and now you have to help us as local car mechanics drive the race car!”
To this, another NVIDIA executive replied that the company “can’t just sell” AI Factory hardware but also needs to offer a software solution that helps customers make the most out of it.
Story continues below this ad
In a statement to the publication, Rumman Chowdhury, a consultant who advises companies on responsible AI, said that “buying GPUs or signing a cloud contract is a business decision; deploying AI is an institutional change. It’s much easier to approve a budget line item than to re-architect workflows, retrain teams, and rewrite governance processes,” She added that the gap between purchasing AI infrastructure and deploying it is common.
Bank of America lacks “MLOps skills”
Looking back at the meeting, an NVIDIA executive said Bank of America lacked the “MLOps skills in house”, which refers to skills used to deploy processes required for implementing AI models in real-world use cases.
He went on to say that the Bank of America thinks NVIDIA’s AI enterprise-focused software wasn’t “ready for their highly regulated banking industry.” NVIDIA also noted the challenges the bank faced in supporting various AI Models and software systems required to meet their needs.
Jumping in on the thread, NVIDIA Vice President Ian Buck said that the chip maker may need to jump in to listen to customer concerns. “Looks like they need help and/or our product is coming up short,” Buck said.
Story continues below this ad
AI may help streamline processes, but its deployment challenges aren’t limited to the banking sector. While banks are the first ones to make use of such technologies for things like credit decisioning, they often run into trouble because the scale of their data and customers is huge.
