BBVA and Goldman Sachs invest in OpenAI’s new venture

• Key insights: BBVA and Goldman Sachs are among the investors in DeployCo, an OpenAI venture for enterprise AI deployments.
• What’s at stake: Banks are ramping up their use of AI, with nearly three-quarters boosting investment in the past year. 
• Forward look: DeployCo will seek clients looking to embed new forms of AI deeper into their organizations.

As banks expand use of new forms of artificial intelligence, collaborations and investments are also accelerating in an effort to add more depth to deployments.

Processing Content

BBVA and Goldman Sachs are backing DeployCo, an OpenAI venture that will sell services to large organizations that want to embed AI into their core technology. The two banks are among the 19 large firms that have invested more than $4 billion in DeployCo, which OpenAI formed following an acquisition of Tomoro, an AI consultancy. The banks are taking a stake in DeployCo amid large AI deployments in financial services designed to transform large swaths of technology, in such areas as payments, investing, customer service, risk and internal processes. Nearly three-quarters of banks increased investment in AI in 2025, according to research from American Banker.

But there are shortcomings. Much of the deployment is still rudimentary, and more work is needed to enable AI to penetrate deeper into organizations, according to bank technology experts.

The continued demand for “consulting” services to implement AI reinforces that the technology as it exists is not capable of simple deployment to replace existing workflows at enterprise scale, according to Christopher Miller, lead analyst at Javelin Strategy & Research. “Investments such as these are what is needed to turn AI into a product that can be consumed by existing enterprise touchpoints and functions such as procurement, project management, and others, and less that they will have truly differentiated skills that will make things happen,” Miller told American Banker.

OpenAI’s venture

OpenAI says it will help large organizations determine where AI can make an impact, and how organizational structure and workflows can evolve with these deployments. OpenAI, which develops ChatGPT and other new forms of AI, reports it has more than one million business clients, and is focused on what it deems is the next step of AI as the technology becomes less experimental and more mainstream.

As the large language models that drive agentic and generative AI improve, businesses can apply AI to larger, more important parts of how they operate, according to OpenAI.

Beyond Goldman Sachs and BBVA, other investors include TPG, Advent, Bain Capital, B Capital, Emergence Capital, Goanna, SoftBank Corp., Warburg Pincus, WCAS, Bain & Company, Capgemini and McKinsey.

“AI is becoming capable of doing increasingly meaningful work inside organizations. The challenge now is helping companies integrate these systems into the infrastructure and workflows that power their businesses. DeployCo is designed to help organizations bridge that gap and turn AI capability into real operational impact,” Denise Dresser, chief revenue officer at OpenAI, said in a release. BBVA and Goldman Sachs did not provide comments by deadline.

A sea of AI

OpenAI is not operating in a vacuum, as there are a number of ventures that aim to add revenue by selling AI’s ability to take a deeper dive into an organization.

Anthropic earlier this month launched agentic AI agents that were built in collaboration with banks, insurers and other large companies. JPMorganChase and AIG are among the partners on Anthropic’s latest partnership, along with Goldman Sachs and other institutions.

“I’m struggling to find a major tech company that hasn’t announced some form of partnership with OpenAI or Anthropic at some stage now, and I expect these to increase as they both start nearing an IPO,” Gilles Ubaghs, executive advisor at Datos Insights, told American Banker, noting DeployCo is focused on taking that next step to full enterprise deployment with tangible ROI.

What is different in this case is BBVA and Goldman Sachs are more than just formal investors—and in BBVA’s case, it’s an entrenched design partner.

“That’ll be critical to getting any financial institutions on board with DeployCo, as in theory this will have particular focus on financial institution needs and realities,” Ubaghs said. “But there are a couple of big caveats. BBVA is a serial innovator and always seems to be first on every new tech wave. That’s great, but I think there is a certain amount of market expectation that they will be first so I question how much that moves the needle.”

Although the vast majority of organisations in developed economies now report they have deployed AI to some extent, IDC research shows that most deployments are broad, but shallow, according to Neil Ward-Dutton, vice president of global agentic automation and AI technologies at IDC. “And the benefits, when they are seen, largely stem from relatively modest personal productivity gains,” Ward-Dutton told American Banker.

An IDC survey from earlier this year showed that only 50% of organizations are able to point to any measurable business value at all from their AI investments so far. “And to be clear, this isn’t about showing positive ROI; it’s about showing any measurable value at all,” Ward-Dutton said.

If AI is going to show meaningful and measurable business value, then implementation has to look less like, “hey, let’s just subscribe to this AI assistant service delivered from the cloud and a lot more like good old-fashioned business transformation,” Ward-Dutton said.

What’s the holdup?

As AI improves, the challenge for banks will be to change internal systems to accommodate the new technology.

“The biggest roadblock to AI in banking is not the intelligence of the models,” Ignacio Segovia, global head AI digital engineering and agentic business transformation and Altimetrik, an AI engineering company, told American Banker. “It is the complexity of the legacy workflows they are forced to navigate. You cannot attach an autonomous agent to a broken, fragmented process and expect transformation.”

What is missing in most deployments is rigorous service blueprinting: mapping the end-to-end workflow, identifying where value is created, and determining which segments are ready for intelligence and which are not, Segovia said.

“The institutions that will scale AI successfully are the ones investing in simplifying the underlying infrastructure so they can clearly delineate which bucket each workflow belongs in. Without that clarity, you end up automating fragments while the human is left stitching together the gaps,” Segovia said.

Success at scale means really digging into gnarly questions around the availability of high-quality enterprise data; business process readiness; what existing software application and established automation tool estates look like; how to assure quality and ensure security; how to manage the people and skills changes associated with potentially upending business processes, working patterns and operating models; and more, according to Ward-Dutton

There’s now a huge race on to help organizations address these issues in safe, cost-effective ways, Ward-Dutton said. “And that race doesn’t only involve software vendors competing with each other, or indeed consulting services firms competing with each other; it increasingly involves everyone, including software and services, all competing,” he said.