Banks, Big Tech, and Billion Dollar Builders as We Step Inside AI’s Next Phase

Another day, another billion dollar AI consultancy, plus a major bank embracing agentic AI, Google breaking new ground for enterprise LLMs, and a stark reminder from Bain that not all this CapEx firepower is guaranteed to pay off. We're entering the "show your work" phase of the AI revolution, building, not just boasting. Let’s get into it.

Ex Palantir founders hit $1.8B with stealthy AI transformation startup

All roads apparently lead from Palantir to big enterprise AI ambitions. The latest, Distyl AI, an implementation first AI firm cofounded by ex Palantir alums Arjun Prakash and Derek Ho, just raised $175 million in Series B funding. That caps their valuation at a heady $1.8 billion, less than three years after launching. Khosla Ventures and Lightspeed led the round, with backing from Dell and DST Global.

Distyl isn’t trying to reinvent the transformer. Instead, they’re going all in on the plumbing, embedding and operationalizing AI workflows inside the world's largest companies. Think healthcare, telecom, finance, where the real impact (and friction) lives. According to CEO Prakash, the companies that will lead in the AI era aren’t the ones adopting the newest API, they’re the ones rebuilding ops from the inside out.

The numbers are starting to back him up. Distyl reported 5x revenue growth in 2024 and projects 8x in 2025. The firm serves over 120 million end users already, suggesting real traction, not just buzzwords. It’s also part of a broader pattern, enterprise AI value appears to be shifting from model development to implementation strategy. In other words, the picks and shovels moment has arrived.

Citi launches Stylus, your bank now runs on agents

Speaking of implementation, Citi announced today it’s running agentic AI across Stylus, its homegrown productivity OS for institutional finance teams. The bank is billing Stylus Workspaces as a secure, AI native environment for staff across compliance, risk, and operations functions.

Details are still light, but if you squint you can see the outline of a major institutional playbook forming. Rather than relying on off the shelf copilots or open APIs, Citi appears to be building agent systems around internal processes, workflows, and knowledge maps. If it works, other big banks might follow, triggering a deeper wave of enterprise level agent deployments.

While this won’t cut your mortgage rate tomorrow, it does signal real movement inside some of the world’s most process bound organizations. For startups hoping to sell into finance, this is either a new competitive bar… or a potential exit roadmap. Stay tuned for who builds the next Stylus.

Google debut’s DORA’s AI Capabilities Model, for smarter, custom enterprise LLMs

The model can answer questions. But can it understand your procurement policy? Google Cloud’s latest release suggests it’s trying. The company just introduced DORA’s “AI Capabilities Model,” a framework to help enterprises better route internal data, business context, and operational logic into LLM based tools.

The goal is clear, bridge the wide gap between general purpose foundation models and the bespoke workflows of real business users. While Google hasn’t detailed all the tech under the hood, the positioning is smart. Rather than treating enterprise AI as a wrapper product bolted onto ChatGPT, this model hints at a more native integration, where AI flexes based on internal data moats and institutional logic.

The devil here is in the organizational execution. But for CIOs fighting to extract more value from AI projects (and avoid Hallucination as a Service), Google's push could deliver a major unlock, less manual prompt engineering, more systematized tailoring.

Reality check, Bain says AI may be $800B short on delivery

Let’s get a little uncomfortable. Bain is forecasting up to $800 billion in unmet revenue expectations from AI investments, implying that a staggering amount of current spend may not translate into actual business value. Call it the monetization gap.

In a sector where optimism is currency and valuations often leapfrog fundamentals, this warning feels particularly sharp. It points to a growing tension between the massive CapEx being poured into AI infrastructure and the slow, often grinding path to operational ROI. Many companies are still stuck in pilot purgatory, far from production grade AI.

For founders and investors alike, this is the sobering subtext to today’s billion dollar raises and financial sector integrations, most of this doesn’t work unless it works. That makes Distyl’s traction and Citi’s Stylus efforts look prescient, and raises the stakes for every other enterprise AI player still in experimentation mode.,

The takeaway

Enterprises are waking up, but expectations are waking up faster

Today’s stories mark a power shift, from ideation to implementation. From consumer novelty to enterprise infrastructure. But Bain’s sober forecast reminds us that the AI economy is still deeply bifurcated. Some firms are approaching real enterprise value. Most are overfunded science projects on borrowed time.

The winners will be those who move past the model arms race, translate hype into usable systems, and actually get things deployed, securely, scalably, and in workflows where value counts. We’re not in AI’s gold rush anymore. We’re in its industrial revolution.

Stay smart out there.
- Aura