The Davos Download

I round up the most relevant AI-in-finance news - the deals being done, who’s rolling out what, and what’s actually working on the front lines

…and whilst the headlines were focused more on geopolitical ultimatums, it was still a collection of the world's most powerful people. When they weren't trying to downplay possible global warfare, AI was the next hot topic.

So it's a great opportunity to round up some of the world's tech and finance leaders giving the latest state of play on AI use cases and potential. I've pulled together the most interesting quotes from the week.

Amongst them, Vista Equity also revealed more details on how they're deploying their "agentic factory" strategy across their 90+ portfolio companies. 30 are already generating revenue from agentic AI, with another 30-40 converting in the coming months.

And yes, another 2026 AI in enterprise report landed, this time from Deloitte surveying 3,000+ firms. I read through it and pulled out the novel insights versus everything you're used to hearing by now.

In This Week’s Issue:

The Davos Download

What the leaders actually said, and why it matters.

Amodei didn't walk back his prediction that AI could eliminate half of entry-level white-collar jobs within five years. "It's eerie the extent to which the broader public and politicians, legislators, I don't think, are fully aware of what's going on." He expects "ichthyic" job losses (massive, sweeping) rather than gradual displacement. On timeline to human-level AI: "I would say two to three years is my honest guess."

Why it matters: When the guy building the leading AI model says we have 2-3 years until human-level AI and 5 years until half of white-collar jobs are gone, that’s kind of worth paying attention to.

Hassabis pushed back hard on Amodei's timeline. Current AI systems are "nowhere near" human-level intelligence, and "maybe we need one or two more breakthroughs before we'll get to AGI." He identified specific gaps: memory, continuous learning, and long-term reasoning. "We'd like systems that can learn in the real world, personalise and change over time. That's still not solved."

On China: "They're very good at catching up to where the frontier is, and increasingly capable of that. But they've yet to show they can innovate beyond the frontier." Chinese AI is about 6 months behind Western labs. "Inventing something is 100 times harder than replicating it."

His advice to undergrads: become "unbelievably proficient" with these tools.

Why it matters: The scientist vs. the entrepreneur. Hassabis says 5-10 years to AGI, Amodei says 2-3. The gap matters. But both agree the impact on jobs is coming. They just disagree on how fast.

"We expect over the next years, in advanced economies, 60% of jobs to be affected by AI, either enhanced or eliminated or transformed. 40% globally. This is like a tsunami hitting the labor market."

Entry-level jobs are being eliminated first, making it "harder for young people to get to a good placement." The middle class is "inevitably going to be affected" as wages stagnate for roles not enhanced by AI.

Her appeal: "Wake up. AI is for real, and it is transforming our world faster than we are getting a handle on."

Why it matters: This isn't a tech CEO talking his book. This is the IMF saying 60% of jobs in advanced economies will be affected. The "tsunami" framing is deliberate.

"We've seen it in our own organization where we've already replaced some roles, including in our analyst program, with AI." BlackRock has hired a "substantial" number of people to focus specifically on AI implementation.

Why it matters: The head of the world's largest asset manager confirming they've already replaced analyst roles with AI. Not planning to. Already done.

"I'm not in the job apocalypse camp. Technology has been disrupting jobs, changing the way people work, destroying jobs, and forcing us as a vibrant economy to create new jobs for decades."

But he warned AI implementation might go "slower than people now think" as companies realize it's harder than expected. Goldman's "One GS 3.0" initiative is reimagining six core processes through AI, including onboarding and KYC. The goal isn't headcount reduction. "If we get this right, I don't think it significantly lowers the number of people we have. It gives us capacity to invest in growth."

Why it matters: Solomon is threading a needle. Best M&A year ever, AI everywhere, but don't panic about jobs.

AI valuations are "at a bubble" level. "You cannot value a $50 million ARR company at $10 billion." But the software selloff is creating "huge buying opportunities."

On what makes software valuable: "Software is not at all about the code or about the technology. Software is about your domain knowledge. Most software companies know a specific vertical, a specific process, a specific function so well that there are three to five companies in the world that know it, and about 20 individuals in the world that really, really know it. That is the franchise. That is the value. That is what you cannot replicate."

He's working harder than ever in 30 years of PE because of AI.

Why it matters: The guy who's done more software deals than anyone thinks AI application valuations are a bubble, but he's buying aggressively anyway. His "domain knowledge is the franchise" thesis is a direct counter to the narrative that AI eats all software.

Dismissed AI bubble fears. "There's no question demand is just insane. It's the largest infrastructure buildout in human history." Called robotics a "once-in-a-generation" opportunity. Noted there's a "great shortage" of workers for AI infrastructure buildout, with electricians, plumbers, and steelworkers seeing rocketing salaries.

Why it matters: Huang's incentives are obvious. But "largest infrastructure buildout in human history" is hard to argue with when you look at the numbers.

"You didn't do a particularly good job making the world a better place," he told the Davos crowd. On AI: JPMorgan now has 200,000+ employees with generative AI access and 2,000 people building AI systems. The bank is in a hiring freeze despite record years, reinvesting efficiency gains into technology rather than headcount.

Why it matters: Dimon's bluntness cuts through Davos pleasantries. JPMorgan is the template for how big institutions are deploying AI: massive scale, significant investment, but using gains to fund technology rather than expand headcount.

For the first time, acknowledged AI could become a bubble. "A telltale sign of if it's a bubble would be if all we are talking about are the tech firms. If all we talk about is what's happening to the technology side, then it's just purely supply side."

Warned that leaner companies will adopt AI more easily. Larger organizations risk being "stuck" if they don't translate resources into new production functions. On what determines winners: "Energy costs will be key to deciding which country wins the AI race."

Why it matters: When the CEO of Microsoft admits AI could be a bubble, that's notable. His framing matters: it's only a bubble if adoption doesn't spread beyond tech. The burden of proof is on everyone else.

News Digest

Vista's "Agentic Factory" Vision for Portfolio Companies

Vista Equity Partners unveiled plans to transform every portfolio company into an "agentic factory" by 2027. The $100 billion software-focused firm, which already announced plans to cut up to a third of its 700-person workforce with AI, is now pushing the same playbook across its entire portfolio.

CEO Robert Smith outlined the vision at Davos: standardized AI infrastructure across all investments, shared agent libraries, and centralized governance frameworks. The goal is turning portfolio companies into "AI-native" operations where agents handle routine work and humans focus on judgment calls.

The details:

Why it matters: Vista isn't just cutting its own headcount. They're building the infrastructure to do it across every company they own.

My take: This is the PE AI playbook made explicit. Standardize the stack, share the learnings, push it everywhere. If Vista succeeds, expect every software-focused PE firm to follow. The firms that figure out how to systematize AI deployment across portfolios will have a structural advantage.

Read more + Read more

Deloitte State of AI: The Gap Between Ambition and Redesign

Another enterprise AI report dropped this month. Deloitte surveyed 3,235 business and IT leaders globally for their annual State of AI in the Enterprise. A lot of the takeaways will be familiar by now: AI coming for entry-level jobs, pilot-to-production challenges, governance lagging behind deployment.

What's interesting to me is the gap between what companies want to do and how they're actually preparing for it. 74% plan to deploy agentic AI within two years. But 84% haven't redesigned jobs or workflows around AI capabilities. Most are focused on "educating employees to raise AI fluency" rather than rethinking what work actually looks like. You can't deploy autonomous agents into processes that were designed for humans clicking through screens. That's a recipe for expensive disappointment.

The sovereign AI data caught my attention too. 77% now factor an AI solution's country of origin into vendor selection. 58% build their AI stacks primarily with local vendors. Where technology is built is starting to matter as much as what it can do.

The details:

Why it matters: The ambition is there. The groundwork isn't. Deploying agents without redesigning workflows is like hiring a new analyst and giving them no training, no context, and no clear responsibilities.

My take: The 84% stat explains why so many AI deployments underwhelm. Firms are layering AI on existing processes rather than asking whether those processes should exist at all. The sovereign AI shift is worth watching too. Geopolitics is now a procurement consideration.

Read more

Other Interesting Things I’ve Read of Seen This Week:

Pokemon tests AI-generated Pokemon with fans (Jan 22) - The Pokemon Company showed AI-generated creatures to fans to gauge reactions. They're not replacing artists yet, just "exploring." Sure, and Vista is just "exploring" headcount reduction.

OpenAI building hardware device with Jony Ive (Jan 23) - Former Apple design chief working on dedicated AI hardware. Because what the world needs is another device to talk to robots with.

PwC: Only 10-12% of companies seeing AI benefits (Jan 20) - Latest CEO survey shows 56% report getting nothing from AI investments. Meanwhile 95% of gen AI pilots failed in August. The gap between AI hype and AI value remains a canyon.

Acquisition Intelligence is a weekly newsletter on AI in M&A for finance professionals, private equity investors, investment bankers, corp dev teams, and deal-makers.

For questions, feedback, or to share what you're seeing in the market, reply to this email.

P.S. I'm Harry, co-founder of DealSage. We're building an AI-native deal intelligence platform to help professionals turn their institutional knowledge into better decisions. If you're curious what we're up to, check out dealsage.io or just reply here