Nobody gets fired for buying IBM.

It is one of the oldest lines in enterprise technology.

Back the market leader, minimise the risk, and move on.

But what happens when there is no IBM yet?

One of the things that stood out during two weeks in the US meeting wealth managers, asset managers and advice technology firms was not just the technology. It was the conviction behind the decisions being made.

LPL Financial — a listed firm with ~29,000 advisers and US$1.9 trillion in assets — has embedded AI startup Jump into its adviser platform.

Hightower Advisors — a High Net Worth firm with US$354 billion under management — ran a competitive pilot before signing an exclusive partnership with Zocks | AI for Advisors.

These are large, sophisticated firms making enterprise decisions on businesses that only came into existence a few years ago.

And that is the point.

These are not mature vendors with large enterprise sales machines. In many cases, the founders are still in the room, still leading the demos, and still making the sales calls themselves.

One founder I contacted on LinkedIn agreed to meet me for breakfast in New York. He arrived after closing a US$150 million funding round that morning.

He still showed up.

The technology is impressive and they have attracted plenty of venture capital. But the bigger observation is that leading wealth management firms have already decided the risk of moving too slowly may now be greater than the risk of backing the wrong provider.

There will be acquisitions. Some firms will pivot. Some will not make it.

But waiting for certainty carries its own risk.

The firms moving fastest have already reached that conclusion.

Phoenix had never been on my radar. Known for golf and red rocks, the CBD was a ghost town when I got there — 40-degree heat will do that. But hidden somewhere among the cactus were 2,500 advisers, asset managers and vendors gathered for Envestnet Elevate — one of the biggest wealth management and advice tech conferences in the US.

The platinum sponsors were BlackRock, Fidelity Investments and State Street. I assumed they were there to sell funds.

They weren't.

Every conversation was about models. Asset allocation frameworks. Outsourced CIO capability delivered directly to financial advisers — in some cases, free, with the economics captured in the underlying assets. The fund was no longer the product. The relationship architecture was.

It makes sense when you think about the context. As passive scaled, fees compressed. ETFs democratised access and destroyed pricing power in the process. The next frontier for the largest asset managers isn't a better fund. It's owning the portfolio construction decision itself.

The booths were full of advisers. Curious about what was on offer — and, to be fair, the white Owala water bottles and free golf balls.

The infrastructure for this already exists in Australia. Hub24 and Netwealth have been building managed account capability at scale for years — $293 billion in FUM, used by nearly 60% of Australian advisers, expected to reach 75% by end of 2026. The difference is who's driving the asset allocation. Here it's largely been the asset consultants. In the US, it was the asset managers doing it themselves.

The largest asset managers aren't waiting to be allocated to. They're now designing the allocation.

On the fringe of downtown Dallas I joined a queue in a back alley. Freight trains honked in the distance as we waited for security. The venue shares its name with an iconic Honkey Tonk — used to hosting rockstars. This morning's headliner was different. One that has exploded from relative obscurity to the top of the charts in under 2 years.

Through the black curtains, a familiar orange glow blanketed the room. The bar had been replaced with egg white omelettes and granola. The crowd was mixed – suits, sweat pants and everything in between – but they all had something in common – laptops. As the lighting dimmed, the orange screen faded to reveal, a chat box. Today's star was Claude.

Anthropic had run a multi-state roadshow to introduce Claude for Financial Services, specifically Co-work a plug-in, built around a group of skills and MCP connections. The Presentation was slick, the demo was slicker.

A live build workshop where attendees experience co-work first hand, a bespoke task, a show and share, and prizes. AI consultants demonstrated how to transform raw data in Excel to a board quality PowerPoint, with one prompt.

I heard from talking heads for 2 years that this technology is generational, seeing it live made that claim even more real.

But the most interesting thing I took out of the room, had nothing to do with the demo.

Enterprise software stocks are at multi year lows. The 'SaaSpocalypse' is apparently a foregone conclusion.

I'm not so sure.

What I saw in Dallas – was an LLM focused on productivity looking to equip the enterprise with skills and agents, not vertical disruption. The software businesses who are investing to improve their workflows and their data, don't appear to be the obvious targets.

It's clear this is a company who will go to market differently. No convention centre, no external branding – a milk bar poster in the car park with a small orange logo, a submission form requiring approval, limited seats, strict door times. It's all about the product.

It's also still a startup. Despite the $47 billion revenue run rate, and reported valuation close to a $1 trillion, the business is ~5 years old. Wifi was patchy, and there were a few bugs in the build. The technology is moving fast. The business around it is catching up.

The level playing field surprised me the most. The person next to me was from a Dallas based AI start up – absorbing the same demo, using the same support. Nobody had a meaningful advantage. That probably won't last.

As I left the dark room and stepped into the midday Dallas sun, a driverless Waymo stopped on the other side of the street.

For a moment I felt like I was in the future. A car ride without a driver, my day planned, meetings prepped and analytical work complete — as long as you had a wifi connection and a PowerPoint.

Then my phone buzzed. Flight cancelled. Time to find a real person who could actually solve the problem.