Signal over Noise with Ulrike Hoffmann-Burchardi

Hello and welcome to Signal Overnoise. I'm Rikov Mahbohyadi,

CIO for the Americas and Global Head of Equities for UBS 12 Management.

This week I would like to focus on a signal that has not made the financial headlines last week,

but that I believe will likely become one of the most important market signals soon.

And that is that we are experiencing the AI super-exponential.

And the super-exponential comes in two forms, AI capabilities and AI adoption.

What makes super-exponential growth different from exponential is that the growth rate itself

accelerates. It's quite difficult to come up with an intuition of what super-exponential means,

because we are rarely encountered in the physical world.

Losely speaking, exponential is like a snowball rolling downhill.

Super-exponential when the slope gets steeper on the way down.

So let's talk about AI adoption first.

After Cheshire PT, we now have the next AI application with super-exponential adoption.

OpenClaw is a free, open-source AI assistant platform

that you can run and install on your own computer.

OpenClaw lets you create AI agents that can manage your inbox,

read your emails, highlight what's important, draft replies.

Put together briefings, a daily summary of your calendar,

unread messages, weather news and other tasks.

And shop and book for you online from groceries to concert and travel bookings.

Think of OpenClaw as a personal AI consumer service.

OpenClaw can support running multiples specialized agents simultaneously.

Users often run five or more agents and more than 1,000 agents

in some experimental setups.

These agents can work together and make sure that you get to the desired outcome,

such as buying the required equipment for your next AI personalized vacation adventure.

OpenClaw is the first open-source project in history that has gained 100,000 stars

on software collaboration platform GitHub in just 12 months.

Less than half the time it took the previously fastest adopted software

TensorFlow to have the same traction.

OpenClaw is now the highest start GitHub project.

The reasons for its viral success are that anyone can see and improve the code.

Your data stays local, you can add skills and customize it,

and it works with many AI providers.

So secondly, besides AI adoption,

AI capabilities have been growing rapidly.

The clearest demonstration?

AI coding.

In the last few months, AI coding has made huge leaps.

Models can now handle big projects in many languages.

Debug their own code and even write tests automatically.

They now contribute to software projects almost end to end.

So why the sudden jump?

A few things.

Bigger and better trained models.

Large amounts of example code and smarter ways for AI to learn from feedback

and improve on its own.

Quantifying these advances is difficult as the benchmarks themselves are struggling

to keep up with appropriate measurements.

In my view, Meeters Long Horizon performance benchmark for software tasks

is best suited to evaluate software performance.

It measures multitask performance, not just a single task.

And Meeters shows signs of super exponential growth

in AI software capabilities.

AI software capabilities previously have been doubling every seven months for six years.

But recently, Claude Opus 4.6 more than doubled.

It's predecessor's time horizon in just two and a half months.

So what do super exponential AI adoption and super exponential AI capabilities

mean for the markets?

Let's start with adoption.

The fact that AI applications are adopted at an unprecedented speed

poses a risk to existing digital platforms

and the durability of their cash flows.

The rise of agents also changes how consumers interact with the digital economy.

The largest profit pools in the digital economy are tied to advertising,

e-commerce, and software.

Currently, we interact directly with these parts of the digital economy.

But increasingly, agents are now becoming the middlemen

that guide consumer choices.

With agents autonomously executing tasks,

the power of existing digital platforms and brands may diminish.

For example, it does not matter how difficult it is to check out on a webpage

if an agent does it for you.

Price and performance will likely matter more than brand for many products.

Brand and platform trust will likely give way to agent trust over time.

This poses a risk to the modes of existing digital platforms

and may give rise to new, agentech platforms.

And secondly, uncapabilities.

AI coding is likely a harbinger of what's to come in other parts of knowledge work.

If there's data that can train AI,

AI is poised to become exponentially better.

Last week's release of GPT-F is a good example.

It's now on top of the leaderboard on Mercos APEX agent benchmark.

This benchmark measures model performance for professional services work.

GPT 5.4 excels at creating long horizon deliverables.

Such as typical investment banking tasks,

like slide decks, financial models and legal analysis.

As a result, productivity will improve,

but it also poses a risk to the employment of knowledge workers.

A job light economic growth scenario seems plausible,

further rate cuts because of a sluggish labour market,

which in turn lifts economic growth.

So what's our investment advice based on these conclusions?

On equities, we favor sectors in physical parts of the economy

that are less susceptible to the immediate AI disruption risks.

We see most parts of the digital economy

at risk from reimagination by AI.

One exception is cybersecurity software

that is needed to secure and guardrail AI agents and the data they use.

Devoursification with different asset classes,

including fixed income and commodities remains key.

Super exponential growth phases tend to end when they hit constraints.

When this happens, we will have more clarity

about the winners and losers in the public and private markets.

Until then, stay well, stay diversified and stay ahead.

UBS Chief Investment Office's investment views are prepared and published

by the Global Wealth Management Business of UBS AG

or its affiliate UBS.

This material has no regard to the specific investment objectives,

financial situation or particular needs of any specific recipient

and is published for informational purposes only.

As a firm providing wealth management services to clients globally,

UBS AG and its subsidiaries offer both investment advisory services

and brokerage services.

Investment advisory services and brokerage services are separate and distinct

differ in material ways and are governed by different laws and separate arrangements.

In the USA, UBS Financial Services Inc.

is a subsidiary of UBS AG and a member of THINRA SIPC.

For information, please visit our website at ubs.com

forward slash working with us.

For a full legal disclaimer applicable to the independent investment

used produced by UBS, please visit our website at ubs.com forward slash

CIO-disclaimer.