BitsOfDream logo
Blog IT Call LinkedIn
← Articles

AI everywhere: hardware, money, rules and uncomfortable questions

Robots entering factories, billions from governments, upcoming IPOs — and one question nobody can answer yet: who actually gets the efficiency dividend?

airoboticsgovernanceindustry

Today, after a rainy weekend spent reading, I finally managed to gather my thoughts and write.

Robots leave the lab

Let’s start with something concrete: Siemens has begun testing an Nvidia-powered humanoid robot on the floor of a German factory. This isn’t a demo — it’s a production trial, presented by the company as a fundamental step toward fully AI-integrated operations. Hours later, Chinese vendor Agibot claimed the first large-scale deployment of industrial embodied AI, declaring the technology has officially moved out of the pilot phase into industrial use. Two pieces of news that, taken together, say one thing clearly: physical AI is no longer an experiment — it’s becoming infrastructure.

A Capgemini report confirms the trend: companies are rapidly shifting physical AI from experimentation to concrete implementation. The distance between the lab and the factory floor is collapsing.

Where the money goes

The UK government just launched a $675 million fund dedicated to AI startups, joining a wave of sovereign AI initiatives being pursued by countries around the world. What’s interesting is that this isn’t just about technology funding — it’s a geopolitical positioning statement. Whoever controls AI controls a significant part of the industrial and economic future (and not just domestically).

At the same time, chipmaker Cerebras filed for an IPO, backed by deals with OpenAI and AWS. The AI semiconductor market keeps posting numbers that look almost unreal.

Rules — from Europe and the US

On the regulatory front, two mirrored developments. In Europe, a group of experts issued a formal call to help the EU regulate AI-based platforms, with particular emphasis on transparency and accountability. In the US, in the absence of federal AI legislation, California and the federal government are using an alternative lever: purchasing power — public procurement rules as a form of indirect governance. No law? Govern through contracts. It’s pragmatic, perhaps necessary — but it leaves many questions open.

Samsung and the uncomfortable question

At Milan Design Week, Marco Porcini — Samsung’s global Chief Design Officer — said something that stuck with me: human thinking should be augmented by AI capabilities. Not replaced — augmented. It’s a distinction that matters, and it resonates with the broader debate about the role of these cognitive tools.

But this is exactly where the most interesting thread of the day comes in, courtesy of Repubblica Tech: AI saves us cognitive time — but that time is automatically reabsorbed by the work system. Gianluca Spolverato puts it plainly: technological efficiency doesn’t necessarily translate into real freedom. It’s the paradox of progress: more tools, more expected productivity. The freed-up time gets immediately filled with new tasks.

My take

Looking at all of this together, it seems we’re living through a pivotal transition: AI moving from experimental technology to systemic infrastructure. Robots enter factories, governments commit billions, stock markets wait for IPOs. It’s a consolidation cycle that looks a lot like what we saw with the internet between 1998 and 2003 — except this time the pace is far more compressed.

The question that none of today’s news can answer — and probably can’t — is Spolverato’s: who actually gets the efficiency dividend? Workers? Companies? Shareholders? It’s an economic and political question before it’s a technological one. And until there’s a shared answer, we’ll keep building increasingly powerful tools on foundations we haven’t yet decided how to distribute.

In the meantime, I keep reading, working, experimenting and writing.

Sources: