The dual-customer thesis in point 4 is the sharpest insight here and it has an implication that goes further than the article draws out. When AI purchasing agents start evaluating suppliers, the criteria will be transparent and documented. Every supplier can see exactly what the agent needs, which means matching the spec becomes table stakes almost immediately. The competitive advantage moves upstream to something less visible.
Whoever builds the purchasing agent controls the evaluation architecture. The restaurant chain buying through an AI agent isnt getting an unbiased comparison, the agent was configured by someone with preferences and weighting priorities baked into the system design. The real sales fight moves from convincing the buyer to influencing how the buying system was built. Procurement AI relocates bias from the buyer's gut to the builder's prompt. the companies that figure this out earliest will focus less on optimising product pages and more on building relationships with the teams designing procurement agents. thats where the next generation of enterprise sales actually lives.
The AI boom isn’t innovation anymore.
It’s a power grab.
And almost no one is saying it out loud.
The IPO of Cerebras Systems didn’t just make headlines
It exposed the real game:
A handful of companies racing to control the infrastructure of intelligence.
Not apps.
Not products.
The foundation layer.
Because once you control compute
You don’t just lead markets.
You shape reality.
Everyone talks about AI like it’s democratizing technology.
That’s the biggest myth right now.
In reality?
It’s concentrating power faster than anything we’ve seen in decades.
Fewer players.
Higher barriers.
Bigger stakes.
Even giants competing with NVIDIA are burning billions just to stay in the race.
This isn’t a startup wave.
It’s an arms race with a financial moat.
And here’s the part that should worry you:
AI is now actively finding ways to break systems.
Using Anthropic models, researchers discovered new methods to bypass macOS protections.
Think about that.
We are building systems.
That can outpace our ability to secure them.
This isn’t evolution.
It’s asymmetry.
Attack scales faster than defense.
Meanwhile, governments are pretending they’re in control.
They’re not.
The US China rivalry has already shifted:
From trade → to technology → to AI dominance
This is now about:
Who controls chips
Who controls models
Who controls cyber power
Not ideology.
Not diplomacy.
Infrastructure supremacy.
And while all this is happening
The real world is starting to crack under the pressure.
At Samsung Electronics and across the semiconductor supply chain:
Demand is exploding.
Workers are strained.
Systems are stretched.
Because AI isn’t just software.
It’s physical.
Energy. Labor. Materials.
And we’re scaling it faster than the system can handle.
Regulators are already behind.
They’re delaying rules
While quietly tightening control over:
Data
Content
Cybersecurity
The dual-customer thesis in point 4 is the sharpest insight here and it has an implication that goes further than the article draws out. When AI purchasing agents start evaluating suppliers, the criteria will be transparent and documented. Every supplier can see exactly what the agent needs, which means matching the spec becomes table stakes almost immediately. The competitive advantage moves upstream to something less visible.
Whoever builds the purchasing agent controls the evaluation architecture. The restaurant chain buying through an AI agent isnt getting an unbiased comparison, the agent was configured by someone with preferences and weighting priorities baked into the system design. The real sales fight moves from convincing the buyer to influencing how the buying system was built. Procurement AI relocates bias from the buyer's gut to the builder's prompt. the companies that figure this out earliest will focus less on optimising product pages and more on building relationships with the teams designing procurement agents. thats where the next generation of enterprise sales actually lives.
Nice article. A lot of folks are still afraid of doing the hard work required to get an AI transformation done well.
Agree 💯
Could you share a few examples of healthcare-related projects you have worked on?