The speed of light and the speed of AI

The Tuesday Column by Eric Lefebvre, today on Thursday.

11:45 – Restaurant Time, Zürich HB

Last week, I was waiting for Binci for lunch.

At Zürich HB, in the restaurant called Time, which in retrospect may have influenced the direction of my thoughts more than the menu did.

Above the hall, the large station clock said 11:45.

Everyone trusts that clock. Nobody really looks at it. Trains orbit around it, people synchronize unconsciously with it, yet it disappears from awareness, like gravity or interest rates.

So, I watched it.

Near the large departure screen, a traveller stood motionless with her suitcase, eyes fixed on the track numbers. A man ran past as if thirty seconds had personally offended him. Same station. Same minute. Completely different speeds.

And then the question appeared:

What is speed, really?

The Boundary Einstein Discovered

We all know the number.

300,000 kilometres per second.

The speed of light.

We memorized it at school, probably next to the name Einstein, and then never thought about it again.

But the speed of light is not interesting because it is fast.

It is interesting because it is a boundary.

You cannot exceed it not because engines are weak, but because beyond that point cause and effect stop making sense. Push further and motion turns into reversal — arriving before leaving, effects preceding causes. The famous paradox: go fast enough and you could erase the conditions that produced you.

Einstein’s unsettling insight was simple: rather than allowing that contradiction, reality changes time itself.

Move faster → time slows down

Move faster → distances shrink

The universe protects order.

When Organizations Accelerate

Companies rarely think this way.

They believe speed creates progress.

• Faster execution

• Faster scaling

• Faster innovation

• Faster AI adoption

Until speed outruns comprehension.

Organizations are meaning-processing systems. Understanding must travel inside them, and it has a maximum propagation speed. Beyond it, action continues but causality breaks:

• Work starts before understanding

• Strategy explains decisions already taken

• Investment precedes business model

From outside this looks visionary. From inside it becomes disorientation.

Oracle – Financing Tomorrow

Right now, the AI industry is testing this limit in real time.

Take Oracle.

The company plans to raise roughly $45–50 billion to build AI data-centre capacity and openly accepts years of financial pressure (losses) in order to secure a position in a future computing market. Debt increases, cash flow weakens, and investors debate whether demand will justify the capacity being constructed.

Oracle is effectively saying:

We will finance tomorrow before tomorrow exists.

Not incompetence. An attempt to move economic causality forward in time.

Musk, xAI, and Burning the Present

Now look at Elon Musk and xAI.

xAI is reportedly burning around $1 billion every month just to train models and expand computing infrastructure, and Musk merged it with SpaceX to build planetary-scale compute capacity, even imagining data centers in orbit.

The logic is identical but more explicit:

We must build intelligence first, economics will follow.

Present losses purchasing future reality.

The industry-Wide Bet

These are not isolated cases. They are the same bet expressed differently.

Oracle finances infrastructure before the market matures. xAI consumes capital before the product stabilizes. Both are rational, if the future arrives fast enough.

But this is exactly the physics problem.

When you approach the speed of light, time itself must distort to keep causality intact.

In business we are attempting the opposite. We are trying to distort causality to keep speed intact.

When Cause Comes After Effect

An organization normally lives in this order:

experience → understanding → product → revenue

AI allows:

investment → infrastructure → hope → explanation

You reach the future first. And only afterwards ask why it works.

Sometimes it will. Sometimes you arrive in a world where the company that financed it no longer fits inside it.

In physics you cannot outrun causality. In business you can. But only once.

12:00 — Time Resumes

I looked again at the clock.

11:59

The traveller at the departure board moved the moment her platform number appeared, perfectly synchronized with reality. She disappeared toward her train, probably not thinking about trillion-dollar infrastructure or machines consuming the energy of small countries to predict the next sentence.

At 12:00, Binci arrived.

Lunch began.

The world returned to normal speed.

Final Thought

The speed of light protects reality from contradiction. AI forces companies to ask whether their own speed still protects meaning. Because the danger is not moving too slowly into the future. The danger is arriving before your business still exists to understand it.

Eric Lefebvre

References articles:

Oracle – massive financing of AI data centers: Bloomberg, Bloomberg, Reuters

Merger of SpaceX and xAI / Computing power on a planetary scale: Reuters, Reuters

Read also: Virality as a risk factor: when 5.9 trillion dollars evaporate in 30 minutes

---