I. The last time I felt this way was in 2007
I still remember the heat, the buzz, and the buzzcut of the guy next to me in line. It was Philadelphia in 2007 and I was there! The original iPhone was about to launch, and the Apple Store had the atmosphere of a rock concert. People were buying a future they did not think could exist. That sleek, glass-and-steel object promised to rearrange everything: how we communicate, work, flirt, and waste time. And it did.
Since then, nothing quite matched the emotional intensity of that moment, until now. (Oh, well, it was a bit in 2017 with the promised Tesla Model 3, but only for a small fraction of people, due to costs, mainly). But now the AI moment feels even bigger. Not just because of ChatGPT’s viral release or Midjourney’s visual magic or even OpenAI’s 800 million weekly users. It’s because AI is no longer a tool we use, because it’s quickly becoming the infrastructure we depend on. The new default. The silent OS of modern life.
Most people think of technological progress as a straight line. But anyone who’s worked in innovation, or taught it, as I do at the Bucharest University of Economic Studies, knows better. Innovation works in S-curves. You have the slow grind of experimentation, the sudden vertical climb of mass adoption, and the eventual plateau where returns diminish and the next curve starts. We’ve already seen several of these: the mainframe, then the PC, then the Internet, then the smartphone & cloud. Each of these was transformative. But none of them replaced thinking.
AI does.
And that makes it fundamentally different.
II. This time, it’s an arms race
Comparisons to electricity are cute. They’re also misleading. AI is the first technology that absorbs and then improves upon human cognition. This makes it less like electricity and more like nuclear energy. Hence the title: AI is the Manhattan Project of our lives.
We’re building an infrastructure stack that touches every domain of power: geopolitical, economic, intellectual. You can already see the pressure mounting:
-
Microsoft and OpenAI are co-designing 10 gigawatts worth of data centers (enough to power multiple cities), treating energy capacity like GPU allocation.
-
Nvidia’s H100 chips are already deployed like strategic assets, with US export controls effectively turning semiconductors into foreign policy.
-
China is hoarding smelters for legacy chips, just to avoid being cut off from 14nm-class tooling.
-
Sam Altman is reportedly trying to raise trillions to redesign the global AI supply chain.
-
Every startup deck now opens with some variation of: “What if we were OpenAI for X?”
This is what happens when a software company becomes an infrastructure player out of necessity. Just like the Manhattan Project required breakthroughs in physics, materials science, and logistics, AI requires breakthroughs in compute, energy, talent pipelines, legal frameworks, and trust. We’re scaling a revolution and revolutions need operating systems.
And just like Oppenheimer’s team, we’re building this faster than we can fully understand its implications.
III. S-Curves and Strategic Moats
Let me bring in a concept I teach my MBA students: S-curves are about strategy, not technology. When a company rides a steep S-curve, traditional moats erode. Switching costs vanish. Distribution advantages collapse. Linear org charts become liabilities. New winners emerge.
The AI S-curve is doing something wild: it’s not just disrupting industries, but it’s compressing them. Take what used to be a 10-step process, say, writing a business strategy report or building a landing page, and reduce it to two prompts and a click. You just eliminated layers of white-collar labor, from junior analyst to mid-level manager.
What used to take a 5-person team, now takes one founder with GPT-5 and a Stripe account.
The moat now is not scale, or capital, or brand. The moat is taste, speed, and distribution. Knowing what to ask, when to ask it, and how to ship it. This is why OpenAI is more than an R&D lab. It’s a distribution machine. 800 million users a week. 100 million per day. This looks like the next Google.
From my work with global corporations and European startups alike, I can tell you this: those who treat AI like a bolt-on will get swallowed. Those who rebuild around it, who replatform their operations, teams, pricing, and even their value propositions. will win.
And here’s where it gets personal.
In my courses and boardroom sessions, I often say this: you don’t need to outrun the bear, you need to outrun the guy next to you. In the AI race, the guy next to you is either sleeping on AI or overhyping it. Both are dangerous. The real advantage lies in understanding the compounding effect of early moves.
The best time to rebuild your company for AI was two years ago.
The second-best time is now.
Instead of a conclusion: We are the system prompt
What makes this moment so different is that we’re not waiting for AI to trickle down. It’s already here. In your inbox. On your Teams calls. Writing your code. Grading your student papers. Suggesting your business decisions.
And it doesn’t sleep.
The Manhattan Project changed the world because it forced humanity to confront the consequences of exponential knowledge. AI is forcing us to do the same, but in offices, classrooms, hospitals, and Slack threads.
If you’re a leader, educator, founder, or policymaker, the question is no longer “Should I adopt AI?”
It’s “What am I training my organization to become in an AI-native world?”
Because make no mistake: we are the system prompt and the next line is ours to write.
