Thread by @ayushbanerjee

Infosys once sat at the same table as Elon Musk, Peter Thiel, and AWS. Backing a scrappy AI startup called OpenAI.

10 years later? Microsoft owns the prize, Infosys rents the future.

How did they blow $45B in a matter of days? 👇

December 2015:

OpenAI launched as a nonprofit and published a list of donors who collectively “pledged $1B” to support the mission.

Infosys was explicitly named among those early backers.

The driver? Vishal Sikka, then-CEO of Infosys.

PhD in AI (Stanford, 1996), advised by John McCarthy (coined “AI”), mentored by Marvin Minsky.

Built SAP’s flagship HANA before joining Infosys in 2014.

His thesis: “AI will change how systems work and firms must own IP.”

Sikka’s strategy was simple and brutal:

Automate everything that you can - because sooner or later - we will fire the unproductive idiots.

Repetitive engineering across a mass of billable hours - pivot from labour arbitrage to platforms, products and IP.

But Infosys never got the chance to finish the pivot.

The boardroom turned into a battlefield.

Vishal Sikka vs N.R. Narayana Murthy.

Silicon Valley-style risk-taking vs. Bengaluru-style conservatism.

By Aug 2017, Sikka quit. Infosys lost its only AI-native CEO.

Then came 2019.

The year that changed everything.

OpenAI flipped from “nonprofit lab” to “capped-profit company.”
Translation: donors could now become investors, partners, insiders.

That was the conversion moment. The golden window to turn goodwill into ownership.

Satya Nadella didn’t blink.

Microsoft wired $1B, locked in exclusive rights, and tied OpenAI to Azure.

Compute + IP + distribution.

Result? Microsoft went from cloud also-ran to the default enterprise AI platform in just 3 years.

That’s the difference between backing an idea vs owning the stack.

Infosys donated to the revolution. Microsoft bought the keys to the kingdom.

This was capture. Microsoft didn’t license models - it integrated them into products and channel reach.

That’s AI at scale.

Fast-forward:

OpenAI is now valued north of $80B, with whispers of a $150B+ valuation.

Microsoft’s stake? Rumoured at ~30%

Infosys’s stake? A resounding Aryabhatta.

Had a certain early donor converted their investment into equity, the win would likely have been massive.

Let that sink in:

A company worth ~$70B today could have sat on an AI asset worth $45B+.

Not in paper gains, in real leverage.

Instead, they are now paying Microsoft for access to the same tech they once seeded.

So what did Infosys actually lose?

Optionality.

The chance to own platform economics, influence roadmap, and capture long-term software margins instead of selling labour to those who do.

And, the lesson is brutal:

You miss paradigm shifts because you’re stuck in the boardroom.

When disruption is timed, internal feuds are fatal.

• Boards that prize short-term optics over long-term optionality kill all bets
• Strategic tech capital must be allocated as partnerships when mission-critical.
• Public founder feuds are time bombs for time-sensitive plays.

Sikka sketched a blueprint for an Infosys that led the AI century.

But the company tore it up.

Now Microsoft builds the house, Infosys rents a room, and Indian IT employees with 70-hour work weeks pray its billable hours do not get automated away.

Corporate history doesn’t punish in quarters.

It punishes in decades.

Infosys blinked. Microsoft pounced.

And Indian IT learned:

If you do not own the platform, you will rent the future at shrinking margins.

@threadreaderapp unroll