Black Swan Destroys The Data Center Narrative

When any data center application can be shown to run on an Apple mini - the scam is up.

America does not need new data centers to lead in A.I.

Today there are multiple streams of new technology bursting forth, unseen by the Wall Street Journal tech writers, the MIT Review rendering data centers dinosaurs which will be abandoned like JC Penney shopping malls in 4 -5 years.

Let’s start at the beginning - so why in the hell do we need new data centers?

Better yet, let’s go one layer deeper and ask the more pertinent question - what does a data center actually do? Bet we are the first ones to ever ask that question.

Why do we have data centers?

There could be lots of responses but if we boil them down to first principles, the answer is to process I/O.

What is I/O?

Every transaction - when you hit that enter key goes into the data center and grabs info giving you an answer to your query.

What is the difference between a Tahoe and a Lexus GX 460?

That answer isn’t sitting in a handy digital envelope - the computer did not know you were going to ask that question - so it didn’t prepare the answer.

The computer sends little messages all over the place, grabs pieces of data from what can be hundreds of places in the “database” and assembles the answer.

Those little messages are called I/O - meaning “input/output” and they consume over 95% of what that data center does.

Not 2%, not 14% - they consume the data center.

I/O was not a big deal with current technology doing queries.

Then A.I. jumped into our lives and instead of queries, A.I. gives opinions - which SUV is better, why it’s better, what tests show - and I/O explodes.

Current data centers cannot handle that volume of I/O so the ancient, obsolete software companies like Oracle tell the President they need half a trill of dough for new data centers.

A.I. does not need new data centers - Oracle does - Palantir does, SAP does - Microsoft and META need massive data centers.

Why?

They are I/O factories.

Why are they I/O factories?

These companies pretty much all use some variant of the relational model down deep - which was created to run HR and billing systems 40 years ago - when I/O was not a constraining factor.

A.I. made their underlying technology either obsolete - like Oracle - or needing huge new data centers - like the rest.

For 40 years they built their technology blind to I/O - and now they are stuck with I/O intensive tech stacks when the world just made a big turn.

The A.I. explosion of I/O caught these guys technologically asleep and they had to scramble fast to prove they were not obsolete.

Trillions in valuations are at stake when the world learns there is an alternative to these software firms - and their data centers - and that time is at hand.

Thus, in every state in America, the Data Center Madness game began.

At first, everything seemed pretty cool - put a data center in our wilderness area and a previously anonymous town gets great electricity, cool hip jobs, probably a Starbucks - and huge tax revenue.

Like all fairy tales, the truth hit hard - with town after town learning why having clean water is a good thing, data centers create almost no jobs, above ground AC wires need warnings for pregnant women - cataclysm.

Today - over $64 billion in data center construction is stuck in the mud with citizen oppo - we know because we personally are testifying in some of those lawsuits about why the whole data center thing is a scam led by obsolete tech companies who got caught not innovating.

Here’s some of the testimony from our team coming to a fight near you.

Any application requiring a data center, can be built with low I/O technology, and run on one or more Apple minis or Intel tiny computers, using less power than a Tesla, no water cooling, at 1/10th the cost and will run 100 to 1,000 to 1 million times faster - no data center needed.

Quantum computing companies are soon to be a dominant force in high compute problem solving. Trillions of dollars are being invested in quantum computing. Quantum computing does has no need for current data centers - none.

New technologies like virtual MESH are being demonstrated today, where the contents of an entire data center can run, instantly, on existing processors from a phone, tablet, server, Raspberry Pi - duplicating a data center with virtual no additional energy - or water - or environmental devastation.

A.I. applications are moving from LLM-centric “capture all knowledge” to where the money is - constrained applications, with small models, solving revenue producing problems like “how does my customer respond to this incentive?” These do not need big data centers.

Entirely new families of compute process device, signal, meter data from moving entities like autos, missiles, which require instant decisioning.

The latency of sending info to a data center is architecturally fatal.

Such compute systems cannot by definition operate with a centralized data center.

Oracle, Palantir and other legacy big data center vendors are scamming the press - elected officials - and you - that the future is massive energy-consuming data centers.

They are public companies and their valuations depend on people who do not know better buying their stock - for their over-40 year old technology.

Remember, these tech dinosaurs use a tech tech stack Gartner - the undisputed authority on tech futures, said is obsolete in 2025.

Well, if you haven’t noticed, it’s 2026.

Every current data center runs today’s conventional technology - the current obsolete techs like Oracle, VMware, SAP, Palantir and other 1980s technologies built to run HR and billing systems.

These technologies were never built for A.I.

They will never be optimized for A.I. because they are 40 year old dinosaurs.

There is a golden rule in software: You cannot run faster than your infrastructure - so you can build 100 data centers and they will ALL operate at the same speed as ANY ONE of them.

That speed is the length of an I/O wait state, times the number of transactions, pretty much.

The Black Swan just appeared.

A Black Swan is when everyone says all swans are white - everyone agrees - the press claims there have never been black swans - PhDs get grants to write pieces on why swans cannot be black.

But one day, a black swan just shows up.

You cannot deny you saw it.

It cannot be unseen.

The press cannot say say it is an anomaly - because it is - and if one Black Swan can exist - any number can exist.

There are companies - like in the Intel video below - who are regularly demonstrating massive, data center consuming applications - moved to low I/O wait state architecture - and the need for the data center disappears.

So coming to a town near you, will be the Black Swan team - an Apple mini in hand - demonstrating live to the entire audience a data center application - that needs 90 hours to run - that’s 5,400 minutes - plus power, water, staffs, lights, solar panels.

Yet, here it runs on our little Apple mini - not in 5,400 minutes - but in 9. Nine minutes, plugged into a wall. $2,000 computer.

The Black Swan - proving there is no application requiring a new data center - is here.

It cannot be unseen.

This week Intel caught on and did an interview with one of the low I/O wait state companies.

As the video notes, once you achieve low I/O wait states - any application can be built pretty much without a data center:

Follow the Black Swan Files Substack - because we are just beginning this fight.