The Coming Quantum Big Short
Is there a future for quantum without transactional applications?
Does quantum compute have a future, at stratospheric valuations, if 95% of all business applications can run today, at quantum speed, on current hardware?
That is the quantum bet.
The quantum bet, now underlying a trillion dollars of investment - is that current computers cannot be made to run faster.
And speed is everything, particularly for A.I.
But, what if current software can run at quantum speed - on current hardware?
Oops, that’s a big effing problem nobody thought to figure out.
Well, we are about to find out - and if you subscribe to this newsletter, since we are the guys doing it, you will be the first to see that future.
Quantum computing on special hardware has a place where it will add major value.
That place is where massive computation needs to be done which cannot be done today.
There is one and only one question concerning valuation.
Is the massive compute market (as opposed to quantum speed market) a market with an 11 zero valuation?
Bet nobody told you there are 2 quite separate quantum markets - the quantum brain market and the quantum speed market.
Let’s explore together.
Let’s separate these two babies before birth.
Quantum compute versus quantum speed.
We are going to use the term quantum brain instead of quantum compute - to keep the lines from tangling.
These are quite different concepts - missed by 100% of the tech press and 99.999% of the investor world.
We met that .001% - because they are the 3 hedge fund guys who called us about how to short Honeywell and a bunch of other soon-to-be brought-out quantum COMPUTE companies JPMorgan and NVIDIA are funding.
That little distinction is going to cost some early quantum investors a ton of money and make fortunes for some smart Wall Street types who told us about their upcoming put bets (it’s a stock option thing).
Quantum compute (brain) enables a problem like a big, ugly formula - perhaps a crypto key or something else - often scientific in nature - to be run and the quantum compute brain is exponentially larger than today. It is compute intensive.
For simpletons, it’s a bit like what happens going from that early Dell PC with a 4 mg memory to a game machine with NVIDIA boards hanging off the side. The brain is bigger. Don’t beat me up on the metaphor here - this is a new concept.
Quantum brain business problems are NOT the same as quantum speed business problems.
Aha, bet we are the first guys to wake you up to that little gem.
Quantum brain computers probably operate at quantum speeds, we haven’t seen one so who knows?
The problem for the quantum brain guys is that it does not matter.
It is speed, not compute/brain that everyone needs, particularly for A.I.
95% of all corporate applications today are transactional and do not need quantum brain - they need quantum speed.
Let’s take a company with a massive billing system. It runs for 90 clock hours in a dedicated data center.
The corporation can choose from two alternatives - a quantum brain solution with all new hardware or a quantum speed solution running on its current hardware.
Both run at quantum speed.
The quantum brain solution requires new hardware, new software, PhDs all over the place - it isn’t going to happen - probably ever - IF the customer can get that damn billing system to operate not in 92 hours, but in minutes, on a computer that plugs into a wall costing $6,000.
And, as the chart - FOR A REAL CUSTOMER - shows, moving applications to quantum speeds - not quantum brains - delivers all kinds of other nice things - like less storage, getting rid of those Indian body shops who add bodies to your maintenance bill and eliminating Oracle from your life forever.
We think, and our growing Wall street followers agree, if you can deliver quantum speed, on current hardware, there is NO NEED for quantum computers for 95% of all corporate applications - PARTICULARLY ARTIFICIAL INTELLIGENCE.
Why?
Because nobody is going to move their billing system to a new hardware platform, with all the security, all the stuff of 40 years of IT experience - if they can move it to quantum speed software - on hardware they know, they own, they have in their shop, and they know they can secure - any app converted in less than 90 days.
Then there is the energy challenge.
Who knows how much energy someone needs to run a quantum computer? Probably a lot.
But the quantum speed on current hardware solution is fast not because it computes any faster, but because it optimizes the 95% of what computers do - optimizes I/O wait states.
So the quantum speed application runs 1,000 to a million times faster with NO INCREASE IN ENERGY USAGE. The data center uses LESS energy.
Like nobody is going to choose quantum brain for corporate and government transactional applications - if they can get quantum speed on current hardware - now.
Once the speed and energy problems are solved, there is little need for quantum brains for transactional applications.
And remember 95% of all compute today is transactional.
The problem to solve in A.I. and almost all corporate and government applications is not bigger brain stuff.
Just last week NVIDIA agreed with our long stated position that LLMs are not the future - the future is small language models, doing constrained things - to solve specific problems.
That NVIDIA position undercut any need for “big brain” LLM quantum compute - opening the door to super fast, aggregated Small Language Models - in a MESH - operating without a data center, as a super computer, at quantum speed - built and put into production in days, not years.
Read more about it on our website: TheSustainableComputingInitiative.com
The A.I. application can run stuff from 1,000 to a million times faster - and eliminate the need for a larger data center in the process.
The largest corporate and government opportunity is converting any legacy application to quantum speed - on current hardware - in a quarter or less - saving 50% or more on data center energy usage - and flipping those savings to A.I. investment.
That is the business proposition with the 11 zeros behind it now being investigated by some clever hedge fund guys who told us about their bets.
Remember the quantum bet, above?.
The quantum bet, now underlying a trillion dollars of investment is that current computers cannot be made to run faster.
Well, that seemed like a pretty good bet.
But the question nobody chose to ask is what if software - NOT SILICON - can be make to run faster?
What if software, available today - can be made to run at quantum speed?
Computers may not run faster - silicon is silicon.
What if software can be made to run faster?
Maybe someone asked, we did, but not too many others thought to look.
The key to make software quantum-fast is to identify the building blocks constraining transaction speed - the I/O wait state.
The Fractal guys found clever ways to optimize I/O wait states and now every Fractal application - which we demonstrate on weekly Zoom calls - runs 1,000 times faster than the legacy application it replaces.
Then, due to the system architecture optimizing I/O wait states - making every application tiny and running in massively parallel tiny Fractals - even more speed.
Currently every Fractal application runs at 200 million transactions per second.
It’s a completely arbitrary number because all we have to do is add more Fractals and the speedometer looks like a Gartner Magic Quadrant - up, up to the right.
If you are one of the quantum investors who recently asked for a demonstration - which we gladly showed you - we are sorry for how you felt afterward.
We want everyone to join Fractal on this journey - of how software can be made quantum-fast - on current hardware, and set you up for making some money in your investment portfolio as Wall Street continues to add zeros to quantum brain companies.
The Fractal bet is 95% of all A.I. applications need the speed, not the brain.
100% of all A.I. applications need to run with far less energy - not more data centers.
Every CIO needs to make his or her data center 1,000 times more productive today - for A.I. without additional energy - so the I/O optimization strategy looks pretty compelling.
95% of all corporate applications are transactional - needing quantum speed on current hardware - over quantum brains - with new hardware almost nobody understands.
When any application can be moved to quantum speed, on current hardware, in a quarter, why would a company move it to quantum brain - taking who knows how long?
The question for the quantum brain guys is how can they support valuations that look like the national debt if 95% of the compute market may be out of reach?
We are about to find out the answer to this little question so subscribe here as we uncover the future together.
We continue to welcome new partners delivering quantum speed solutions, on current hardware, today, without a data center.
Subscribe and join us for a fun ride.
FractalComputing Substack is a newsletter about the journey of taking a massively disruptive technology to market. We envision a book about our journey so each post is a way to capture some fun events.
Subscribe at FractalComputing.Substack.com
Fractal Website: Fractal-Computing.com
Fractal Utility Site: TheFractalUtility.com
Fractal Government Site: TheFractalGovernment.com
Fractal Sustainable Computing: TheSustainableComputinginitiative.com
@FractalCompute
Portions of our revenue are given to animal rescue charities.
We are not publicly traded. We are building an ecosystem of software companies, each of which will take investments, delivering Fractal solutions to the market. Follow us at FractalComputing.Substack.com and we will be updating that site with partners when they are raising funding.
Are you publicly traded? I'd like to invest.