> Oracle totally shyt the bed thanks to their AI investment

VP_Spiro_T_Cheney wrote:

They’ve burned over $20B in AI capex over the last 6 months only to see the cost of providing compute services rising faster than the revenue it generates.  Went from $11.2B cash flow positive to $13.2B cash flow negative in just six quarters.

Like crypto, AI is a bubble floating on a lake of rancid shyt.

The fact that your analysis is basically a 6 month analysis shows how flawed it may be.  We don't know yet.  Larry thinks this was a smart long term play.  You think he fuqed up.  One of you is a business genius.  We will see who's who in about 5 years.

wrote:

Oracle is a legacy business and has no practical use for AI. AI and ERP doesn’t compute.

The play is creating hyperscaling datacenters within the Oracle Cloud for AI itself, mostly with OpenAI being the client.  It doesn't have to be for ERP applications.

What the industry really needs is more foundry right now because of all the silicon going to these projects.

wrote:

That is just the circular revenue pipe dream. Meanwhile, Oracle has real customers with nothing in AI to cross-sell.

And IBM has real customers with nothing in mainframes to cross-sell either.

And Micosoft has real customers with nothing in Xbox to cross-sell either.

These are large companies with multiple multi-billion dollar products.

Spanky/Spiro is grinding out a 5% dividend investment play while the market is up 30%.

I'd be bitter too.

SamplesBoi wrote:

The play is creating hyperscaling datacenters within the Oracle Cloud for AI itself, mostly with OpenAI being the client.  It doesn't have to be for ERP applications.

What the industry really needs is more foundry right now because of all the silicon going to these projects.

To date the various players have spent roughly two trillion dollars on “compute” as Sam Altman puts it… and what they can charge for leasing out that compute doesn’t even pay the operating costs, let alone provide ROI on that two trillion dollar investment.

And the killer data point for Oracle is the fact that the operating costs rose faster than the revenues.

Even worse, unlike the railroad or fiber optic buildouts, most the assets they bought with those two trillion dollars have a useful life of only 2-4 years.

LLM AI is a bubble, what remains to be seen is who’s the sad fuqer that end up as bagholder.

SamplesBoi wrote:

The fact that your analysis is basically a 6 month analysis shows how flawed it may be.  We don't know yet.  Larry thinks this was a smart long term play.  You think he fuqed up.  One of you is a business genius.  We will see who's who in about 5 years.

I think Larry's information is of a higher quality than OLG's - if Larry thinks AI is real, it probably is. 

When choosing whom I should believe, either one of the four-five most powerful tech titans globally or an idiot on this board, I'm going to trust the titan. 

ORCL is being punished because their free cash isn't great - they're too levered.

VP_Spiro_T_Cheney wrote:

To date the various players have spent roughly two trillion dollars on “compute” as Sam Altman puts it… and what they can charge for leasing out that compute doesn’t even pay the operating costs, let alone provide ROI on that two trillion dollar investment.

You're finally right, Pillowbiter.
None of these AI operations do anything but burn cash.

Whomever thinks the future of AI is giant data centers moving giant LLMs is a fool.

All that will trickle down to bespoke small models that have specific uses. And an absolute shyt ton of hand-coding. AI will eventually need to give in and spend its billions on human beings to refine every word, phrase, and idea it generates.

They think they will get masses of ignorant paeans to do the refining for cheap while they bank the cash for hard assets being bubbled now.

VP_Spiro_T_Cheney wrote:

To date the various players have spent roughly two trillion dollars on “compute” as Sam Altman puts it… and what they can charge for leasing out that compute doesn’t even pay the operating costs, let alone provide ROI on that two trillion dollar investment.

And the killer data point for Oracle is the fact that the operating costs rose faster than the revenues.

Even worse, unlike the railroad or fiber optic buildouts, most the assets they bought with those two trillion dollars have a useful life of only 2-4 years.

LLM AI is a bubble, what remains to be seen is who’s the sad fuqer that end up as bagholder.

None of that matters at all.

AI is very close to being smarter than every single person on the planet. It will do a 10x in the next 5 years (at a minimum). It is already able to design engineering projects in seconds, correctly, that would take a team of seasoned engineers years to do. It feels like magic working with it most of the time.

People are complacent as to what this means to humanity and the world. "It draws 6 fingers". Yeah, that was 2 years ago. "It hallucinates". That was last year. "It can't operate in the real world". So many humanoid robots are on their way, I can't even count them. They are stronger than any human, don't sleep, don't get sick, don't die and only get better over time.

Oralce is investing in what it knows will take over the world. Thinking about short term profit is like trying to push the ocean back with your hands. Totally pointless.

The main tech firms know what is coming. They need to have the resources for it.

Everything else will be completely worthless. What is the value of anything if it can be made for nothing by robots including the mining, design, manufacturing, etc.

wrote:

the only reason AI isn't learning everything on its own, without humans, is because it has literally no way to do that because they are training it not with the real world but with information about the real world.

the next step is robotics where AI will interact directly with the world just like we do. it will be able to understand physics by its own movements no by reading about it or watching videos. it will be able to physically upgrade itself and as it does, it will very quickly be better in every way than every human.

that's the next step, like it or not. it is inevitable at this point.

If I told you 5 years ago that would be able to write any kind of document, in any language and create any kind of software program, in any software programming language, be able to diagnose medical issues, create videos that are so realistic that people can't tell if they are real or not, create music that is better than any band could make in seconds, create art and pictures, you simply wouldn't believe it. It can talk in any language and translate into any language.

It can look at a video and tell you what it sees and what it thinks will happen. It can do the same with images and audio. All that and it doesn't even have a physical presence yet.

Everyone not on the bandwagon is going to be left behind. Anyone not using it is already being left behind. They are totally useless because their productivity is zero compared with someone using AI.

nah. I envisioned a lot of that stuff   Asimov envisioned it 70 years ago. The bubble is in the idea that we need huge data centers for AI, and the resultant energy burn to operate them.

We don't. We don't need a deep thought that tells us 42 is the answer based on the collective intelligence of the internet.