Cooled EGR is not an either-or to DEF. You can have both.
That's even more weight, space, and complexity that no-one wants. It's hard enough fitting the EGR cooler for 4,500hp loco in the loading gauge. And the reason they went with EGR in the first place is because they don't want to deal with the logistics of supplying loco fleets with DEF. Using EGR and DEF would completely defeat the purpose.
I was saying that the local data centers don't affect the residents. They use closed-loop cooling and they are sited properly. What part didn't YOU understand?
And closed loop cooling uses zero water. Oh it uses less water, not zero water. What about power? Does datacenters use closed-loop power? That does not exist?
Show evidence.
Truckee, California. That datacenter did not build their own power. They are just buying all of Truckee's power. Screw the locals.
DEF systems on heavy vehicles work, but they're fairly, well, heavy. Among other thing, they use electrical heat to get up to operating temperature. Then there's the issue of needing the fluid. For earthmoving equipment and railway locos, they'd rather not deal with that and have gone with complex EGR systems with liquid cooling instead.
Nice joke for people who remember Rolls Royce cars before they became BMWs, but I always remember it being "sufficient", not "adequate".
You're missing the joke - Rolls Royce cars (before the brand was sold to BMW) never used to quote power output in marketing material, they just put "sufficient" where a figure would usually go.
The cost though... They did it thinking it would be cheap, and it turned out to be the opposite.
I'm not sure this is really true. They sold it to the public as being cheap, but de Gaulle was absolutely adamant that France needed independent nuclear deterrent capability, because he didn't fully trust the USA. A domestic nuclear industry was a prerequisite for that. So they were probably lying about the cost to make it more palatable to the public all along.
The economics for reprocessing only work if you put an arbitrarily high value on plutonium. In the early days, when countries were building up nuclear weapons stockpiles, this sort of made sense. The nuclear weapons organisations bought plutonium from fuel reprocessing for high prices, and the UK was even selling plutonium to the USA. But now there's more plutonium floating around than anyone really needs, so reprocessing isn't cost-effective. But you don't want to completely dismantle the capability in case you need it for a future arms race, or for maintaining your stockpile.
And then they screwed up their environmental goals by promoting diesel cars, again not knowing that they were so bad when the decision was made.
Diesel cars are still better than petrol cars on every metric besides NOx. The focus was on reducing CO, unburnt hydrocarbons, and then CO2. Diesels are better on all those counts. But NOx eventually became a huge problem, particularly with the way VW, BMW, etc. cheated on emissions.
Apple does not manufacture their own CPUs. They contract TSMC to make them. Apple also does not manufacture their own SSDs, cameras, etc. Of all the components, RAM is a commodity component that survives on thin margins relying on high volume for profits.
Building a chip fab would take years if Apple had the site, the personnel, plans, permits, etc. today. Then it takes an experienced foundry like TSMC about a year after construction is complete to start making enough acceptable product in volume. So 3 or 4 years from now, Apple might have Apple RAM for their devices. By that point, if the RAM crisis is over, all the existing companies could sell their RAM for less than it costs Apple to make as Apple has to recoup capital costs. In the end, Apple will lose money. For what?
Almost every time a company makes promises that seem too good to be true, it turns out too good to be true. Sometimes they even get prosecuted for it. Pivoting from AI slop factory to medical imaging, where being correct actually matters, doesn't inspire confidence.
You don't buy a RAM company, you start one.
How long do you think it would take to "start a RAM company"? If Apple had the personnel, site, plans, equipment, etc, it would take years for them to build the plant. Then the plant does not make 100% sellable product on day 1. That might take months to a year. So 3 years from now, Apple might, maybe have a few chips they could use.
The existing companies refuse to expand to meet demand, which is the whole reason for this mess.
Um no. They existing companies are being lots of money to make specialized memory for AI. They are meeting demand. They are meeting demand of people who are paying them the most. They are not meeting the demand of us peasants who can't afford to throw money at them.
And spending $1T for memory is crazy. Instead like the MP said, "They could build a memory fab of their own from petty cash if they actually wanted to... "
Building one would take years if Apple had the personnel, the expertise, site, etc. And at the end of the it, Apple built a manufacturing plant that does not fit into their strategies. After all, Apple does not manufacture their own CPUs. They contract TSMC to make them.
If they can develop their own processors, they can certainly do their own memory!
The problem with memory is manufacturing them at the lowest costs not developing it. Memory is standardized and considered a commodity due to the huge number of patents surrounding it. The problem right now is the RAM manufacturers are getting lots of money not to make consumer grade RAM like DDR5. They are getting lots of money to make HBM for AI servers. Where would Apple make this RAM? Certainly not at Micron, SK Hynix, or Samsung. TSMC could manufacture if they were not fully booked making AI CPUs, AMD CPUs, Apple CPUs, Intel CPUs, NVidia GPUs, AMD GPUs, etc.
Sell a few $100B in bonds, issue new stock, talk with a few bankers,
This isn't a small business loan. The number of banks that could lend hundreds of billions is very small.
. Antitrust laws (not just in the US) might prevent an acquisition but not the financing.
There are other obstacles. Samsung's RAM business is a part of their chip business. They would have to separate it from their foundry business. It is highly unlikely Samsung would ever agree to that. Samsung and SK Hynix are South Korean companies. I would think the country of South Korea would object to the sale. That leaves Micron which is in the US. That's where Dell, HP, Lenovo, Asus, etc would all sue to keep Apple from buying a key supplier.
That said, they probably don't want to. A large acquisition can be very, very distracting to management and is usually a new loss over time.
Also RAM was a commodity industry before the current crisis: Cutthroat competition that survives on thinnest margins and highest volume. That is not the business model Apple wants to invest in. That's why they (like every computer manufacturer) have bought their RAM instead of contract manufacturing it like their other chips.
You are lost in the Swamps of Despair.