NASA still needs to explain why. Never in history did we do science or major projects like this for no reason. Even Neil deGrasse Tyson talks about this; the original effort to go to the moon had geopolitical context of surpassing the Soviet Union and the dual-purpose military use of rockets-turned ICBMs. Columbus sailed around the world not to prove it was round, it was to cut out the Middle-Eastern middle-men in the trade of Asian goods in Europe; it was about money and geopolitics. Every major advancement in history has either an economic driver or a geopolitical driver. With that, the Moon, Mars, and the whole bit are just too far away, too costly, too dangerous, and don't generate enough economic or geopolitical benefit.

Or, if they do, they aren't articulating it well. Without that driving purpose, this moonbase or Lunar Gateway or whatever just simply won't happen.

It is not illegal to announce intended plans. The articles themselves say it will require approval by Congress.

I know it's knee-jerk to hate on the executive branch and Trump, but using terms like "illegal" becomes hyperbolic and meaningless if its overused and used in a factually incorrect way.

It's not ASML that'll be the problem. They'll buy a system and it'll be there in crates ready to be isntalled long before:

A) the permits are in place. This is Texas, so maybe "faster than California", but it'll cost.

B) The water and power is secured; again it might go faster but it's 18-24 months. However,

C) This is Texas, the power grid has been unreliable. Since foundries run 24/7 fully automated to produce the kinds of chips he's talking about, they'll need reliable backup power. Backup generator lead times are, as of today, around 48-90 months. This is very hard to go faster, as these gensets have specific bottlenecks, namely the castings for the crankshafts are typically done at specialized facilities, cannot be made to go faster, and have slots secured years in advance. Maybe you pay through the nose to buy someone else's gen set, or you're robbing them off of old recycled ships and turning them into gen sets, but they'll be pieces of junk; a real risk. I bet the EUV system will be ready years before the power requirements and backup gensets are ready.

D) labor - thanks to Elon's MAGA tendencies, we're now starting to face a labor shortage. We have a major skillset mismatch in that the kind of people to operate these tools typically aren't trained in America, so to go fast you'll have to, shocker, bring them over via H1-Bs. Heck even construction to build these fabs is facing labor shortages, particularly in places on the Southern border like Texas because construction is heavily staffed with Hispanic people, who are getting harder to find today thanks to Trump and ICE.

So yeah, odds are this is going to cost 2-5X as much as it should and take 2-5X longer than it needs to, and probably still won't be able to operate when "done".

I remember when Tesla was a capital P Progressive brand. It's stunning how much the world has turned.

I agree with you wholeheartedly, however if you're in Uber's position you can't afford to look at the now. The fact is, they attempted to build their own robotaxi and failed, which left them in the position of being a ride-hailer app only. That makes them vulnerable if there is only one player to provide automated taxi services. And so far, Waymo is really the only successful player in this. Tesla on the other hand, as much as I dislike the company has had naysayers against it for many years and while it never beats those naysayers when they say they will, the fact is that A) they are profitable, B) they have a large chunk of cash, and C) Elon has made many investors a lot of money, so if he asks for more investment he'll get it, and he is directly talking about muscling in on Uber's territory.

So while I think Musk is a clown, you can't discount what he can achieve when he stays focused. Right now he's vulnerable because the political focus made him distracted, which gives Uber time to support more players coming into the market. On top of that, if they make 12 bets on alternatives to Waymo or Tesla and only 3 or 4 of them make it, those 3 or 4 who are successful will be much smaller; operating through Uber's platform will give them revenue and allow them to survive, and that gives Uber control over them the way they'll never have over Waymo or Tesla.

So it's the smart move on their part.

There are 2 up for election this year, and 3 in 2028. That's 2 years, not six. Justice seats go up for election on even numbered years. In 2 years you could have 4 to 3 Democrat majority in the Ohio Supreme Court.

I think you're overplaying it a bit. I do see that the Republican crowd has elected a 6-to-1 Republican/Democrat set of justices in the Ohio court. But frankly if the Ohio Supreme Court overrules something like this, then A) Ohio got what it voted for, and B) it would directly fly in the face of the MAGA crowd who elected those guys. Ohio state supreme court justices do not serve for life like the US Supreme Court, they ahve 6 year terms and mandatory retirement at age 70. So if they clearly ignore a Constitutional Amendment that allows tech bros to come in and go against the wishes of rural AMerica, I think you'd see a MAGA backlash in the next election cycle. Some lessons have to be learned the hard way, and if that happens then Ohio got an expensive lesson in hard-right populist propaganda.

So this question is exactly the question people should be asking; its perhaps the most important question and few people are asking it.

You're exactly right to be concerned, if this was OpenAI or Anthropic or xAI. In essence these companies are announcing multi-billion dollar deals to build data centers that can't be built, to buy GPUs in a volume that can barely be made, requiring infrastructure that hasn't been built, to build compute power that may or may not be needed, for a market that hasn't yet been proven.

In OpenAI, xAI, or Anthropic's case, they need to build big data centers. There's a general consensus that bigger compute is needed for more accurate model training. I saw a good analogy the other day; in essence LLMs work on relative location to other words in language. For example, to know any point on Earth, you need two relative data sets - latitude and longitude. Throw in altitude and you get a pretty darn accurate map of the Earth. LLMs use that similar logic, but with tens of thousands of data points per word. To use the mapping analogy, the finer you slice latitude and longitude, the more inaccurate it's going to get. So for example, GPS can find your phone pretty accurately with the granularity we use today, but if it had to catalog every grain of sand with latitude and longitude, it's going to make mistakes as that is too granular.

So if you build a much larger data center, and go from tens of thousands of data points per word to millions, then the theory is you get far more granularity and can improve accuracy. There are some disputing this, saying that the data math doesn't work out that way, but it's the logic behind big data centers.

The startups like OpenAI, Anthropic, and xAI, don't have cash to pay this. So they do this in exchange for equity, or in deals for equipment. So for example, nVidia invests $10B in OpenAI, but it doesn't give cash, it gives GPUs. Say they're priced at $10,000, so it's giving OpenAI 1,000,000 GPUs for a $10B investment. The trick is it probably cost nVidia $2,000 to make each GPU, and if OpenAI goes public at a higher valuation than nVidia's "investment", let's say 3X, then nVidia gets to recognize a "sale" of $10B on $2B in cost and $8B in gross margn, marks the equity on it's balance sheet, then sells it for say $30B, recognizing a $20B gain. nVidia's got the cash, so it can weather this, but if OpenAI crashes it'll drop nVidia's stock and create a major write-down. This is the big concern, because if there is a drop in value and OpenAI can't pull off it's lofty aspirations, it'll bring down the value of many entities at once. ANd i would say that's a higher probability than we would like, because OpenAI has yet to prove a model where it can sell a product with a positive gross margin, it's conversion rate from it's free to paid is atrocious, and it's operating costs are enormous.

So the startups are a serious concern because of how they're financing these things. It's all equity transactions and "future promises" which create "accounting" revenue and value, but not real cash.

Now let's get to this story. Meta, Microsoft, Alphabet, these guys are different. They are building the data centers, but the difference is, they have other products which print money. While this is $27B contract over 5 years, Meta has $87B in cash on the books right now; they can pay for it. And notably it's not for building data centers, this is for contracting time on other people's data centers; Meta is making it other people's problem to build, own and operate the data centers, it's focused on training it's model. I think that's important, because data centers take about 5-7 years to build (the backup gensets for example are 90 month lead time), by that time others will have come out with dedicated AI chips, nVidia GPUs won't be the only chip on the market ideal for LLMs. So they won't be saddled with legacy server hardware, they can just swithch to someone with the latest and greatest and let infrastructure be someone else's problem. Also, Meta doesn't have to scrape data from the internet (like how Encyclopedia Britannica just sued OpenAI for wholesale copywrite infringement). it has Facebook and Instagram to do that.

But the key is, this is clearly an expensive business to be in. Meta, Alphabet, Microsoft; they all have other revenue streams to support this spend. If they fail at AI it'll hurt their stock price, maybe some people get fired, but the company moves on. Not so with OpenAI, Anthropic, and xAI and some others; this is all they have. And if it's a war of attrition on how much money can be spent to who dominates this market, the legacy FAANG companies can finance on revenue, not equity; they will win that every day of the week and twice on Sunday.

But to fully answer your question, it is insane the size of these deals, but given it's Meta, it'll come from their balance sheet; Meta's got the cash on hand to do it.

When I was in college, we had hundreds of CS majors. Many wanted to go into video game development. And that was just one university; hundreds of universities were churning out gamer nerd programmers every year, making tens of thousands of potential new game devs.

Labor is a market. If you're part of the oversupply, you'll be taken advantage of. I'm not making a moral judgement about it, it's simply the way of things.

They're not going after him for lying on his resume. Your resume is simply your profile, and investors readily assume what it says is who you are; his LinkedIn says the same thing.

They're going after him for fraud on stock sales, embezzlement (that's the misallocating expenses to the company thing), and taking his emails and such. The resume thing is most likely to establish a personality profile in their complaint that would make the actual charges seem more plausible; ie he lied about a PhD, so yes his personality and behavior suggest he would commit similar fraudulent actions.

I think the short time frame says more about nVidia than anything else. nVidia happened to be in the right place at the right time; they were the only ones doing multi-threading because that's necessary for GPUs, with are for graphics. No one else bothered because it was seen as a niche market. Then blockchain, and later LLMs, burst on the scene and it turns out multi-threading is pretty darn useful over traditional processor logic; suddenly nVidia who specialized in niche stuff is the hottest thing in town.

But that also means they have to make a very careful dance. A) they can't just give to one AI company or the other; they're better served being neutral and encouraging the rapid growth and pouring in of billions of dollars as it'll come back to them. So quickly moving away from being seen as too close to OpenAI over the other companies makes sense. But also B) multi-threading isn't proprietary, and GPUs are just what's useful and available. In less than 5 years there will be dedicated AI-PUs, not GPUs, designed to manage heat and power better and multi-threading better, and nVidia won't be the only game in town. So they have a delicate balance to keep themselves seen as the market leader here, and given how the AI "leader" switches so quickly, it doesn't make sense for them to be tied to any one company.

Actually, he probably wants cash.

They make things; they're not investors. Often these kinds of "investments" from companies like NVIDIA are not cash injections into companies, they're in-kind. So if their chips and GPUs are priced at let's say $10,000, and they make a $1B "investment", they're really giving their investment 10,000 GPUs, not $1B in cash. But note, they're priced at $10,000; it costs NVIDIA likely $2,000 to make and the rest is overhead and marketing. So an "investment" in OpenAI comes back to them in the form of sales, that are non-cash sales but still accounting sales.

Now they go public, and let's say their investment nets $3B on the $1B. That's essentially like making $30,000 on a $10,000 product that cost them $2,000 to make; a real windfall.

To be fair, there is a long tradition that predates Trump in the DoD if using the work performing system for defense contracting.

Yep. Just what we need: an AI capable of making kill decisions for autonomous weapons that is deep down deeply racist, fascist, and used primarily for NSFW images.

America!

And that's not a great thing. I worked 6 years for one of the primes, and it's clear how they operate. Anthropic's problems would go away if the leadership starts acting like Lockheed; kowtow to the DoD officials by calling them DoW, and say "yes sir, whatever you want sir". Hire a bunch of retiring O4s to O6s who aren't going to get their stars and give them cushy mid-management jobs, and then stack on top margin after margin while delaying projects left and right.

It's funny because having worked at a Prime before and seeing this very thing happen, I also don't believe very heavily in the military industrial complex. The concept was that industry would push the military to war because sales were driven by weapons usage, but that never really materialized. Rather, it was a welfare state. The DoD is a terrible customer, buying things in fits and starts, changing requirements in the middle of a program, and squandering R&D budgets on pet projects and nonsense. Meanwhile the contracting officers are too lazy to go direct to a Tier-2 or Tier-3 supplier for an interesting idea, and instead farm it out to a Prime who just subcontracts to the Tier 2 and puts an overhead fee on it. Meanwhile the warrant officers for a given technology are reluctant to change anything without 10X the proof of capability and safety studies than would be normal, meaning half of our military's subsystems are so legacy compared to what's available commercially that our military is eminently hack-proof because no modern hacker knows how to hack an abacus and a hamster wheel in code written in ancient Egyptian that is the backbone of many sub-systems.

The Primes on the other hand have regular bills to pay and workforces to maintain and see this insane way of doing business that the Pentagon does, and adapts to it milking as much overhead as possible so they can level out their monthly payroll expenses without too much labor disruption. All the while Congress has no idea what to do to fix the issue, so they impose restrictions on government employees where they have to report even a lunch meeting over $25 or get investigated while we squander billions with bad bureaucrat managers in DoD.

The one thing the Trump Admin is doing somewhat right is targeting this exact issue; somewhat right in that it's an issue that needs solving so they got that right, but wrong in that I don't think they know how to fix it.

Sorry for the rant. I loved my time working at a Prime and I still cherish it, with some great colleagues that I still keep in touch with. But once I started seeing how it all worked, I was just stunned with the ridiculousness of it all. And now we see effectively mid-tier at best contracting bureaucrats trying to manage something as fast moving as AI with all the subtlety of the Titanic, and with likely similar outcomes. And what's sad is that the DoD should cave to Anthropic. Claude is good, and the military does lots of things that don't involve weapons; it has the world's most complex logistics chain, a huge healthcare system, major R&D programs, huge humanitarian programs, it led to the development of game theory, it has (or had until Hegseth) one of the world's best leadership training programs; all of those aspects of the military could benefit, and if they really want killer-AI weapons, as bad as that is, I'm sure Musk will sell it to them with Grok. It's painful to watch what could be amazing utilization of AI become a giant s-show.