Are Tech Stocks Rolling Over?

[Zathras]

How are they pushing the crash cart up to service the 10,000 Starlink satellites in orbit now?

Elon did a video a while ago explaining how the "AI1" satellites will work. He said that the AI1 is basically the same tech as Starlink, except actually easier to do. AI1 will have a larger solar panel facing the sun than Starlinkj and will have a radiative cooling panel knife edge 90 degrees to the solar panel. They are easier because they do not need all the complicarted radio antennas that a Starlinkj has, just about one rack's worth of NVDIA GPU's and some laser comm links to connect to the Starlink network.

There will be a large number of small nodes, so when one satellite fails it will just be deorbited and replaced with another. If they weigh 2000 pounds, that means one Starship with 100 ton payload capacity could easily launch them, probably limited by the "pez dispenser" that can hold up to 60 satellites.

I would suggest the Starlink satellites don’t require cutting edge chips that a data center will.
Between the replacement rate of 1-3 years and the added damage of space based radiation. SpaceX would be replacing the entire group every 2-3 years.
Cutting edge data centers run 100,000+ GPUs. SpaceX, would need 1500 satellites to equal that.
In addition, by deorbiting the satellites there is zero recycling of the old chips/cabling or other materials.

So every 3 years, SpaceX is building and launching the entire structure.
Land based companies are replacing chips.

It is fun to think about, but I don’t see how this works.

 

[Markola]

Makers of chips are booming while the hyperscalers that buy them are tumbling.

“Bank of America frames it starkly. Hyperscaler capex has climbed from 70% of operating cash flow in 2025 to nearly 100% in 2026.”

Microsoft’s Worst Month Since 2000: Why Is This Happening?

Shares of Microsoft Corp. are down over 20% in June, on pace for the steepest monthly drop since December 2000. Twelve months ago, the Redmond, Washington-based company’s market cap hovered around $4 trillion. Today, it’s at $2.65 trillion, behind Nvidia...

finance.yahoo.com

 

[surprising]

Makers of chips are booming while the hyperscalers that buy them are tumbling.

“Bank of America frames it starkly. Hyperscaler capex has climbed from 70% of operating cash flow in 2025 to nearly 100% in 2026.”

Microsoft’s Worst Month Since 2000: Why Is This Happening?

Shares of Microsoft Corp. are down over 20% in June, on pace for the steepest monthly drop since December 2000. Twelve months ago, the Redmond, Washington-based company’s market cap hovered around $4 trillion. Today, it’s at $2.65 trillion, behind Nvidia...

finance.yahoo.com

Interesting times. Nearly every company that can design a semiconductor is trying to dethrone Nvidia with their own custom chips. It's pretty crazy because the barrier to entry used to be pretty high to develop a chip but what's happening is a chip for inferencing, which runs the models, is relatively simple. It's a bunch of matrix multiply units and a ton of memory. So the space is becoming so fractured and insane. All these companies are getting billions to build the same thing. Not all of them will survive.

 

[Al18]

Not all AI chip companies are performing the same. NVDA is only up just under 5% YTD. By comparison, ARM is up 214% YTD

 

[surprising]

AI related stocks have been on quite a roller coaster lately, but so far it hasn't spilled over to a general market selloff. It seems to be a battle between institutional investors selling AI and retail investors buying the dip.

Meta recently said they are exploring renting out unused datacenter capacity. SpaceX is already doing the same. Implication is there isn't enough work to keep the hyperscalers busy, so perhaps they over-built.

People say this is nothing like dot.com because all these hyperscalers are massively profitable. Well, as far as I recall, the Ciscos of the world back in dot.com era were also highly profitable and also over-building the internet infrastructure. It sure seems like the same setup for a pretty big crash, BUT perhaps retail investors are going to keep the party going or even prevent a larger crash with steadfast buy-the-dip.

 

[oiseux]

Between the replacement rate of 1-3 years and the added damage of space based radiation. SpaceX would be replacing the entire group every 2-3 years.
Cutting edge data centers run 100,000+ GPUs. SpaceX, would need 1500 satellites to equal that.
In addition, by deorbiting the satellites there is zero recycling of the old chips/cabling or other materials.

They are constructing a modular system so they can swap out the electronics, because they will need an accelerated life cycle. The satellites last a bit longer. So they are proposing a system of maintenance robots (which will have their own similar issues with space, even if tethered).
There are physical damage issues from space but also a massive increase in memory soft errors and hard errors (due to displacement). Improved shielding technology needs to be developed, and the systems may require extensive system redundancy in addition to other mitigations.
I think needless to say it will take longer than expected.

Nearly every company that can design a semiconductor is trying to dethrone Nvidia with their own custom chips. It's pretty crazy because the barrier to entry used to be pretty high to develop a chip but what's happening is a chip for inferencing, which runs the models, is relatively simple. It's a bunch of matrix multiply units and a ton of memory.

Yes, the barriers to chip design have lowered. Custom primary compute chips (SOCs) have proliferated across industries, aided by foundries, IP houses and partners. The goal often is to produce a total architecture that is more tailored to their needs and can give a company an edge in the market (a specialized tool vs. a swiss army knife). The next expanding wave are modules and chiplet based designs. The very large SOCs, unlike the ASICs of the past, have a high economic hurdle, costing hundreds of millions to develop on a leading edge technology node. Countering that, there is also a tool-chain and software infrastructure advantage of the general purpose incumbants.

Meta recently said they are exploring renting out unused datacenter capacity. SpaceX is already doing the same. Implication is there isn't enough work to keep the hyperscalers busy, so perhaps they over-built.

No, these two players have looking to leverage their internal infrastructure and core competency in that area to expand into an adjacent market. This is typical when a player achieves market penetration that limits their future growth, or begins to stall. Usually the challenge though is scaling a support, service and sales structure for something so orthogonal to your main business.

 

[Markola]

It’s a jittery tech market, and no one is talking about potential IPOs for Anthropic and OpenAI anymore, that I’m hearing anyway:

Today, SpaceX closed below the $150 initial trading price last month.
Microsoft is down nearly 18% year to date
Tesla down 8% ytd
Palantir (probably recording our keystrokes) down 20% ytd
Meta down 5% ytd

Apple, Amazon, Google, Nvidia in low single digits to mid-teens positive ytd.

Chipmakers are up strongly ytd, though Micron and Intel are coming back to earth. Samsung posted blockbuster record profits today, yet the stock fell 6%.

I’m way down on another tech stock.

Bah humbug? Summer swoon?

 

[Markola]

Wow, SPCX is around $138.

 

[Fermion]

Wow, SPCX is around $138.

I got my one share purchased on first day at $167! My next buy is at $30.

 

[Whakamole]

Microsoft is down nearly 18% year to date

Microsoft is also poorly managed.

 

[joesxm3]

Wow, SPCX is around $138.

Whoops. I bought another 50 shares this morning.

I had hurt my wrist over the weekend so I can't really do anything. I think that had something to do with me not getting out of bed until 11am this morning.

Last week when SPCX was around $145 and I was assuming my $150 PUT would get assigned (break even $144.70) I put in a limit order for $40.27 which I figured would never be reached in case SPCX bounced and my PUT did not get assigned.

As I write this I realize what I am writing as my reason does not make sense. But that is the way it is

I hope the Starship test happens on schedule this Thursday and is a roaring success, but $12 is a long way to go to save my PUT.

 

[jim584672]

SPCX sucked all the energy out of the market.

 

[NW-Bound]

I got my one share purchased on first day at $167! My next buy is at $30.

You made me look. Jan 2027 put at 30 is $0.20 bid, $0.30 ask. Yet, there are 666 contracts outstanding. Are those yours?

 

[NW-Bound]

SPCX sucked all the energy out of the market.

Someone said "energy"?

My oil and gas stocks went up good today. I just don't have enough of them to cancel out my other pummelled stocks.

 

[RetiredHappy]

On a day like today, I usually don't even log into my account to see what it looks like.

 

[Fermion]

You made me look. Jan 2027 put at 30 is $0.20 bid, $0.30 ask. Yet, there are 666 contracts outstanding. Are those yours?

Probably some morbid person is making a play that Elon OD's on Ketamine or something, which would be the only thing that could drive the price down that low that fast.

 

[RetiredHappy]

Probably some morbid person is making a play that Elon OD's on Ketamine or something, which would be the only thing that could drive the price down that low that fast.

Good point. If Elon falls sick or dies for whatever reason, both Tesla and SpaceX stock prices will collapse. I don't keep up with the news to see if he has capable COOs and visionary leaders in both companies that can keep it going. I don't have that much invested in both companies to lose sleep over it.

 

[NW-Bound]

Probably some morbid person is making a play that Elon OD's on Ketamine or something, which would be the only thing that could drive the price down that low that fast.

Using the lower bid price of $0.20, I compute the premium of these 666 contracts to be $13,320. The cash to secure the puts is almost $2M. That's 0.7% gain in 6 months. Not a great return.

Nah. You can do naked puts, but will your broker allow it?

 

[Fermion]

Using the lower bid price of $0.20, I compute the premium of these 666 contracts to be $13,320. The cash to secure the puts is almost $2M. That's 0.7% gain in 6 months. Not a great return.

Nah. You can do naked puts, but will your broker allow it?

Surely you have looked at spreads before right? Say the $25 puts were $0.10. You could then sell the $30 puts and buy the $25 puts for a net credit of $0.10. 666 contracts would net you $6,660. You would need to have available $330,000 but it gets the return up to 2% in 6 months.

 

[NW-Bound]

Surely you have looked at spreads before right? Say the $25 puts were $0.10. You could then sell the $30 puts and buy the $25 puts for a net credit of $0.10. 666 contracts would net you $6,660. You would need to have available $330,000 but it gets the return up to 2% in 6 months.

I have looked at spreads once or twice, but decided that it was not for me. I have never done spreads or iron condor, or butterfly, etc... Only straight calls and puts, and 1 week out. I make good enough money with these so far, and have not felt the need to do other things.

In your example, 4% annual return is not good enough for me. The risk is indeed quite low, but I would want a bit more return though.

 

[Fermion]

I have looked at spreads once or twice, but decided that it was not for me. I have never done spreads or iron condor, or butterfly, etc... Only straight calls and puts, and 1 week out. I make good enough money with these so far, and have not felt the need to do other things.

In your example, 4% annual return is not good enough for me. The risk is indeed quite low, but I would want a bit more return though.

Yes in that example it would be a silly risk but if you could get the return up to 15% in 6 months on a put spread that might be worth it for some stocks.

 

[Markola]

Ooof. SpaceX is trading below the IPO price.

 

[Diogenes]

Good point. If Elon falls sick or dies for whatever reason, both Tesla and SpaceX stock prices will collapse. ...

Disagree. Once a high-flying company matures, the pseudo-founder is generally a detriment. That is particularly the case for flamboyant, celebrity CEOs who care more about public-impact than on running their company. I say "pseudo" because at least in the case of Tesla, there was already a promising company; Elon came-in, in what amounts to a hostile takeover. Elon's departure would allow the actual engineers to take over. Then maybe Tesla could resume innovating and actually build things like the Tesla Roadster, instead of strumming the AI-and-robotics banjo.

 

[GrayHare]

Ooof. SpaceX is trading below the IPO price.

Not uncommon with IPOs, of course. There will be a buying opportunity before long.

 

[FD1000]

Why bother with single companies? Just compare VOO, QQQ(tech), SMH(chips) for 1-5-20 years.