400k Investment Project (@400kProject) header
400k Investment Project (@400kProject) profile picture

Investor time-stamping my portfolio moves on Savvytrader. Falcon Heavy up +1014%, Falcon 9 up +501%, Stairway to Heaven up +462%.

Joined January 2023
Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)Photo by 400k Investment Project (@400kProject)
400k Investment Project retweeted
I still find it totally wild that the chief scientist of OpenAI posted this on the company blog, and we all just moved on to the next day's news
35
140
14
1,083
151,247
$IREN spooky… $4B debt on $4B ARR… yikes 🤦‍♂️ Ask the US government how they’d feel about debt = 1yrs tax revenue
1
203
So to build 1 TW compute: $3-6 Trillion terrestrial compute $NBIS $IREN $20-70 Trillion orbital compute including radiators and launch cost $SPCX See previous thread for details…
1
3
346
And remember we only have 15-20GW AI compute online right now in North America. Low hanging fruit is not orbital…
76
So let me get this straight, ignoring the hardware costs (GPUs, racks, solar panels, radiators) starship (not Falcon 9) would need to do 200,000 trips to space to get 1 TW of compute up there?? So the estimated… best case launch costs are around $2 trillion excluding the cost of the orbital data centre hardware. And that’s cheaper than building on earth? Grok: “Grand total: about twenty one million tonnes to orbit. Falcon 9 gets you roughly a million launches. Starship, a hundred tonnes a pop, still means over two hundred thousand flights” $SPCX vs $NBIS $IREN etc…
2
1
2
253
So when you factor in the enormous radiators required for just one rack in space you’re looking at a base cost for the hardware of absolute minimum 2X the cost of terrestrial compute before the $2T minimum cost of 200,000 launches and the time and logistics of getting them there and of course hoping there isn’t a massive chain reaction in orbit destroying everything floating around up there ala “Kessler Syndrome” $SPCX 🧐🧐
1
76
And sorry just to be clear. - $2T is best case launch costs. Currently $50-100M per launch of starship (est) so $10-20T in launch costs until prices come down… a lot.
59
Hey @elonmusk what’s the plan here? $SPCX $TSLA $NBIS $IREN $CRWV “Thermal engineer here. I have done a lot of work on thermal management of orbital systems, and I just wanted to chime in and say that you are absolutely right. When people started talking about orbital data centers my first thought was that there is no way they could cool them without an absurd number of radiators. I did some simple hand calculations similar to yours and came to the same conclusions you did. My numbers were a little different (solar panel efficiency on typical orbiting panels is more like 40% but solar panels also absorb a lot of heat from the sun without converting it to electricity and the comparatively warm Earth partially blocks radiator view to cold deep space). It is common to use fluid loops to carry heat away from electronics but the fluid loops must dump that heat somewhere and that place is typically the radiators. Ultimately all heat/energy that goes in at the solar panels must be radiated away as you state. Orbiting data centers are not impossible, but they are very, VERY impractical. The Stefan-Boltzmann constant puts an upper limit on how much heat you can radiate away from a given radiator area and there is no getting around that” youtu.be/S1umCJfbJ0w?si=3TF-…
2
460
Been doing a lot of searching around for causes for the BTC sell off this weekend. Looks like stop hunting at the point where liquidity is lowest + geopolitical uncertainty around Iran, hacks, FAA shut down etc. bid was thin. Seems most likely. Similar story for equities on Liquid - though possibly just selling off in sympathy to risk proxy BTC.
1
211
Appreciate the details you were trying to give in the spaces. @jimpowers100 can you post the thesis you were trying to get across re: photonics/800v architecture changes etc. I'll repost. Hosts are brutal in these spaces. Any space with Jaguar and Joe Schmuc etc. is brutal to listen to.
141
Basic math, Elon claims 1-1.5yr payback period for a GW of compute @$50B cost @Pregory1 @mg10012
1
1
179
And yes, most is not rented at spot rates typically. However, spot prices are unlikely to collapse due to all of the uses of intelligence - think humanoids, world models, science etc. and power remains the constraint. Not anyone can stand up a data centre even if they want.
57
Truth. Bullish.
"A rate hike is no longer certain. We didn’t foresee Jensen having multiple side chicks."
180
AI is getting out of control. Look at what you’ve enabled Jensen.
If you’re going to pop the AI bubble, do it with style baby
1
489
This market is rather annoying. Pump.. fade, pump, fade. Iran, Jensen, Jensen…
1
1
379
400k Investment Project retweeted
$EOSE Joe alluded on multiple occasions to 3.5GWh of near term opportunities. Specifically referring to the next 4-5weeks. “We're currently working 3.5 gigawatt hours of late-stage opportunity more than today's entire backlog. More than half is waiting on financing, the rest is in contract negotiations” Joe on 27 “We've got near-term deals here that we'll know a lot more on here in the next 4 to 5 weeks that are going to determine what that looks like.” These ops are Currently in the financing stage and contract negotiations. Now I don’t know exactly what these projects are. Whether they’re FPUSA related, independent developers or state related. But I am going to assume that they’re serious ops. This would mean they are obviously happy with the tech and the data they have seen. Otherwise they would be at the financing stage or in negotiations. I believe that delays in these ops are also the reason line 1 was moved earlier than previously thought. Another point is that the spotlight is definitely on Joe. Given the leadership change at $flnc it’s clear that C suite can be on the chopping block for underperforming. Given the history of setting expectations and not meeting them I find it interesting that Joe would kind of push this kind of rhetoric a few times within the call. It only sets investors up for disappointment if nothing eventuates. No doubt this is another significant piece of the puzzle and I don't want to be entering 2027 with ~$650m in backlog.
1
1
42
2,175
Wouldn’t $SPCX SpaceXSI be most likely to be impacted by permitting hold ups etc with massive data centre expansion proposed for 2027/2028? Blue leadership doesn’t care about a random data centre providing high paying blue collar jobs out in West Texas… $CIFR, $IREN projects for example
1
453