Impulse Space

[FutureSpaceTourist]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

[TrevorMonty]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

Tom thinks cloud servers will move to space as they are so power hungry.

He also developed on NG's TR107 1,1Mlbs LOX/RP1 engine which unfortunately never found a LV. 

Sent from my SM-T733 using Tapatalk

[FutureSpaceTourist]

https://www.cnbc.com/2023/06/27/spacex-vet-tom-mueller-impulse-space-target-service-economy-in-orbit.html

SpaceX veteran Tom Mueller targets space service economy with tug business
PUBLISHED TUE, JUN 27 202312:54 PM EDT
Simona Riccardi
@IN/SRICCARDI/
@SIMONA__G
Morgan Brennan
@MORGANLBRENNAN

KEY POINTS

Tom Mueller, who once spearheaded SpaceX’s rocket engine and reusability development, is betting on the in-space services economy with his new company.

Impulse Space, founded in 2021, builds space tugs that can move cargo to different orbits.

So far, Impulse Space has raised $30 million in seed funding last year from investors such as Peter Thiel’s Founders Fund and Lux Capital. It’s currently embarking on a Series A round.

[greybeardengineer]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

Tom thinks cloud servers will move to space as they are so power hungry.

I would always respect Tom's opinions on anything aerospace but I come from the EE side and this comment puzzles me. One of the more costly elements of server farms is heat rejection which becomes far more problematic in space. For a cost sensitive, power hungry, cooling intensive commodity service it really puzzles me why he thinks any aspect of it is helped by being in space. Heck, has he ever tried lifting a basic 2U rack server? 

[JayWee]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

Tom thinks cloud servers will move to space as they are so power hungry.

I would always respect Tom's opinions on anything aerospace but I come from the EE side and this comment puzzles me. One of the more costly elements of server farms is heat rejection which becomes far more problematic in space. For a cost sensitive, power hungry, cooling intensive commodity service it really puzzles me why he thinks any aspect of it is helped by being in space. Heck, has he ever tried lifting a basic 2U rack server? 

I'd love to see a good study of TCO of compute-in-space.
_IF_ you start with a demand for a 100% renewable powered compute (ground solar/wind), you need to take into consideration the day/night cycle, yearly insolation variation, associated energy storage, possibly transmission, etc.
When you put it into space, you can pack everything into one optimized package. I'm curious how competitive it'd be.

[Robotbeat]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

Tom thinks cloud servers will move to space as they are so power hungry.

I would always respect Tom's opinions on anything aerospace but I come from the EE side and this comment puzzles me. One of the more costly elements of server farms is heat rejection which becomes far more problematic in space. For a cost sensitive, power hungry, cooling intensive commodity service it really puzzles me why he thinks any aspect of it is helped by being in space. Heck, has he ever tried lifting a basic 2U rack server? 

Extremely low cost space launch means you can ALSO launch heavy radiators cheaply.

I’ve thought a lot about this. A basic budget 1U server has a value density of $100/kg, one stuffed with high performance components is $1000/kg, and if you actually build the server like you’re trying to make it lightweight, you can build them even denser like a cellphone is about $5000/kg, whereas Starship aims for under $10/kg, so this is definitely within the realm of possibility.

Starlink satellites are rejecting heat from their buses directly, they also have super cheap solar panels, and their cost of electricity might be comparable to terrestrial power when you put them in the right orbit for the 2nd generation satellites on Starship.

[greybeardengineer]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

Tom thinks cloud servers will move to space as they are so power hungry.

I would always respect Tom's opinions on anything aerospace but I come from the EE side and this comment puzzles me. One of the more costly elements of server farms is heat rejection which becomes far more problematic in space. For a cost sensitive, power hungry, cooling intensive commodity service it really puzzles me why he thinks any aspect of it is helped by being in space. Heck, has he ever tried lifting a basic 2U rack server? 

Extremely low cost space launch means you can ALSO launch heavy radiators cheaply.

I’ve thought a lot about this. A basic budget 1U server has a value density of $100/kg, one stuffed with high performance components is $1000/kg, and if you actually build the server like you’re trying to make it lightweight, you can build them even denser like a cellphone is about $5000/kg, whereas Starship aims for under $10/kg, so this is definitely within the realm of possibility.

Starlink satellites are rejecting heat from their buses directly, they also have super cheap solar panels, and their cost of electricity might be comparable to terrestrial power when you put them in the right orbit for the 2nd generation satellites on Starship.

So maybe instead of 1000x more expensive than terrestrial a server CPU hour in space might be only 200x more expensive. What is the BENEFIT of a CPU hour in space vs on the ground that is so great that makes anyone even consider this?

[Robotbeat]

Because ultimately you can get power and cooling and perhaps globally accessible data via lasers for cheaper than terrestrial. You can use lasers to directly access the Starlink constellation. If space based solar power is even close to closing, you can just access that energy directly in space, eliminating the need for a power transmitter, power receiver on the ground, transmission on the ground from that receiver to the broader terrestrial grid. You save a factor of 2 or 3 transmission inefficiency plus the need for a kilometer scale transmitter and receiver and all the expensive RF electronics associated with that.

I think it’s feasible for radiators to be cheap enough to make all that viable, and potentially competitive with terrestrial servers.

Ultimately, it will be necessary if you assume terawatts of electricity being used for computation in the future.

[greybeardengineer]

Because ultimately you can get power and cooling and perhaps globally accessible data via lasers for cheaper than terrestrial.

LOL, ok. At this point let's just agree to disagree because we aren't even in the same solar system in terms of world view.

[Robotbeat]

Because ultimately you can get power and cooling and perhaps globally accessible data via lasers for cheaper than terrestrial.

LOL, ok. At this point let's just agree to disagree because we aren't even in the same solar system in terms of world view.

Have you updated your worldview on this? Just run some numbers. Starlink gen2s may cost around $100/kg. Starlink v1 was $1000/kg.

They’re using commodity solar cells probably costing about 20-30¢/Watt, and using a German built stringer (used for making commodity terrestrial utility scale solar modules for 30-40¢/Watt) to solder them up into spacecraft arrays. They’re not manufacturing the satellites by 3D printing or expensive CNC machining, but by mass manufacturing stamped metal or molding, like how you mass-manufacture automobiles. We ARE in a different solar system already…

[gongora]

https://twitter.com/GoToImpulse/status/1676025311744622593

Flight preparations are underway!

The photo shows one test axis of the Structural Protoqualification Random Vibration and Sine Burst test campaign to qualify our Mira spacecraft serial number 2 for flight in October 2023.

[Jer]

Interview with Tom Mueller about Impulse, SpaceX etc

https://podcasts.apple.com/gb/podcast/manifest-space-with-morgan-brennan/id1680523433?i=1000618013896

Tom thinks cloud servers will move to space as they are so power hungry.

He also developed on NG's TR107 1,1Mlbs LOX/RP1 engine which unfortunately never found a LV. 

Sent from my SM-T733 using Tapatalk

It hard to find something if it's buried  it 6 feet under.

[edzieba]

Because ultimately you can get power and cooling and perhaps globally accessible data via lasers for cheaper than terrestrial.

Power very arguably. For comms, terrestrial optical fibre links beat satellite laser links up and down the street any day of the week, and all your interconnects are also terrestrial. Your server remains local to your customers rather than whipping out of view every few minutes (and stuck on the other side of the planet for a good chunk of the orbit) negating any local latency advantages, etc. If you want to avoid the "server runs away all the time" issue by putting a sever on every satellite and porting server state between each to 'hover' the session in place, you have no added even harsher bandwidth requirements to the system for the ISLs (need to synchronise possible terabytes to petabytes between clusters every few minutes) and insane architectures to ensure coherency.
Plus needing to actually access your servers occasionally for maintenance and upgrades. If you're willing to forgo easy access to hardware, you may as well skip the rocket launch and dump your enclosed server room in the ocean a'la Project Natick instead, and take advantage of free cooling (power savings!), and avoid the disadvantages of power and comms. And you can still pull the whole thing back out for re-use afterwards too.

Putting servers in orbit basically adds a bunch of extra headaches and makes everything more difficult, based only on a potential saving on bulk solar power (which you then probably need to waste on driving active cooling anyway).

[Asteroza]

Because ultimately you can get power and cooling and perhaps globally accessible data via lasers for cheaper than terrestrial.

Power very arguably. For comms, terrestrial optical fibre links beat satellite laser links up and down the street any day of the week, and all your interconnects are also terrestrial. Your server remains local to your customers rather than whipping out of view every few minutes (and stuck on the other side of the planet for a good chunk of the orbit) negating any local latency advantages, etc. If you want to avoid the "server runs away all the time" issue by putting a sever on every satellite and porting server state between each to 'hover' the session in place, you have no added even harsher bandwidth requirements to the system for the ISLs (need to synchronise possible terabytes to petabytes between clusters every few minutes) and insane architectures to ensure coherency.
Plus needing to actually access your servers occasionally for maintenance and upgrades. If you're willing to forgo easy access to hardware, you may as well skip the rocket launch and dump your enclosed server room in the ocean a'la Project Natick instead, and take advantage of free cooling (power savings!), and avoid the disadvantages of power and comms. And you can still pull the whole thing back out for re-use afterwards too.

Putting servers in orbit basically adds a bunch of extra headaches and makes everything more difficult, based only on a potential saving on bulk solar power (which you then probably need to waste on driving active cooling anyway).

While this is all becoming a bit off-thread, the only thing wrong with GEO cloud computing specifically is latency (and the requisite terrestrial transmitter power).  You have an easier time with sun tracking and terrestrial receiver tracking which means you avoid the moving cluster problem, and you are in the sun all the time. Doesn't solve the cooling problem per se but it might make shading/radiator orientation easier.

But Microsoft is correct, putting submersible datacenters near coastal cities is easier in the short term. Putting it beyond territorial waters makes licensing easier too. Also, not being visible means most NIMBY beachouse owners won't complain either since it won't affect property values. I remember there was some SPS opposition due to the visible light reflecting of SPS sats at GEO.

[FutureSpaceTourist]

https://twitter.com/timfernholz/status/1683511289287684096

Impulse Space, the rocket engine company from SpaceX vet Tom Mueller, says it raised a $45 million Series A...some start-ups are still finding VC backing!

Edit to add:

https://techcrunch.com/2023/07/24/impulse-space-is-flying-high-with-new-funding-led-by-rtx-ventures/

[tbellman]

Space.com has an interview with Tom Mueller (founder and CEO of Impulse Space, and former SpaceX rocket engine designer):

Impulse Space CEO Tom Mueller talks early days at SpaceX, moon bases and a booming space industry (exclusive)

[FutureSpaceTourist]

https://twitter.com/lrocket/status/1720493154292629632

Saiph thruster candy corn @GoToImpulse

[FutureSpaceTourist]

https://twitter.com/gotoimpulse/status/1723082114269430202

Any company’s first launch of their product is always an incredible feat. But for Impulse, ours is a little more special because of the journey it took to get here. We moved into a new 55,000 sq ft facility in Redondo Beach in March of 2023... (1/6)

The building was an old furniture warehouse and was the farthest thing from a space factory. The final epoxy wasn’t even laid yet, but there was no choice, the 7,000 sq ft facility in El Segundo was incapable of supporting 50+ employees, machining equipment & engine testing (2/6)

On top of this, just 2 months before in January, Mira, our 1st vehicle, wasn’t even designed yet. Not only did the Impulse team renovate an entire building and make it space-ready, but we designed, built, tested and shipped a vehicle within ~6 months (3/6)

We also had to finalize a Series-A raise in an economy that has been less than ideal. We also doubled in size and now sit at over 85 full-time employees.   On the eve of launch, it’s important to look back and reflect on the incredible hard work, (4/6)

hours, rigor & sweat that the Impulse team has put in and how we got to this point. Building an incredible team is half the battle; but execution is the rest of it. And that starts this weekend. (5/6)

To the employees shown below, and those not, thank you for your incredible efforts to get us to this point.
   
Tune in on November 11th at 10:47am PT to see Transporter-9 lifting off out of Vandenberg. Go Impulse!

#MiraFirstFlight #ImpulseSpace

[Cheapchips]

https://twitter.com/lrocket/status/1723769027435405633

Great news for our Mira LEO Express 1 mission- we are receiving data from Mira and spacecraft health is good!  I want to thank the Impulse team, our investors and all of our partners and suppliers.

[FutureSpaceTourist]

https://twitter.com/sciguyspace/status/1724191232425910604

Tom Mueller has moved on from SpaceX and rockets to apparent success with in-space propulsion.

https://arstechnica.com/space/2023/11/impulse-space-appears-to-succeed-with-its-first-spacecraft/

SpaceX founding employee successfully moves from rockets to in-space propulsion
"We want to make it cheap and easy to get anywhere in the Solar System."

by Eric Berger - Nov 13, 2023 10:07pm GMT

SpaceX launched its ninth "Transporter" mission on Saturday from California, carrying dozens of small- and medium-sized satellites into low-Earth orbit.