David Kwon @dkwon98

Partner @AbstractVC SF Joined August 2016

Tweets

1K
Followers

2K
Following

771
Likes

4K

David Kwon @dkwon98

13 hours ago

@allisonkiang @neilmovva Congrats Allison!! Lfg!!

0 0 1 89 0

View Details

David Kwon @dkwon98

18 hours ago

@neilmovva @blintzbase @sailresearchco Huge congrats team! Amazing progress

0 0 6 240 0

View Details

David Kwon @dkwon98

2 days ago

@sailresearchco 🔥 🔥

0 0 2 365 0

View Details

The XDOF team is a world-class team of researchers who deeply understand the problems their customers face. They are open-sourcing the largest teleop dataset and want to accelerate those in pursuit of general purpose robots. We're excited to be supporting them, congrats to the team!

Philipp Wu @philippswu

a week ago

Announcing @xdofai: We’ve raised $70 million to build the core robotic infrastructure ecosystem for robot foundation models. My cofounders Fred (@YideShentu), Nemo (@itsnemojin) and I have been pursuing the dream of general purpose robots for our entire lives. After work at

86 71 697 170K 280

1 0 20 2K 1

View Details

David Kwon @dkwon98

a week ago

@cattttang Agreed, mentioned in point 1!

0 0 1 104 0

View Details

David Kwon @dkwon98

a week ago

Let’s count the wins here 1. SpaceX gets a frontier coding lab, with the highest quantity and quality of coding data outside of Anthropic 2. Cursor is at $4b ARR and targeting 7.5b+ in EOY ARR. At $60b, that’s accretive to SpaceX’s EV/revenue multiple 3. One of Cursor’s only deficiencies was long term access to compute, which they now secure. Who knows what they’ll cook 4. Cursor was the largest customer of several inference neoclouds, meaning SpaceX slows down or handicaps competitors in the process 5. SpaceX gets Cursor level AI talent with a meaningful lockup, at a time when they face huge attrition in xAI Win x 5

SpaceX @SpaceX

a week ago

SpaceX has exercised the option to acquire @cursor_ai in an all-stock transaction with the goal of building the world’s most useful AI models. For the past few months, SpaceXAI has been jointly training a model with Cursor, which will be released in Cursor and Grok Build soon.

2K 4K 37K 25.8M 3K

20 40 1K 105K 131

View Details

David Kwon @dkwon98

a week ago

@tdsfixer @grok Answer is probably yes. Cursor is already a preferred enterprise solution because it works well and you can be model agnostic. Lack of stability associated with Anthropic models is a tailwind for any other coding company

0 0 1 49 0

View Details

David Kwon @dkwon98

a week ago

@Timmetie Horrendous take

1 0 2 199 0

View Details

David Kwon @dkwon98

a week ago

@OosGergVer you're wrong if you think all of the attrition was intended

1 0 2 717 0

View Details

David Kwon @dkwon98

a week ago

@calitrav Yes, but the conversion price will be dependent on SpaceX's average share price closing to the closing date (unless I'm mistaken)

0 0 13 609 0

View Details

David Kwon @dkwon98

a week ago

@GerritD They had definitive offers for this most recent funding round at ~$50B. So it wasn't out of necessity. But they made the hard (and imo correct) decision that they could go further and faster as part of SpaceX. Everyone won here

1 0 21 999 0

View Details

David Kwon @dkwon98

a week ago

Group chat alpha is a real thing

0 0 8 861 0

View Details

Revel.io @Revel_Software

a week ago

@GoToImpulse is on a mission to accelerate humanity's future beyond Earth. To get there, they are building the spacecraft that will advance the orbital economy, and need to test as fast as they build. Before Revel, standing up a new test stand took 1–3 weeks, first-time deployments could take up to 3 months, software changes required looping in a specialist, and the team was locked into a single DAQ ecosystem. Now, stands are up in days, sometimes even hours. Engineers self-serve on sequence changes, the program runs multiple test fires per day instead of one every other day, and with the ability to mix and match hardware across vendor ecosystems, the team saves over 100x on hardware alone. Today, Revel runs across hundreds of Impulse assets — engine stands, vacuum chambers, avionics test stands, facility monitoring, and more — creating a single codebase that can be maintained and deployed consistently at scale. Thanks to @lrocket, @_Eric_Romo, Dane Sarcone, Julia Vyborny, Trevor Hanken, and the rest of the Impulse Space team for the partnership. Read the full case study at revel.io/impulse

3 10 48 18K 7

View Details

David Kwon @dkwon98

a week ago

Hard to overemphasize how good of a deal this is for SpaceX. People laud Sam Altman as the deal king, but Elon is a wizard for this. Very rarely does a deal make this much sense strategically AND financially (due to SpaceX’s high stock price). And it’s uniquely enabled by Elon’s aggressive decision to build a frontier lab and eventually Colossus 1+2 at a time when that was non obvious, and people thought the game was over

0 0 43 3K 3

View Details

David Kwon @dkwon98

a week ago

Embedded in 3 is that if SpaceX can’t rent out Colossus 1+2 for favorable terms or if Ant/Google back out of their deals, they have a productive use for that excess compute (Cursor). That gives them more leverage in every negotiation because they have a true fall back plan