When competition heats up, you train harder.
Therefore, compute demand is going to increase.
LLM models aren’t done and development or improvement probably won’t ever stop
Why I think we’re not due for a huge crash just yet.
> The A.I race is still on, with $META now saying that their newest model outperforms GPT 5.5. Do you think Sam Altman, someone with a tiny ego wants to lose this race? He needs this to work, if not his IPO is gonna fail in 2027.
> KOSPI is recovering and bounced on key moving averages today (50MA).
> SPY also looks like it’s just consolidating here. It actually printed a bullish divergence on the daily chart, with the 10/20EMA bullish cross occurring as well. Break out coming with a break above the down sloping resistance, tested 3 times so far.
What are you thinking about all this?
Do you know who else builds satellites and has robotic arms?
$RKLB
NASA awarded the Swift Boost contract via SBIR Phase III in September 2025 specifically to Katalyst Space for a $30M, sub-1-year development of LINK: a spacecraft with three robotic arms for autonomous rendezvous, mechanical capture (Swift has no docking port), and slow multi-month orbit boost. Katalyst beat proposals from Starfish Space and a Cambrian Works/Astroscale JV.
The mission demanded specialized grappling tech for an aging, non-cooperative government satellite
A @NASA space telescope is falling, and a robot is gonna try to catch it tonight! The NASA Swift Observatory has studied some of the most violent events in the universe for over 20 years: gamma-ray bursts, deaths of massive stars, neutron star mergers. It's one of the telescopes that studied the kilonova that gave us our first direct evidence that neutron star mergers forge the heavy elements like gold and platinum! It's in the middle of a mystery that could upend our current models of these events. It's an important ~$500 million telescope doing important science. That's why we want to save it.
Now, why it's sinking. Satellite orbits in LEO naturally decay due to atmospheric drag. But there's more to Swift's story. Remember the big solar storms and aurora in 2024? That was during the solar maximum, when solar activity ramps up (more flares, space weather, aurora). The extreme UV radiation the Sun spat out heated Earth's upper atmosphere, causing it to expand. So it got thicker at Swift's altitude (373 miles), thus more drag, leading to Swift dropping to just 220 miles high.
So, NASA gave a startup called @katalystspace Space $30 million to rescue Swift. Katalyst built a robotic spacecraft, LINK, in just ~9 months (normally takes YEARS to do this)! It'll fly on the very last Pegasus rocket @northropgrumman has.
The plan: LINK will follow Swift for about a month and snap photos to find points where it could grab it, then it'll latch on with its three robotic arms, and slowly boost Swift back up to its orbit (like several months slow). If this works, Swift gets years of life back, and it'd be the first time an American commercial spacecraft captures an uncrewed US government satellite. Pretty awesome for this to be the Pegasus final flight!
It's incredible to watch as the space economy unfolds. A glimpse of what it looks like to repair our satellites instead of letting them burn. $30 million to repair Swift instead of hundreds of millions to replace it.
LINK was supposed to launch today but it was scrubbed, so next attempt is tonight (July 1 at 2:43am PT). No livestream, but I'll be watching NASA's updates with my fingers crossed tightly. 🤞
Image: Artist concept from Katalyst
@aleabitoreddit Alexander Wang says their new “Watermelon” model uses an order of magnitude more compute than “Avocado” to me, and order of magnitude would be at least 10x the compute used. How would they have excess? They will need to buy more compute to sell it at a premium to the market
@OracleNYSE $ONDS completed full Acquisition of World View Enterprises on April 1st, 2026; World View Enterprises is one of the providers eligible for future task orders under the contract. Ondas is not directly named.
@BrettKrieger12 Imagine the amount of worldwide umanned drones and assets that will need L-band communications for precise position, navigation, and time data today and in the future... from health and tracking collars for cows, construction and farming equipment, as well as military assets.
@wholemars The idea is that they'll be so plentiful, you wouldn't need to own one unless you want to manage them as a business owner to cater to your own customers
@aleabitoreddit Augmented reality glasses or headsets that overlay a face or cosmetics to a robot. Like a realtime Snapchat face filter. No need for physical cosmetics that smear or create more work to fix, clean up, or permanently stain the underlying material.
The value of L-band will be increasing with the growth of drones and robotics.
Low frequency L-band is needed for military defense operations and asset management especially if higher frequency bands fail or are actively blocked.
Drone operations are going to be conducted with massive numbers to overwhelm enemies. Each drone will need to be accounted for to ensure they don’t get into the wrong hands to be reverse engineered, or for planned maintenance, management including planning supply chains in advance of an asset breaking down. Warfare will require a complex system to network and manage the swarm with precision of position, navigation, and timing; taking advantage of the properties of L-band.
The market for $RKLB is getting bigger
Rocket Lab 🤝 Iridium – a powerful combination.
Our founder and CEO @Peter_J_Beck is joined by @IridiumComm’s CEO Matt Desch @IridiumBoss and Rocket Lab CFO Adam Spice to discuss how this acquisition will unlock and accelerate our future in space applications.
Links in the comments.