Boulder, CO
Joined November 2024
Ari Newman retweeted
IF you’re paying attention to financial markets and/or overall US economy THEN this dialogue below is gold standard to digest. Very high signal to noise ratio and better than all the ‘click bait headlines’ you will be served in coming weeks/months trying to further explain this trend.
Replying to @chamath
Pretty much. As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers. SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue. That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.
27
31
2
747
165,113
I see a ton of defense company pitches with "working with Ukraine" - real-world deployments are valuable and prove the tech works. It's not clear whether any of these will become venture-scale businesses. One temporary customer doesn't = global market.
1
35
Ari Newman retweeted
Turns out the future does have flying cars
48,033
100,034
32,749
834,245
79,215,368
My 18yr old just said… WTF we got rocket league IRL b4 GTA6!
Configurations – Performance: road ready, seats 2 Active aerodynamics & multi-mode suction system for next-level performance on & off the track – Thrust: race ready, seats 2 Multi-mode suction system paired w/ rear-mounted cold-gas thrusters that enable 0–60 in <1 second – Rocket: air ready, seats 1 Multi-mode suction system plus perimeter-mounted cold-gas thrusters supporting forward & upward acceleration plus lateral rotational movement. So...yes, it can fly Range >300 miles Top speed up to >230mph (Rocket) 0-60 Performance (<1.6) Thrust (<0.9) Rocket (<0.59) Reserve now – Tesla.com/Roadster
22
It’s almost like @AnthropicAI is asking to be shut down but can’t stop advancing itself due to lust for power.
16
⚠️ We are now entering dot-com risk territory with AI credit. CoreWeave CDS just crossed 1,000bp - ~63% implied default odds, up from 30% a year ago. Oracle is at 300bp. Meta's A+ data center bond is down 20% from issue. Credit leads equities. The why 👇
1
30
The AI buildout's dirty secret: it's circular. NVDA invests in the hyperscalers, who turn around and pre-buy billions in GPUs for datacenters that don't exist - and may never. 55% of AI investment is now AI-on-AI. Vendor financing is exactly how telecom ended in 2000. The other pin: a step-function in inference efficiency, or a faster jump to open-weight models, turns frontier training economics into commodity hosting overnight. Different business, different math, same debt. The accelerant is the profitability wake-up call. Are any AI datacenters profitable? Not the pure-plays - yet. The model works if utilization stays high and GPU-hour prices hold. They haven't: H100 rental is down ~80% in a year against 3-5 year debt. Gartner: $3.5T of AI infra spend is impairment-exposed by 2030 One writedown headline - say NVDA takes $100B on future revenue - and everyone takes a much harder look at the economic merry-go-round. The unwind starts at the weakest balance sheet and moves inward. That's why CoreWeave's credit matters more than its size. We've been here before. Unbridled optimism, banks backing up the story, and no one making conservative decisions.
60
For those of us who work on computers all day - here's the real unlock with @Starlink on flights. Instead of in-air hours being a productivity loss that compounds while on the road, it is now flipped to a gain. No office distractions; pure grinding on tasks and deep work. I get ahead before I land, and I can catch up on the return instead of the backlog getting worse and worse. thank you @elonmusk, @Gwynne_Shotwell @united!
10
We are barely 2 years into agents. Let this sink in. 🤩
This is f*king insane. SpaceXAI engineer Nick Miller (ex-Cursor): "99% open one Grok Bot and call it automation. I run 20+ bots with a Chief of Staff, PM, QA, and 15+ workers." In this 30-minute demo, he shows how to turn one bot into a coordinated team Grok Bot → Chief of Staff → PM → QA → Workers More bots don't make a team. Handoffs and verification do Worth more than most $1000 agent engineering courses Bookmark and watch it today Then read my 4-Bot Grok Bot setup with full prompts below ↓
13
Ari Newman retweeted
Colorado has something awesome called the Taxpayer Bill of Rights. Among other things, it protects a simple principle: everyone pays the same state income-tax rate. If your income doubles, your tax doubles. Seems reasonable, right? Amendment 87 would repeal that protection and nearly double the top marginal state tax rate while modestly reducing the lowest rates. Under the new rules, doubling your taxable income could mean a state income-tax bill nearly four times what you’d owe on your current income today. To me, this looks like a blatant attempt to buy votes: the majority votes themselves a tax break funded by much higher taxes on a small minority. And it hits startups especially hard. Founders and early employees often spend years building something before receiving a substantial payoff from selling their shares after an IPO or in an acquisition. A decade of work with no guarantee of success can produce a payout counted as income in a single year—and taxed at much higher rates. I’ve spent a decade building Boom in Colorado, on track to create hundreds of high-paying jobs. Colorado’s tax fairness was one reason I chose to build here. Now voters are being asked to change those rules and take a much larger share of a potential taxable payoff from that work. Should I have built Boom in Texas, which has no state personal income tax? Should the next ambitious founder choose Colorado or somewhere else? Colorado is at a crossroads. Do we want to stay true to our OG spirit of rugged independence, freedom, and fairness—or make Colorado a less attractive place to build? Punishing success is morally wrong and bad policy to boot. We should be celebrating the people who build new things and create new wealth for themselves and others. Their work gives us companies, products, and jobs. Let’s keep Colorado as one of the best places to live, build, create, and innovate—and to enjoy the rewards of our hard-fought efforts. I’m voting no on Amendment 87.
148
189
22
1,902
84,938
Time to short the stock.
Delta Airlines CEO Ed Bastian in new interview on Delta not adopting @Starlink: "We know what we’re doing in the space, and we’ve been the pioneer in Wi-Fi. We set that standard launching free, fast Wi-Fi that forced the industry to adopt it. We’re happy with where we sit." He said he wasn't worried about being at a competitive disadvantage by not having Starlink. Delta said it will adopt Amazon's LEO satellite internet, but won't start installations until 2028. (via WSJ)
14
Huge deal. RIP terrestrial carriers
We announced an agreement to acquire a nationwide low-band spectrum license portfolio that will pave the way for @Starlink to become a major mobile carrier in the US. With this new spectrum and our Gen2 constellation, Starlink Mobile can ensure Americans have access to high-speed mobile broadband no matter where they are → spacex.com/updates#starlink-…
3
I suspect we'll see more VCs quit the industry and go back to building things, or try for the 1st time - in the next 12-24 months. There has never been a better time to build given the state of AI tools. It's rewarding and 10x faster than ever. Venture is in a strange place w/ the lion's share of the money going to mega funds that will never outperform (statistically speaking) and a lot of emerging managers' 2020-23 era funds underperforming making the next raise even harder.
1
17
Ari Newman retweeted
I read the fine print on Colorado Amendment 87, the graduated income tax. The ballot table shows a tax cut for everyone under $1 million. The most anyone saves is $325 a year. Earn under $15,000? You save $1. In reality, nearly every income-earning Coloradan will pay more in taxes over time if 87 passes. Paired with Prop NN, the state keeps more of what you pay. 87 will also drive founders, companies and capital out of Colorado, and younger Coloradans will pay the price. Either way, a no vote means a stronger Colorado for every Coloradan. #Colorado #EngageColorado #Amendment87 Read the fine print in today's Engage Colorado newsletter in the replies below ↓
3
1
3
280
Well deserved! congrats @LunarOutpostInc
Lunar Outpost has been selected for @FastCompany's 2026 Next Big Things in Tech list! Fast Company's Next Big Things in Tech has recognized our Pegasus Lunar Terrain Vehicle, the next generation rover Astronauts will drive on the Moon in 2028, as a new technology breakthrough that will help define the future of the space industry. Pegasus also supports a range of mission profiles including lunar site exploration, foundational science operations, resource prospecting, and surface site preparation as NASA lays the groundwork for a permanent Moon Base by 2030. We are honored to receive this award alongside industry leaders advancing the next frontier of technology and innovation. Read more: fastcompany.com/91612630/spa…
1
28
I've been having major problems with a service provider. I escalated the issue. Suddenly, all the sloppy, slow work gets fixed in 24 hours. This only reinforces that I have to move to another team or provider. The fix affirmed my suspicion that the problem is a person, not a process.
6
Is any other space company even in the same league? Not even close. $SPCX
SpaceX’s launch week has been absolutely insane In just FOUR DAYS, SpaceX launched every vehicle in its current orbital lineup: • Falcon 9 • Falcon Heavy • Starship 3 different launch vehicles 4 different launch pads 3 different states: Florida, California and Texas And they basically covered every major category of payload: • Starlink • Crew • Commercial • Military / national security Then on October 1 alone, SpaceX completed THREE launches in less than 13 hours SpaceX is simultaneously operating across almost every major part of orbital spaceflight now Different rockets, different pads, different coasts, completely different missions...all running at the same time No other launch company today is operating across this combination of vehicles, launch sites, mission types and cadence
60
In deep tech, revenue isn't the leading indicator. It's the lagging one. The moat comes first. These are "Fat Startups"... Judging these companies on SaaS timelines is the wrong lens. Wrote about it, and about why the slowest feedback loops matter most: arinewman.substack.com/p/the…
1
10
I keep having conversations with very smart people about the physics / thermal problems of data centers in space but Starlink V3 is starting to prove the question is price performance not “will this be possible”. Thermal is an engineering and cost problem, not a physics problem. The radiator is smaller than the solar array, and V3 is already rejecting a fifth of the target load in production. The open question is dollars per kW-year, and that's a launch-cost bet starship can answer. Going to be very interesting over the next few years!
17
Everyone's posting about agents this week. Almost no one is talking about how this validates the underlying infrastructure markets. Every agent that takes action burns compute, pulls data, opens an attack surface and draws power. The demo is the cheap part. What I'm wondering is what happens when this level of agentic work goes mainstream. Will we see the loads spike? Will foundation inference slow down? That's what I'll be looking at.
10