@StockManOnX Agree. Deal risk is pretty close to 0%. I model it at 10% conservatively because why not.
English
Kish
6.3K posts
Kish
@KishavanBhola
trading; math + physics education; ex: ML eng @googleAI brain, quant @blackrock;
🇺🇸 Katılım Mayıs 2017
605 Takip Edilen2.6K Takipçiler
To be clear, the $SATS - $SPCX discount can grow.
I've seen a lot of people on here claiming $SATS is some slam dunk long, but I don't know. A 32% discount seems reasonable to hold through a 23x float increase.
Some people claim the discount is larger but they don't consider (a) time-value of money, (b) deal risk, and (c) taxes.
Kish@KishavanBhola
EchoStar's ($SATS) deal with SpaceX gives a fascinating market-implied price for SpaceX in 2028 — after the massive lock-ups and index buying. On my model it's ~32% below SpaceX ($SPCX) today. SPCX floated only ~4% of itself, so as lock-ups roll off through 2027 the tradeable float grows ~23×! My estimated forced index buying (Nasdaq-100, total-market, Russell/MSCI) is only ~1.4× today's float. That −32% is the price of being forced to hold SPCX through all of it. SATS is essentially a SPCX wrapper; its legacy business nets out small after debt, so the SpaceX stake is most of its equity value. But it doesn't receive the shares until the deal closes (~Nov 2027) and likely can't sell until 2028. So I model SATS conservatively, as a taxable, deal-contingent forward on SPCX delivered in 2028. And −32% is conservative. It credits a ~10% deal-break risk and full corporate tax on the gain. Strip those and the raw gap is north of 40% for buying SATS vs SPCX.
English
And taxes are unclear –– maybe SATS figures out how to save those –– but assuming you receive the post-tax equity value as a one-time dividend is the liquidation model and the most conservative approach.
The market always sells these forward asset wrappers at a discount.
English
EchoStar's ($SATS) deal with SpaceX gives a fascinating market-implied price for SpaceX in 2028 — after the massive lock-ups and index buying. On my model it's ~32% below SpaceX ($SPCX) today.
SPCX floated only ~4% of itself, so as lock-ups roll off through 2027 the tradeable float grows ~23×!
My estimated forced index buying (Nasdaq-100, total-market, Russell/MSCI) is only ~1.4× today's float. That −32% is the price of being forced to hold SPCX through all of it.
SATS is essentially a SPCX wrapper; its legacy business nets out small after debt, so the SpaceX stake is most of its equity value. But it doesn't receive the shares until the deal closes (~Nov 2027) and likely can't sell until 2028. So I model SATS conservatively, as a taxable, deal-contingent forward on SPCX delivered in 2028.
And −32% is conservative. It credits a ~10% deal-break risk and full corporate tax on the gain. Strip those and the raw gap is north of 40% for buying SATS vs SPCX.
English
No, the market isn’t valuing $SPCX more than $AMZN.
Reminder:
Kish@KishavanBhola
Tiny float, not shortable, lockup for weeks, index inclusion buying ahead, boomers and market bearish into IPO, humanity’s only path to the stars In the near term, $SPCX could trade anywhere. The market cap is essentially meaningless
English
Apparently this is a "deep research" benchmark from Perplexity. Having this kind of result for agentic software would be much more interesting (and economically valuable).
The other thing is, a fusion of 3 cheap OS models plus an Opus 4.8 judge (that they conveniently don't mention) *still* underperforms solo Fable 5. Everything that outperforms it is more expensive.
OpenRouter@OpenRouter
Introducing the Fusion API, the smartest compound model in the market. Fusion achieves Fable-level intelligence at half the price. How it works 👇
English
I feel totally justified running up insane Fable 5 bills on my projects as soon as it was released.
Going back to Opus 4.8 is painful.
OpenRouter claims that multiple models + judge can match Fable 5 at lower cost. I don't trust benchmarks but worth a shot.
Kish@KishavanBhola
I'm probably not buying these IPOs (already have exposure) but I do know that I will pay crazy amounts for frontier model intelligence. My exploding Fable 5 bills are proof of that. It's not like Opus solves the same problems but slower; it simply can't accomplish what Fable does. For some problems it's worth it to pay literally 10x or even more.
English
Kish retweetledi
As a result of a US government directive, we are suspending access to Claude Fable 5 for all users. You can continue to use all other Claude models.
Here’s what this means for you:
Across Claude products, new sessions will run on your selected default model or Opus 4.8, and existing Fable 5 sessions will end with an error.
On the Claude Platform, requests to Fable 5 will also return an error. Please update your integrations to other Claude models.
We know this is a disruption to your workflows; we appreciate your patience and support.
Anthropic@AnthropicAI
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…
English
Don’t get me wrong — I get it.
Starlink can become the internet.
AI inference compute in orbit at scale
SpaceX is building the Dyson Sphere.
Who cares about valuation also. A lot of people would donate their money to make Mars happen.
150x revenue is unprecedented but so is SpaceX.
English
I'm probably not buying these IPOs (already have exposure) but I do know that I will pay crazy amounts for frontier model intelligence. My exploding Fable 5 bills are proof of that.
It's not like Opus solves the same problems but slower; it simply can't accomplish what Fable does. For some problems it's worth it to pay literally 10x or even more.
Jeremy Raper@puppyeh1
So you can use the 5th/6th/7th best LLMs, getting 80-85% of the top guys' performance, but at an 85-95% discount in price? You know what we call that? A commodity... exactly what happened with LCD TVs, OLEDs, solar panels, electric cars, phones, etc good luck with your AI IPOs!
English
This model is SOTA on everything and available in two flavors:
1) Fable 5 for agentic systems
2) Mythos 5 for cyber and biomedical research acceleration
Claude@claudeai
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.
English