Paul | Bitcoin, Crypto & Equities | Macro | Charts & On-Chain | NFA | 馃嚚馃嚘
Ontario, Canada
Joined November 2025
- Tweets44.4K
- Following691
- Followers3.5K
- Likes234K
Not Another Quant retweeted
So many are wondering what鈥檚 going on with $BE after such a great quarter. Bloom hasn鈥檛 been paid for a lot of what they sold yet and permits are now slowing down some of their projects.
Bloom made $250M of profit in Q3 but their operating cash flow was negative $27M. In the 10-Q you can see a lot of that comes from $464M of Q3 revenue they haven鈥檛 been able to bill yet. They only get to bill once a project reaches certain stages or the customer signs off on it.
Q1 and Q2 were both positive so operating cash flow is still $273M for the year. Bloom鈥檚 CFO said they don鈥檛 expect Q3鈥檚 negative cash flow to keep going.
But that depends on projects staying on schedule. Bloom has paid suppliers $337M in advance this year and if projects get pushed back they could end up with more inventory than they need and less cash.
They say permits are now delaying some of their projects. They usually can鈥檛 sell the systems to the investors paying for a project until it has its permits so a delay can mean Bloom gets paid later.
Their biggest project is the $ORCL data center in New Mexico and it still doesn鈥檛 have its air permit. The permit hearing isn鈥檛 until December 7 and an October 26 hearing will look at whether to throw out the application.
Not Another Quant retweeted
The yen continues to strengthen with inflation numbers coming in above BoJ target. The BoJ has been saying it鈥檚 worried about waiting too long to hike. The issue is hiking enough before inflation gets stuck above 2%.
And now with headline and core both above 2% I think a December hike is more likely.
Global stocks also sold off today and that helps the yen. People buy the yen as a safe haven when stocks drop and some close out carry trades which means buying the yen back.
US yields also dropped today because money moved into US Treasuries. People hold dollars over yen because US rates pay more and when US yields drop that extra pay gets smaller.
It looks like the carry trade is fairly calm for now. Yen short positions are a lot smaller than they were in July after most of them were closed out in August and September.
Not Another Quant retweeted
The new $SPCX 10-Q shows one customer SpaceX doesn鈥檛 name made up 1/3 of their revenue in Q3. That鈥檚 about $5.0B and it鈥檚 2/3 of their AI revenue.
In Q2 the same customer was about $1.5B so it more than tripled in one quarter.
It looks like it could be Anthropic. Reuters reported Anthropic鈥檚 contracts with SpaceX are worth up to $84.5B into 2029 which is nearly double the original deal.
We don鈥檛 have that confirmed but if Anthropic files publicly soon then the S1 should show their SpaceX contracts.
Not Another Quant retweeted
Yields keep moving with AI stocks. When AI sells off people run to bonds and yields drop and when AI bounces yields go back up.
The market is asking if all the money going into AI will pay off. It鈥檚 most worried about companies borrowing to pay for it. About 40% of US growth is coming from AI spending so yields falling seems to show the market is starting to question that growth.
Those growth concerns would feed the recession risk. When people start losing their jobs that鈥檚 when the music stops.
But so far the job market is ok. Layoffs are still low but hiring looks like it could be starting to slow down.
Yields aren鈥檛 falling because inflation is cooling. We need this war to end to send oil down which will help cool headline inflation.
Core still seems sticky and the longer oil stays high the more it risks spreading into other prices. If it pushes wages up that could keep inflation from coming down but so far wage growth is slowing.
I don鈥檛 think many doubt AI pays off eventually. We鈥檙e being promised trillions in AI revenue and it鈥檚 growing fast. But the spending comes first and the revenue comes later. The wait in between is where things could get uncomfortable especially for companies that borrowed heavily.
Not Another Quant retweeted
The yen is gaining on USD because US yields have come down and the Fed is expected to hold rates next week.
Japan鈥檚 inflation is around the BoJ鈥檚 2% target but rates are only 1.25% so borrowing is still cheap. With a weak yen and high oil pushing prices up the BoJ wants to keep hiking before inflation goes past 2%. The market continues to price in a December hike.
Bessent has been pushing the BoJ to hike since the summer because higher rates in Japan support the yen without Japan having to sell US Treasuries to do it.
But the market thinks the Fed will hike in December as well so the difference between US and Japanese rates may not change much.
Japan鈥檚 inflation numbers come out tonight and a higher than expected number would make a BoJ hike in December even more likely.
Not Another Quant retweeted
$META is doing almost the same thing $GOOGL did in 2014. Google gave shareholders a non voting Class C share for each share they owned so Page and Brin could keep control while Google kept issuing stock.
The Class C shares traded lower than the voting ones and Google ended up paying $522 million to make up for it.
The big difference is Meta wants to use these shares to help raise the hundreds of billions they say they need for AI. Google mostly used theirs to pay employees and buy companies.
And it is also for Zuckerberg. If California passes its billionaire tax next month he can sell some of his new shares to pay it without giving up votes.
$META is creating a new kind of share with no voting rights. Every shareholder will get one for each share they own. It鈥檚 a stock dividend but it works like a 2 for 1 stock split.
Zuckerberg holds most of the voting power. Since the new shares don鈥檛 come with votes Meta can sell them to pay for AI and he keeps control.
Zuckerberg will get the new shares as well. He can sell 20% of them without losing voting power. With the other 80% he gives up some voting power for every one he sells.
That rule goes away over 8 years if he stays full time on Meta鈥檚 executive team or as executive chair.
Not Another Quant retweeted
Layoffs are still really low but hiring looks like it鈥檚 slowing down.
US jobless claims fell to 194K last week which is the lowest since July. But ADP鈥檚 weekly jobs number had been going up since August and just dropped to 18,500 from 26,000. More people are also staying on unemployment which can mean it鈥檚 taking longer to find a job.
Claims can stay this low while ADP drops because they count people applying for unemployment and ADP counts how many jobs companies added overall.
Most companies still have enough business to keep their workers. But fuel and tariffs are costing them more so it looks like they are being careful about hiring. It seems some companies don鈥檛 need to hire as many people now that they鈥檙e using AI.
There are still more jobs being added than in July and August according to ADP. They revised their last two numbers up so this drop might end up smaller.
We will know more on November 6 when BLS reports October and revises September.
Not Another Quant retweeted
$SPCX expects revenue to more than double in six months and most of that growth comes from AI.
Part of it is from deals they have signed or expect to sign and part is from their own products like Cursor and Grok Bot.
Either way SpaceX needs a lot more servers running to do this. That鈥檚 why they鈥檙e spending so much in Q4.
They say demand isn鈥檛 the problem so it depends on how fast they can build.
Not Another Quant retweeted
Bessent wants US long yields lower but Japanese investors have sold foreign bonds four weeks in a row. Japan holds more Treasuries than any other foreign country.
A US 10 year used to pay about 4 percentage points more than a Japanese 10 year and now it pays about 2 more. A lot of Japanese buyers hedge the dollar and that costs them more than 2 points so a Japanese bond pays them more. That鈥檚 been true since 2022 but Japanese bonds now pay almost 3% so they are finally worth buying instead.
AI companies are borrowing a lot right now and a big part of it is long term debt. They sell it to the same pension funds and insurers that buy long Treasuries. The US government is competing with AI companies for those buyers at the same time Japan has less reason to buy US bonds.
But US auctions are still getting strong demand at yields above 5%. Bessent needs that demand at lower yields and so does anyone hoping for lower mortgage rates. When the 10 year goes up mortgage rates go up with it and the average rate is 7.4% now.
Not Another Quant retweeted
ETF investors have pulled more money out of ethereum:native this month than in any other month this year but Ethereum is still holding its weekly support.
JPMorgan put their dollar token on Ethereum on Monday. Two of their funds are already on it so it looks like they鈥檙e building more of their business on Ethereum.
The suits are getting on board. Are you?
Not Another Quant retweeted
$SPCX is getting paid more to rent out their AI servers than they expected and customers are paying more when their deals renew.
They said they keep their deals short because they expect prices to keep going up. When SpaceX went public the Anthropic and $GOOGL deals were worth about $2.2B a month. Anthropic鈥檚 original deal can already be ended with 90 days notice and Google鈥檚 can from next year.
But if prices fall customers could leave or pay less when their deals renew. Then it would take SpaceX longer to make back what they spent on the new servers.
Right now they say new servers pay for themselves in under a year which helps them pay to build more.
Not Another Quant retweeted
$SPCX still has all 18,712 of their bitcoin:native . They didn鈥檛 buy or sell any through September and the price going up added about $475M to their Q3 profit.
They paid $661M for it which works out to about $35K per Bitcoin. At the end of September it was worth $1.57B.
Interstellar hands?
Not Another Quant retweeted
Over half of $SPCX revenue now comes from their AI segment. In Q3 AI was 52% of their revenue while Starlink was 36% and rockets (space) were only 12%.
Rockets would be higher if SpaceX counted their own Starlink launches but those don鈥檛 count as rocket revenue.
Starlink still makes most of their profit. AI made $431M in Q3 after $2.5B of depreciation.
84% of their capex is going into AI so their depreciation should keep going up. Going forward a lot rides on that AI spending paying off.
Not Another Quant retweeted
$SPCX now expects to spend about twice what they said they would in Q4 because they鈥檙e adding more AI capacity than they planned in August.
They don鈥檛 make enough cash yet to pay for all of that. They have $89.9B in cash and securities left and they鈥檙e looking to borrow more.
They鈥檙e doing it because they think the new AI servers pay for themselves in under a year and they expect revenue to more than double by Q1.
Not Another Quant retweeted
Still monitoring the ethereum:0x4a220e6096b25eadb88358cb44068a3248254675 situation!
Quant builds the tech that lets banks move tokenized deposits between each other. The Clearing House picked them for the US network in September and UK banks already use a platform Quant built.
Canada鈥檚 six biggest banks are looking at doing the same with Canadian dollar deposits. They haven鈥檛 picked a tech company or set a launch date.
$BMO and $TD US banks are already in the Clearing House network Quant is building.
So I wonder if this makes Quant one of the main options for Canada? But BMO is already working with Google on tokenized cash so Google could be another one.
Not Another Quant retweeted
bitcoin:native continues to do well relative to gold since the low in BTC/GOLD in February.
JPMorgan鈥檚 yearly long term forecast came out yesterday and it includes Bitcoin for the first time.
They think Bitcoin will grow about twice as fast as gold over the next 10 to 15 years (around 10% a year vs 5.5%.)
They also expect it to swing about three times as much as gold. But it鈥檚 not in their sample portfolios yet because it doesn鈥檛 move with other assets in a reliable way.
Slowly then quickly.
Not Another Quant retweeted
I鈥檝e been digging into Canadian National Railway $CNR.TO and its Q2 2026 results reinforce why I consider it one of Canada鈥檚 most compelling long-term businesses.
Here are the numbers:
- Revenue: $4.75B (+11% YoY)
- Diluted EPS: $2.06 (+10%)
- H1 free cash flow: $1.84B (+19%)
- Management raised full-year earnings guidance.
What caught my attention?
The operating ratio rose from 61.7% to 62.5%. On the surface, that looks like a step backward, but higher fuel costs were a significant headwind.
Underneath that, execution remained strong, with record fuel efficiency and solid freight volumes, particularly in grain and energy.
Why I like the business:
An extensive rail network that鈥檚 extremely difficult to replicate.
Pricing power supported by essential transportation infrastructure.
Strong cash generation supporting dividends, buybacks and reinvestment.
Opportunities to improve operational efficiency.
What I鈥檓 watching:
CN is a high-quality business, but it鈥檚 also capital-intensive. I鈥檒l be watching whether operating margins improve, how debt levels trend and whether free cash flow continues growing.
Valuation matters too. Even a great business can deliver disappointing returns if you overpay.
My takeaway:
CN doesn鈥檛 need explosive growth to compound shareholder value. Its competitive advantages, cash generation and essential infrastructure make it a compelling long-term business.
The question isn鈥檛 whether CN is a great company. It鈥檚 whether today鈥檚 valuation offers an attractive long-term return.
Are you buying $CNR.TO at these levels or waiting for a better entry?
Not Another Quant retweeted
The Bank of Canada鈥檚 Business Outlook Survey came out today and it looks like businesses are going to keep raising prices because of the war and tariffs.
About half of them are now planning to pass some or all of their war and tariff costs on to customers. Businesses that had larger cost increases than normal have only passed on 26% of them so far. Those businesses think it鈥檒l be 44% by this time next year.
Most say fuel and tariffs are what鈥檚 raising their costs. Wage increases look about normal and most aren鈥檛 planning to add workers.
In February 11% of businesses expected inflation above 3% over the next two years and now 41% do.
Things are probably worse now than this survey shows. 3/4 of the interviews were done before trade talks with the US broke down on August 21. Since then more businesses have seen the trade war raise their costs and they expect more inflation.
The next survey comes out January 18 and it鈥檒l show how much the trade war has changed things.
Not Another Quant retweeted
September CPI doesn鈥檛 give the BoC a reason to hike on October 28. Inflation rose to 3.1% because of gasoline but inflation excluding gasoline fell from 2.4% to 2.2%.
Trim and median are the core inflation numbers the BoC watches most and they fell to 1.8% and 1.9% YoY. MoM they rose faster over the summer but slowed in September. I don鈥檛 see higher gas prices pushing up anything else in September.
Earlier this year more businesses told the BoC they planned to raise prices but I don鈥檛 see that in September CPI yet. Canada鈥檚 new counter tariffs mostly hit things businesses buy like US steel. A lot of those goods are also made in Canada so for now the BoC doesn鈥檛 expect the tariffs to push inflation up much.
Macklem (BoC Head) said on Saturday that if oil stays around $100 he expects inflation to slowly rise above 3%.
A December hike depends on oil and whether businesses start raising prices faster. If trim and median rise as fast MoM as they did in July and August I think they hike.
Not Another Quant retweeted
Heading into $META earnings on Wednesday the Q4 revenue forecast and any 2027 spending outlook will help show whether revenue is keeping up with their AI spending.
Costs are growing faster than revenue and Q2 capex used almost all their operating cash flow. Meta can afford to spend heavily but the full payoff could be years away.
A big spending increase is already expected for 2027. If growth slows and spending comes in higher than expected investors could start questioning it more. Any Muse revenue numbers would help show how much AI is bringing in outside ads.
The expected $10B legal charge could mean Meta misses their target of beating 2025 operating income. But the charge doesn鈥檛 tell us whether AI spending is paying off.
$CRWV and $NBIS investors will want to see if Meta still plans to add more capacity. Meta already has long term deals with both so a cut back would mostly mean fewer new orders.
If Meta cuts back because returns look weak I think that would raise doubts about AI demand more broadly.
If they cut back because they can run AI more efficiently it could mean less business for CoreWeave and Nebius but it wouldn鈥檛 say much about demand for Meta鈥檚 AI products.
But cheaper AI could get used more so Meta could end up needing more capacity anyway. Meta says they still don鈥檛 have enough compute.