Milk Road Stocks
@MilkRoadStocks
Our PRO analysts find underrated stocks before the market catches on.
We called names like $MU, $NBIS and $BE all before they ran.
Join for $1 👇
160 Following 25.6K Followers
ChatGPT went from 100M weekly users to 1.2B in less than 3 years (Save this).
The milestones OpenAI has announced:
→ Nov 2023: 100M
→ Aug 2024: 200M
→ Dec 2024: 300M
→ Feb 2025: 400M
→ Mar 2025: 500M
→ Aug 2025: 700M
→ Oct 2025: 800M
→ Feb 2026: 900M
→ Jul 2026: 1B
→ Sep 2026: 1.2B
Going from 100M to 200M took about 9 months. Going from 1B to 1.2B took less than 3.
That's roughly 1 in 7 people on Earth using it every week.
Show more
This is the most insane thing from Open AI Dev Day
ChatGPT just hit 1.2 billion weekly users and 2.5M+ businesses are using its products!
They are already generating $70B ARR off this user base and are expanding their ads business with expectations of $2.5B ARR by EOY
How can you be bearish on this company?
They just grew their revenues almost 75% in the last 1.5 months, adding $30B in ARR since early August
I think OpenAI is going to be just fine and the concerns around their business struggling or defaulting is just nonsense
The revenues for AI labs continue to show up, and if they can start to generate revenue from ads or txn fees too, then I think the labs will become unstoppable companies
The only AI lab risk left is Anthropic, but seeing as they are moving ahead with IPO my guess is the numbers aren't as bad as the market is making it out to be
The aI infra trade is going higher, that's why I see from these numbers OpenAI just released
By the way, if you want to see how I'm allocating to this market, you can check out my real-time portfolio at the link in my bio
Good luck out there!
Show more
The S&P 500 has had 11 bear markets since 1950, with the average drop being ~35%, and the average bear market taking ~14 months to hit bottom.
Every drop of 20% or more, on closing prices:
1. 1956-57: -21.5%
2. 1961-62: -28.0%
3. 1966: -22.2%
4. 1968-70: -36.1%
5. 1973-74: -48.2%
6. 1980-82: -27.1%
7. 1987: -33.5%
8. 2000-02: -49.1%
9. 2007-09: -56.8%
10. 2020: -33.9%
11. 2022: -25.4%
2020 was the fastest by far. The index fell 34% in 33 days and was back at a record 5 months later.
1973-74 had the slowest comeback. After the bottom, it took almost 6 years to get back to the old high.
Show more
$10K put into the S&P 500 in October 2006 grew to about $80.8K over the next 20 years.
Miss its 10 best days and you'd have $35.9K.
That's with dividends reinvested, from Oct 26, 2006 to present.
→ Stayed invested: $80.8K
→ Missed the 10 best days: $35.9K
→ Missed the 20 best days: $21.2K
→ Missed the 30 best days: $13.9K
→ Missed the 40 best days: $9.5K
There were about 5,000 trading days in that stretch. Sitting out the best 40 of them, less than 1%, left you with less than the $10K you started with.
All 10 of the best days came during the 2008-09 financial crisis, the 2020 Covid crash or the April 2025 tariff selloff.
The biggest was Oct 13, 2008, when the index jumped 11.6%, two trading days after one of its worst drops.
Show more
“AI token usage is going 500x from here” says @Jason
Anyone arguing that AI usage has peaked or compute demand is maxed out is completely lost and obviously doesn’t use the technology
We’re still at something like 2% of people in corporate America using AI today AND both OpenAI and Anthropic haven’t even launched agents at scale yet
We have only touched the surface of what’s possible with AI and the usage will not just grow (linear) it will ACCELERATE (exponential) from here
I feel like this is hard for people to wrap their heads around but if you want to understand where this market is going, you need to understand this trend.
It’s not stopping, it’s getting much bigger, and the mid-terms, rates or anything else will not stop this
It doenst mean the market goes up in a straight line, but as an investor don’t get lost in the noise. This is the biggest technological change in our lifetime
Think MUCH MUCH BIGGER
Show more
Alphabet $GOOGL spent $61.1B on research and development last year, more than Intel's entire 2025 revenue of $52.9B!
R&D spending in each company's latest fiscal year:
1. Alphabet $GOOGL: $61.1B
2. Meta $META: $57.4B
3. Microsoft $MSFT: $35.6B
4. Apple $AAPL: $34.6B
5. Nvidia $NVDA: $18.5B
6. Intel $INTC: $13.8B
7. Broadcom $AVGO: $11.0B
8. Oracle $ORCL: $10.3B
9. AMD $AMD: $8.1B
10. Tesla $TSLA: $6.4B
Meta's R&D bill jumped 31% in a year, from $43.9B to $57.4B, and it also spent more than Intel brought in.
Nvidia spends less than a third of what Alphabet does on R&D, even though its revenue is more than half of Alphabet's.
Note: Amazon doesn't break out R&D as its own line, so it isn't on the list.
Show more
My estimate is SpaceX spends around $470B in the next 5 quarters… if he’s to get to 13GW
Elon just stepped in and said their capex efficiency is underestimated
My estimate is based on $30b/GW and then $45B with Vera Rubin’s
Not sure how Elon plans to get this cheaper outside of casting his own turbine blades… but looking forward to the next earnings call to learn more
If they can build 1GW for under $30B in 60-90days and immediately lease it for $50-$60B/year, the numbers here are insane
Show more
ChatGPT went from 100M weekly users to 1.2B in less than 3 years (Save this).
The milestones OpenAI has announced:
→ Nov 2023: 100M
→ Aug 2024: 200M
→ Dec 2024: 300M
→ Feb 2025: 400M
→ Mar 2025: 500M
→ Aug 2025: 700M
→ Oct 2025: 800M
→ Feb 2026: 900M
→ Jul 2026: 1B
→ Sep 2026: 1.2B
Going from 100M to 200M took about 9 months. Going from 1B to 1.2B took less than 3.
That's roughly 1 in 7 people on Earth using it every week.
Show more
This is the most insane thing from Open AI Dev Day
ChatGPT just hit 1.2 billion weekly users and 2.5M+ businesses are using its products!
They are already generating $70B ARR off this user base and are expanding their ads business with expectations of $2.5B ARR by EOY
How can you be bearish on this company?
They just grew their revenues almost 75% in the last 1.5 months, adding $30B in ARR since early August
I think OpenAI is going to be just fine and the concerns around their business struggling or defaulting is just nonsense
The revenues for AI labs continue to show up, and if they can start to generate revenue from ads or txn fees too, then I think the labs will become unstoppable companies
The only AI lab risk left is Anthropic, but seeing as they are moving ahead with IPO my guess is the numbers aren't as bad as the market is making it out to be
The aI infra trade is going higher, that's why I see from these numbers OpenAI just released
By the way, if you want to see how I'm allocating to this market, you can check out my real-time portfolio at the link in my bio
Good luck out there!
Show more
Alphabet $GOOGL spent $61.1B on research and development last year, more than Intel's entire 2025 revenue of $52.9B!
R&D spending in each company's latest fiscal year:
1. Alphabet $GOOGL: $61.1B
2. Meta $META: $57.4B
3. Microsoft $MSFT: $35.6B
4. Apple $AAPL: $34.6B
5. Nvidia $NVDA: $18.5B
6. Intel $INTC: $13.8B
7. Broadcom $AVGO: $11.0B
8. Oracle $ORCL: $10.3B
9. AMD $AMD: $8.1B
10. Tesla $TSLA: $6.4B
Meta's R&D bill jumped 31% in a year, from $43.9B to $57.4B, and it also spent more than Intel brought in.
Nvidia spends less than a third of what Alphabet does on R&D, even though its revenue is more than half of Alphabet's.
Note: Amazon doesn't break out R&D as its own line, so it isn't on the list.
Show more
Lyn Alden: a split Congress after the midterms keeps "the 7% of GDP deficits" in place "for the next two years."
Gridlock rules out real spending cuts or tax hikes. She says it also rules out big stimulus or new tax cuts, and probably caps defense spending.
US money supply has grown about 7% a year on average. It's growing slightly under 6% now, and she sees anywhere from 5% to 8% as "business as normal."
"There's very little portfolio decision making I'm doing based on election outcomes."
FT @LynAldenContact @BitcoinJesusETH @milkroaddaily.
Show more
Warren Buffett told Americans to buy stocks in October 2008, and the market fell another 28%.
His op-ed in The New York Times on Oct. 17, 2008 was titled "Buy American. I Am."
He wrote that he was buying US stocks in his personal account, which until then had held nothing but US government bonds. He was following a rule he'd used for years, to "be fearful when others are greedy, and be greedy when others are fearful."
He also admitted he didn't have "the faintest idea" whether stocks would be higher or lower a month or a year later.
The S&P 500 closed at 940.55 that day. It kept falling until March 9, 2009, when it bottomed at 676.53.
By June 2009 it was back above where it was when he wrote the piece, and it's gone up about 8x since.
Show more
Berkshire's portfolio is basically a bet on American brands, and half of it sits in just three stocks (save this).
Its largest position is $AAPL at 22.0%, followed by $AXP at 17.1% and $KO at 10.9%.
Other notable names:
$GOOGL: 9.4%
$BAC: 9.2%
$CVX: 4.7%
$OXY: 4.3%
$CB: 3.9%
$MCO: 3.7%
$GOOG: 3.2%
$KHC: 2.6%
The big move in Q2 was Alphabet. Berkshire added about 24.5M $GOOGL shares (+45.2%) and about 23.6M $GOOG shares (+658.3%). Count both share classes and Alphabet is now 12.6% of the portfolio, more than Coca-Cola.
It also added about 17.5M $DAL shares (+44.0%) and trimmed $BAC by about 30.2M shares (-5.9%). The deepest cuts were $COF (-58.0%) and $NUE (-52.5%).
We track 20+ Insider Portfolios, including Nancy Pelosi, Gavin Baker and Donald Trump.
Want to get an email whenever they make a trade?
Simply sign up for a free Milk Road account and click the bell notification icon on your preferred Insider page 👇
Show more
Cisco $CSCO closed at $80.06 on March 27, 2000, and it took more than 25 years to close above that again (Save this).
That month it briefly passed Microsoft as the most valuable company in the world.
Then the dot-com bust hit. By October 2002 the stock was at $8.60, down 89%.
The business kept growing the whole time. Cisco's yearly revenue went from $18.9B in fiscal 2000 to $56.7B in fiscal 2025.
The stock finally closed back above $80.06 on Dec 10, 2025, about 25 years and 8 months after the peak.
Show more
Ben Horowitz: The dot-com bubble died from too little demand...
But AI compute is sold out until "beyond 2028," according to Amazon $AMZN CEO Andy Jassy.
Jason Calacanis asked if CoreWeave's crash pops the AI bubble.
Ben: "We don't really have a lack of demand in AI so I think the dynamics will be different."
But a16z won't back neoclouds. At LoudCloud, bandwidth fell from $1,500 a megabit in 1999 to $50 by 2001.
"You borrowed money to get it, you're dead," he says.
Is Ben right that demand makes this different from 2000?
FT @bhorowitz @ajassy @Jason @theallinpod.
Show more
SpaceX $SPCX raised $75B in its IPO, more than the next three biggest US IPOs combined.
The biggest US IPOs ever, by money raised:
1. SpaceX $SPCX: $75.0B (2026)
2. SK hynix $SKHY: $26.5B (2026)
3. Alibaba $BABA: $21.8B (2014)
4. Visa $V: $17.9B (2008)
5. Enel: $16.5B (1999)
6. Facebook $META: $16.0B (2012)
7. General Motors $GM: $15.8B (2010)
8. Deutsche Telekom: $13.0B (1996)
Alibaba held the record for almost 12 years. SpaceX broke it in June with more than 3x as much, and a month later SK hynix passed Alibaba too.
Show more
Brad Gerstner: Anthropic's IPO could draw over $500B of demand, compared to SpaceX's IPO which drew about $350B.
The worry, according CNBC: Q3 revenue grew 45% to $16.8B, down from ~140% earlier this year.
Gerstner: "None of the revenue would have happened this year without Claude Code. Claude Code became better than any software engineer at writing software."
He says models already finished in the labs will "unlock other products like Claude Code."
Can new models reaccelerate Anthropic, or is 45% the start of a cooldown?
FT @altcap @AnthropicAI @HalftimeReport.
Show more
If self-driving cars, rockets and gene editing pay off, so does Cathie Wood.
After $TSLA, her next biggest positions are $SPCX (SpaceX) at 5.2% and $TEM (Tempus AI) at 4.0%.
Other notable names:
$NVDA: 3.2%
$CRCL (Circle): 3.2%
$TWST (Twist Bioscience): 3.0%
$CRSP (CRISPR Therapeutics): 2.9%
$PLTR: 2.8%
$TXG (10x Genomics): 2.8%
$COIN: 2.6%
Cathie recently added about 2% to 3% more shares to most of her holdings. The biggest adds were $JOBY (air taxis), up 12.6% or about 885K shares, and $AVAV (drones), up 11.0% or about 77.2K shares.
She trimmed $AMD by about 20.4K shares and $TSM by about 16.9K shares, both down 4.2%, while still adding to $NVDA (+1.8%).
We track 20+ Insider Portfolios, including Nancy Pelosi, Gavin Baker and Donald Trump.
Want to get an email whenever they make a trade?
Simply sign up for a free Milk Road account and click the bell notification icon on your preferred Insider page 👇
Show more
Lyn Alden: a split Congress after the midterms keeps "the 7% of GDP deficits" in place "for the next two years."
Gridlock rules out real spending cuts or tax hikes. She says it also rules out big stimulus or new tax cuts, and probably caps defense spending.
US money supply has grown about 7% a year on average. It's growing slightly under 6% now, and she sees anywhere from 5% to 8% as "business as normal."
"There's very little portfolio decision making I'm doing based on election outcomes."
FT @LynAldenContact @BitcoinJesusETH @milkroaddaily.
Show more
Nobody is better positioned to dominate AI right now than SpaceX (+Tesla)
For 2 reasons:
1. They're about to have the most compute in the world
2. They have multiple products to monetize their compute (hyperscalers don't even have this)
Sure, they can lease 1GW/$50B+ or they can route it to Grok/GrokBot and generate $100B+/year like the frontier models are experiencing (they already have 800k+ paying GrokBot users as of 2 weeks ago)
Elon can also route that compute to FSD, Robotaxis and eventually humanoids at Tesla, all of which will require significant amounts of training and inference compute as they scale up into the millions (FSD already close to 2 million users)
This is the big reason why Tesla needs to merge with SpaceX, it needs access to loads of compute as its products scale. Rather than rent it from SpaceX if they merge than they simply receive it at cost and have the ability to switch that compute from Tesla products back to short-term leases if needed
So effectively, Elon is building 1GW of compute for around $30B ($45B with Vera Rubins) and immediately able to lease it for $50B/year, plug it into Grok for $100B/year or use it for his fleet of autonomous robots
Oh, and this doesn't even start to consider the fact that the bulk of the worlds compute will end up coming from Space, and SpaceX has about a 10 year lead on the space market, in terms of cost effectiveness and load size with Starship
This will only push their position/moat even further ahead than it is today. I just dont see how SpaceX doesn't become the worlds most valueable company in the next few years
I still believe this stock is significantly undervalued and will continue to buy on pullbacks. I originally got our Milk Road PRO members in $SPCX at $115. If you want to see how I'm playing this market and check out my real-time portfolio you can do so here:
Show more
Cisco $CSCO closed at $80.06 on March 27, 2000, and it took more than 25 years to close above that again (Save this).
That month it briefly passed Microsoft as the most valuable company in the world.
Then the dot-com bust hit. By October 2002 the stock was at $8.60, down 89%.
The business kept growing the whole time. Cisco's yearly revenue went from $18.9B in fiscal 2000 to $56.7B in fiscal 2025.
The stock finally closed back above $80.06 on Dec 10, 2025, about 25 years and 8 months after the peak.
Show more
Ben Horowitz: The dot-com bubble died from too little demand...
But AI compute is sold out until "beyond 2028," according to Amazon $AMZN CEO Andy Jassy.
Jason Calacanis asked if CoreWeave's crash pops the AI bubble.
Ben: "We don't really have a lack of demand in AI so I think the dynamics will be different."
But a16z won't back neoclouds. At LoudCloud, bandwidth fell from $1,500 a megabit in 1999 to $50 by 2001.
"You borrowed money to get it, you're dead," he says.
Is Ben right that demand makes this different from 2000?
FT @bhorowitz @ajassy @Jason @theallinpod.
Show more
Through all the fears around Anthropic and OpenAI
It's pretty clear they will both hit $100B ARR each by the end of this year
That's a 10x by Anthropic and 5x by OpenAI in 1 year!
I'm sorry doomers but the internet companies never had this type of growth at the application layer in the late 90s
Sure, the infra stocks did well, similar to the infra stocks this time around with AI. But the difference here is that the application layer is printing money too.
Anthropics revenue jumped from $11.6B in Q2 to ~$16.8B in Q3, and thats during a slow September before the new models launched. October has been a very big month for token usage so we should expect $20B+ for Q4
And ever since OpenAI launched Codex and Astra their revenue have been on an absolute tear as well
But we're going to close out this year with $200B ARR from just 2 companies, in addition to the revenues coming from GrokBut, Muse, and more
To compare today to the .com bubble makes 0 sense
That makes has their panties in a knot anyway though and continues to sell off. I think there will be a violent v-shaped recovery soon when everyone realizes that the AI revenue is there to support the build out
I dont know exactly when it will happen, but it will happen
Show more
If self-driving cars, rockets and gene editing pay off, so does Cathie Wood.
After $TSLA, her next biggest positions are $SPCX (SpaceX) at 5.2% and $TEM (Tempus AI) at 4.0%.
Other notable names:
$NVDA: 3.2%
$CRCL (Circle): 3.2%
$TWST (Twist Bioscience): 3.0%
$CRSP (CRISPR Therapeutics): 2.9%
$PLTR: 2.8%
$TXG (10x Genomics): 2.8%
$COIN: 2.6%
Cathie recently added about 2% to 3% more shares to most of her holdings. The biggest adds were $JOBY (air taxis), up 12.6% or about 885K shares, and $AVAV (drones), up 11.0% or about 77.2K shares.
She trimmed $AMD by about 20.4K shares and $TSM by about 16.9K shares, both down 4.2%, while still adding to $NVDA (+1.8%).
We track 20+ Insider Portfolios, including Nancy Pelosi, Gavin Baker and Donald Trump.
Want to get an email whenever they make a trade?
Simply sign up for a free Milk Road account and click the bell notification icon on your preferred Insider page 👇
Show more
Elon Musk: once there's a million times more digital intelligence than biological, "it's hard to imagine humans being in control."
"I'd probably prefer it if humans were always in control," though he says that depends on the AI.
In the million-to-one scenario, "I hope the AI is nice to us."
Walter Isaacson brought up Larry Page once calling Elon a speciesist for wanting humans in charge. Elon admitted there have been "a few times where I thought maybe Larry Page was right."
"AI and robots are going to decide the future."
"I find it unlikely that humans would remain in control if there's a million times more digital intelligence than there is biological intelligence."
FT @elonmusk @WalterIsaacson.
Show more
Elon Musk: AI and robots are the only way to fix the US debt.
He says no one in Washington wants to cut spending, and debt interest now tops military spending.
"Digital superintelligence probably improves the output of goods and services by 20 to 30%, something like that." Call it $30T a year.
With robots and a mass driver on the moon, Elon goes bigger: "then you can get to, I think, potentially a million times the size of the Earth economy."
Can $TSLA robots outgrow the US debt?
FT @elonmusk @WalterIsaacson
Show more
Andy Jassy: revenue commitments for Amazon's Trainium AI chips are up more than 20x in a year
On Amazon's earnings call last night:
"Our chips business continues to grow triple-digit percentages year over year, and revenue commitments for Graviton and Trainium have each increased more than 20 times year over year."
Jassy also said customers keep pushing Amazon to offer Trainium outside AWS, and Amazon plans to do that late in 2027.
Oracle said this week it will put Google's TPUs in its cloud in 2027, and now Amazon wants Trainium outside AWS on a similar timeline.
Two of Nvidia's biggest customers are turning the AI chips they built for themselves into products they sell to other companies.
Show more
WOW! Nine companies are on track to spend about $1 TRILLION on capex this year, with $AMZN topping the list at about $220B.
That spending pushed its free cash flow over the past 12 months to -$42.4B, and Andy Jassy told investors last night that 2027 will be meaningfully higher.
$GOOGL, $MSFT and $META add roughly $510B between them. $SPCX, Samsung, $TSM, $CRWV and $NBIS make up the other $270B or so.
The other side of that $1T is revenue for the companies selling the chips, memory, power equipment and construction.
Show more
Why is the stock market getting NUKED right now?
Google -12.6%, Amazon -7%, Intel -25%, Nasdaq futures down 2%+... and it's happening after these companies posted some of the best cloud growth we've EVER seen
Here's what you need to know as an investor:
1. Google missed on revenue and EPS, search growth slowed to 12%, and free cash flow came in at NEGATIVE $9.07B after $51.4B of capex in a single quarter. The stock had its worst day in 20+ years and lost about $500B in value
2. Amazon piled on last night. AWS grew 39% (fastest in nearly 5 years) and Amazon posted its highest operating margin ever... but trailing 12-month free cash flow went to NEGATIVE $42.4B, from +$14.8B a year ago. And it signaled capex goes UP again next year
3. Intel finished the job. Weak Q4 guide, a PC slump, and a warning that it'll burn cash this year. Stock -25% premarket, dragging every chip name down with it.
One thing to note from all of this:
Nobody is worried about AI DEMAND. Demand is the one thing every company confirmed. Google Cloud more than doubled, Azure grew 48%, AWS is accelerating and still can't build capacity fast enough
So to be clear, demand is NOT the issue. Any bear who's focused on demand should be muted, they are out to lunch
The fear right now is the BILL
The 4 big hyperscalers free cash flow was -$26B combined. 3 of the 4 are now spending more than their businesses bring in
So they're borrowing more and more to fill that gap. Amazon's long-term debt went from $66B to $173B in just 9 months
These companies used to be the biggest cash machines on earth, buying back their own stock. Now they're issuing bonds to keep building. THAT'S the shift the market is repricing and struggling to wrap their head around... even though FCF is expected to go positive by the end of next year, and significantly higher into 2028
But here's the ironic part. Every dollar of that capex is REVENUE for somebody else: the chip, memory and power companies building this stuff out. Yet $MU is still down ~5% this morning. The market is selling the spenders AND the companies getting paid... it makes no sense!
I'm seeing people on X call this the AI bubble popping. Once again, I don't see it. Bubbles pop when demand dries up. Here demand is accelerating and these companies are spending MORE because they can't build fast enough. What's being repriced is how long investors are willing to wait for the cash
I continue to believe these dips are for buying and we go much higher as the market realizes that this negative fcf is only short-term and the revenue is showing up in many different places from AI
This is not a bubble, this is one of the greatest investing opportunities of our lifetime and the market continues to give us gifts at these prices. Take a look at what dips I'm buying and what my real-time portfolio looks like here:
Don't forget to give me a follow @kylereidhead for more insights on AI and markets
Show more
