The only way things get better for the market from here on is when Trump stops the Iran war
It has dragged on with no visible outcome
2
7
508
$NVDA Today it's make it or break it
( Although already feel pretty broken and broke )
Let's find out!
5
566
On days like this glad I had allocated $ZETA as my largest position in the portfolio
The stock is up 81% YTD and on track to keep growing even more next year
Still believe $ZETA to 10x faster than $NBIS
2
37
5,969
One of the reasons why sentiment was so positive around Neoclouds back in 2025 was before of the huge deals they were signing with big customers.
Each major deal with a hyperscaler moved $NBIS and $IREN more the stock upwards 20% to 30% although the pricing wasn't that attractive.
Now we don't see such headlines since they are getting better deals from smaller AI labs or other enterprise customers although the deal size may not be huge.
I think to flip the sentiment back they should start announcing such deals, unit economics of it, prepayments etc if they can. The only time you hear of these deals that too at a high level are the earnings call and doesn't get much attention
Feel this will do a world of good for the sentiment especially looking at how things are looking doom and gloom although that's not the case for all neoclouds.
It also lets shareholders know that yes the AI build out is on track despite the huge capex commitments made and there is a reason for those commitments ( aka these lucrative deals we are signing with these customers and the build out for them )
Ok Neoclouds that haven't matured might be done likes of Nscale, Lambda and others
Consolidation in works
Bullish $NBIS $IREN $CRWV
NEW: We continue to learn more and more about how Nvidia's plans to finance the AI boom were running into issues—even before Jensen was ousted.
Nvidia’s Days as an Open Checkbook for the AI Economy Are Numbered.
w/ @RWhelanWSJ & @KateClarkTweets for @WSJ
more tips? Signal me: 215-896-1307
wsj.com/tech/ai/nvidias-days…
17
4,876
People need to understand that supply will catch up to demand eventually
Whether that happens in 2028 or 2029 that is yet to be seen but that will happen so Hyperscalers or neoclouds won't have the same pricing power like we are having today
Why I feel it's sooner? After watching Starship getting caught by mechazilla with it's chopsticks it would be unwise to underestimate Elon if he is aiming to bring excess of 20GW of compute online $SPCX and not to mention he plans to put 1GW of orbital compute online from 2028
He may not hit that number but let's say he brings online even half of that 10GW next year that itself is massive supply coming online along with other Hyperscalers and neclouds scaling their capacity too
All I'm trying to say is foolish to assume neoclouds keep getting premium pricing 5 years down the road and add that into your valuation...maybe we are lucky and still get premium pricing and that would be the best case scenario but as someone who is trying to share a valuation 5 years down the road it needs to be conservative so it doesn't mislead people
Replying to @Kajiastic
This just validates that most people have no idea how large the demand for ai really is. We are years away from non-tech companies implementing the same ai workflows that leading tech companies have
2
4
2,253
I'm a $NBIS bull but people should be careful throwing around such extravagant price numbers casually.
25M/MW over the next 5 years is unrealistic... supply will catch up to demand by 2028 with $SPCX and $META aggressively building out compute capacity and then companies won't have the attractive pricing power anymore like they do today
This also doesn't include possible delays in build out
Below accounts for flawless execution in the best case supply/demand scenario
My 2030 price target for $NBIS is $3,000.
Here’s the math.
Nebius expects to deploy at least 5 GW of active compute capacity by 2030.
And the economics keep improving:
• Annual revenue per MW was previously around $10M.
• Long-term contracts now exceed $25M per MW.
• Short-term contracts can generate $40–50M per MW.
And the trend is accelerating.
The gap between AI compute supply and demand is expected to widen for at least another 18–24 months.
For 2030, I’m using a conservative assumption of $25M in annual revenue per MW.
5,000 MW × $25M = $125B in annual revenue.
Now look at profitability.
AWS currently operates at approximately 40% operating margins, suggesting a net margin of roughly 30%.
For Nebius, I’m assuming a 25% net margin at scale.
$125B revenue × 25% = $31.25B in annual net income.
And 2030 is not the finish line.
Management expects to add at least 1.5–2 GW of additional compute capacity every year.
On a 5 GW base, that represents 30–40% annual capacity growth.
A company generating more than $31B in annual profits while expanding its capacity at that rate commands a premium valuation.
I’m applying a 30x P/E multiple.
$31.25B × 30 = approximately $940B in market capitalization.
I’m also accounting for roughly 25% additional share dilution by 2030, reflecting the enormous capital investment required to support this expansion.
The result?
A $3,000 price target for $NBIS by 2030.
Nearly $1 trillion in market capitalization.
And a business that could still be expanding its compute capacity by 30–40% annually.
$3,000 per share is not a fantasy. It’s what the numbers point to.
14
42
12,382
$IREN at $18 is priced like the ramp won't happen.
I ran the same model I did for $NBIS: revenue, capex, net income and FCF through 2032. Video attached.
WHAT IREN HAS GUIDED
• $4B+ ARR, now targeted for January 2027 (could run to end of February)
• 0.8 GW of data center capacity by end of 2027
• ~$5B capex last quarter, ~$15B planned for the rest of FY27
• $20.3B backlog, about two-thirds investment grade
MY ESTIMATE AFTER 2027
• +0.5 GW a year through 2029, then +0.3 GW
• 2.7 GW by 2032
REVENUE
• 2027: $5.5B
• 2028: $17.4B (+216%)
• 2029: $26.7B (+53%)
• 2030: $34.5B (+29%)
• 2032: $45.5B
CAPEX
• 2027: $25B
• 2028: $31.5B
• 2029: $33.1B
• 2030: $21.8B
NET INCOME
• 2027: -$1.3B
• 2028: +$3.1B
• 2032: +$5.8B
FREE CASH FLOW
• 2028: -$20.3B
• 2029: -$15.7B
• 2030: +$0.7B
• 2032: +$6.7B
Similar shape to Nebius. Capex runs far ahead of revenue through 2029, GAAP net income turns positive in 2028, and FCF turns positive in 2030 once the build slows.
WHY I THINK IT'S THE CHEAPEST OF THE THREE
EV to near-term run-rate (13 Nov close):
• $IREN ~3.0x
• $NBIS ~5.8x
• $CRWV ~6.4x
Its latest 5-year deal was ~$22M per IT-MW a year, with 90% of GPU capex prepaid.
WHAT CAN BREAK IT
• Funding: ~$90B of capex over 2027-29 from a company worth under $8B. $6B of this year's plan still depends on new financing. Yes prepayments help but would need to see more of it
• Dilution: cheap on EV doesn't help if the build is paid for in stock.
• Delays: the $4B ARR target has already slipped once.
ASSUMPTIONS
• ~$19M revenue and ~$60M capex per MW
• 72% EBITDA margin at scale (29% today)
• Beyond 2027 it's my estimate. IREN gives no revenue guidance.
Wow $NVDA .... could he come back?
7
643
Follow up on my $NBIS model. I've now split revenue into 4 segments:
-Hyperscalers
-AI cloud
-Inference
-Asset-light partnerships.
With inference and asset light included, GAAP net income turns positive in 2028 instead of 2029, and FCF turns positive in 2030.
You see capex cut in half from 2030 as I project they will reduce the build out from 1GW to 400MW as supply catches up to demand. Could happen sooner too
$NBIS PROFITABILITY PATH (GAAP) - 2029
$NBIS dropped 16% today on 2027 capex guidance of $40–45B, sparking fears that spending will permanently outpace revenue.
I ran the numbers assuming +1 GW/year capacity growth, +10% annual capex, and a conservative $18M revenue/MW on new builds:
• GAAP Net Income: Turns positive in 2029
• Revenue vs. Capex: Revenue overtakes capex in 2030 assuming 1GW build out per year until 2032
2029 Model Snapshot ($18M/MW):
• Revenue: $51.6B ($36B from 2 GW at full run-rate + $7.2B ramping GW + $8.4B legacy)
• EBITDA (62%): $32.0B
• D&A (5-yr server depreciation): -$25.7B
• Interest: -$4.7B
• Pre-tax Income: +$1.7B (shielded by NOLs) vs. $54.5B capex
The entire thesis comes down to monetization per MW:
• At $25M/MW, profitability hits in 2028.
• At $13M/MW (hyperscaler heavy mix), there’s still no profit by 2032 which is unlikely
2029 margins will be razor-thin (~3%), and FCF stays negative as long as the aggressive buildout continues
Bottom line is that people are too focused on the near term and aren't able to extrapolate hoe this will look like over the next few years.
Hopefully below animation helps
12
2,388
$NBIS PROFITABILITY PATH (GAAP) - 2029
$NBIS dropped 16% today on 2027 capex guidance of $40–45B, sparking fears that spending will permanently outpace revenue.
I ran the numbers assuming +1 GW/year capacity growth, +10% annual capex, and a conservative $18M revenue/MW on new builds:
• GAAP Net Income: Turns positive in 2029
• Revenue vs. Capex: Revenue overtakes capex in 2030 assuming 1GW build out per year until 2032
2029 Model Snapshot ($18M/MW):
• Revenue: $51.6B ($36B from 2 GW at full run-rate + $7.2B ramping GW + $8.4B legacy)
• EBITDA (62%): $32.0B
• D&A (5-yr server depreciation): -$25.7B
• Interest: -$4.7B
• Pre-tax Income: +$1.7B (shielded by NOLs) vs. $54.5B capex
The entire thesis comes down to monetization per MW:
• At $25M/MW, profitability hits in 2028.
• At $13M/MW (hyperscaler heavy mix), there’s still no profit by 2032 which is unlikely
2029 margins will be razor-thin (~3%), and FCF stays negative as long as the aggressive buildout continues
Bottom line is that people are too focused on the near term and aren't able to extrapolate hoe this will look like over the next few years.
Hopefully below animation helps
$NVDA The comeback would be epic
3
406



















