Neal Arnold

[Neal Arnold @_SilverFox1]

@TheBigBerbowski $Spcb Supercom winning in electronic monitoring fits the bill. Some good write ups on x. Very undervalued

[Neal Arnold @_SilverFox1]

@yianisz Bb

[Stephen D'Addario @stevdaddario]

SuperCom $SPCB: Heads We Win More, Tails We Still Win.

Thesis was stronger pre ISAP V award, but core thesis logic still applies with the 1x1 award structure. Long.

https://t.co/koVAglbeEp

#Investing #FinTwitt

[Misunderstood Multibaggers @meeijer]

Is $NBIS still a buy? – Here’s my valuation model and price targets

There’s been a lot of debate recently about whether $NBIS is still a buy. To answer this question, I’ve built a valuation model.

First, let’s get the numbers straight.

The Microsoft deal will deliver $17.4B (potentially $19.4B if Microsoft acquires additional services or capacity under the agreement) in revenue for Nebius ($NBIS) over 5+ years, running through 2031. Microsoft will start receiving infrastructure from the New Jersey site later this year. To complete the buildout, Nebius will use a mix of its own capital and capital from the contract.

Assuming the remaining 300 MW in New Jersey costs $12M per MW (since these are high-end GPU clusters, not bare metal), and Nebius funds $2B from its own balance sheet and $1.5B via the Microsoft deal, the total project would generate roughly $15.9B (potentially $17.9B) in revenue over the following 5 years. That translates into about $3.18B in ARR from the deal alone ($3.58B at the high end).

In short, the deal will generate around $10M in ARR per MW, I’ll revisit this point later. Overall, the Microsoft deal demonstrates both credibility and stability, which should be reflected in the stock’s valuation, more on that later in this post.

Now, to estimate Nebius’ total ARR in the coming years, I’ll project their active power (the MW they’re actually monetizing, not just installed/connected):

For 2025, I expect active power to reach 100 MW, as management guided to 220 MW of connected power but only 100 MW of active power by year-end. By the end of 2026, I estimate around 485 MW of active capacity. Here’s how that breaks down:

• Iceland: 10 MW
• Finland: 75 MW
• New Jersey: 200 MW
• Paris: 5 MW
• Kansas: 30 MW
• United Kingdom: 15 MW
• New greenfield sites (management mentioned that multiple new sites will be announced potentially supporting hundreds of MW each): 150 MW

Total = 485 MW

By the end of 2027, I expect all current and in the near future announced datacenters to be fully built out:

• Iceland: 10 MW
• Finland: 75 MW
• New Jersey: 350 MW
• Paris: 10 MW
• Kansas: 40 MW
• United Kingdom: 50 MW
• New greenfield sites: 500 MW

Total = 985 MW

By 2028, I expect Nebius to announce and complete several additional greenfield sites, doubling capacity to around 1.6 GW. By 2029, capacity could reach roughly 2.5 GW. Assuming continued hyperscaler demand through 2030, I see total capacity reaching around 3.5 GW.

Here’s a quick summary of my base case capacity projections:

2025: 100 MW
2026: 485 MW
2027: 985 MW
2028: 1.6 GW
2029: 2.5 GW
2030: 3.5 GW

Now, let’s look at my bear case and bull case.

Bear case:

In this scenario, demand for AI infrastructure slows meaningfully. Model efficiency improves faster than expected, meaning hyperscalers and AI startups can achieve similar performance with fewer GPUs. Hyperscalers also shift focus toward renting cheaper bare metal capacity rather than high-end GPU clusters, reducing demand for Nebius’ premium infrastructure. Funding conditions remain tight, limiting Nebius’ ability to finance new greenfield projects, and pricing pressure across the industry gradually compresses margins.

2025: 100 MW
2026: 300 MW
2027: 550 MW
2028: 750 MW
2029: 1 GW
2030: 1.3 GW

Bull case:

In this scenario, AI demand accelerates significantly, driven by large-scale model training and edge AI deployment. Hyperscalers (Microsoft, Google, Amazon, etc.) continue to expand capex every quarter, pushing the limits of available GPU capacity. Nebius becomes a core infrastructure partner for multiple hyperscalers thanks to its cost-efficient power sourcing, flexible datacenter footprint, and reliability. The company executes flawlessly on buildouts, keeps utilization high, and raises capital efficiently through debt and equity markets to fund expansion. AI adoption broadens across industries, ensuring sustained demand for high-performance compute infrastructure.

2025: 100 MW
2026: 600 MW
2027: 1.5 GW
2028: 3 GW
2029: 6 GW
2030: 9 GW

Now, assuming Nebius maintains $10M ARR per MW through 2026, then $9M in 2027–2028, and $8M in 2029–2030 (as competition increases and pricing normalizes), we get the following ARR outlooks:

ARR projections (Bear | Base | Bull):

2025: $1B | $1B | $1B
2026: $3B | $4.85B | $6B
2027: $4.95B | $8.87B | $13.5B
2028: $6.75B | $14.4B | $27B
2029: $8B | $20B | $48B
2030: $10.4B | $28B | $72B

Note: These numbers are estimates, not facts, but they help visualize how Nebius could scale and what kind of ARR they might generate.

To reach 9 GW (bull case), Nebius would likely need close to $100B in capex, which seems highly unlikely unless the stock price appreciates substantially, allowing for equity raises, and Nebius sells off their Subsidiaries. The base case is far more realistic, with modest capex funded through a mix of debt and subsidiary sales.
Valuation

For simplicity, I’ll focus value Nebius with the P/ARR ratio and assume share count grows modestly each year.

Bear case valuation (shares +10% YoY):

2025: $1B * 25 = $25B → $100/share (16% downside)
2026: $3B * 12 = $36B → $131/share (10% upside)
2027: $4.95B * 10 = $49.5B → $163/share (38% upside)
2028: $6.75B * 10 = $67.5B → $203/share (71% upside)
2029: $8B * 9 = $72B → $197/share (66% upside)
2030: $10.4B * 8 = $83.2B → $206/share (74% upside)

Base case valuation (shares +6% YoY):

2025: $1B * 30 = $30B → $120/share (2% upside)
2026: $4.85B * 15 = $73B → $275/share (132% upside)
2027: $8.87B * 13 = $115B → $409/share (245% upside)
2028: $14.4B * 13 = $187B → $628/share (430% upside)
2029: $20B * 11 = $220B → $880/share (596% upside)
2030: $28B * 11 = $308B → $919/share (676% upside)

Bull case valuation (shares +3% YoY):

2025: $1B * 35 = $35B → $140/share (18% upside)
2026: $6B * 20 = $120B → $465/share (293% upside)
2027: $13.5B * 16 = $216B → $815/share (588% upside)
2028: $27B * 16 = $432B → $1,582/share (1,235% upside)
2029: $48B * 14 = $672B → $2,391/share (2,018% upside)
2030: $72B * 12 = $864B → $2,979/share (2,414% upside)

Before you assume I’ve gone crazy, no, not yet haha. The bull case is aggressive and unlikely to materialize. I personally think reality will land slightly below my base case, which still implies significant upside if Nebius executes on capacity growth and closes additional hyperscaler deals.

I believe $NBIS remains a great buy today, and an even stronger one for those investing with a long-term horizon. NFA.

Thanks for reading!

[Neal Arnold @_SilverFox1]

https://t.co/4KWdygHNSY

[Neal Arnold @_SilverFox1]

@mvcinvesting $SPCB

[Girth Master @GirthMaster21]

Regarding the upcoming ICE contract:

This business is growing at 33% YOY CAGR. $spcb

[JC @j16425646]

Here is a link to the CEO of $SPCB's presentation yesterday. Skip to 20:50, when Trabelsi shockingly acknowledges: "Of course we have been approached by private equity and competitors...doesn’t mean that we are going to take their offers" $GEO $AXON

https://t.co/aZlYRz7jas

[M. V. Cunha @mvcinvesting]

JUST IN: Arete Research raised its Price Target on $NBIS from $128 to $184 and keeps a Buy rating 🟢

Back in April, some of us were loading up under $20. Fast-forward a few months, and we’re looking at a PT nearly 10x higher.

Let that sink in.

[JC @j16425646]

$SPCB is about to go crazy. Microcap leader in offender electronic monitoring (countercyclical, as higher unemployment is tied to higher incarceration) trading at 4.5X run-rate eps, gaining share, high short interest, and acquisition candidate (i.e. $GEO, $AXON). #nfp

[Neal Arnold @_SilverFox1]

@georgegalloway The global sumud flotilla has nearly reached Gaza. Citizens of the world step in where our despicable governments have failed.

[Neal Arnold @_SilverFox1]

They sail tomorrow. Peace and love to all.

Global Sumud Flotilla - Global https://t.co/tQpmxrBe30

[JC @j16425646]

@mvcinvesting $SPCB. $40mn market cap, share gainer in robust market for offender electronic monitoring, recurring revenue model, surging margins and cash flow, gaining massive share in the US and Europe. Comps exited for and avg of >5X EV/sales. $SPCB at ~1X sales and ~4X normalized EPS.

[Neal Arnold @_SilverFox1]

@mvcinvesting $SPCB

[Neal Arnold @_SilverFox1]

@Traderlife_ $SPCB Conservative valuation gives $25 a share over 2.5x from yesterday's close

[Neal Arnold @_SilverFox1]

$SPCB should be trading at 2-3x per share on these earnings. Great risk to reward opportunity

SuperCom Reports Record Net Income of $5.3 Million and Non-GAAP EPS of $1.84 for First Half 2025 https://t.co/fw5CAdP2ck

[Neal Arnold @_SilverFox1]

@j16425646 Breaking $11.........

[JC @j16425646]

$SPCB analysis published this morning yields a price target of $68 - $120, up ~14X from the current price. Allied Universal, $GEO, $AXON (which crushed earnings last night) cited as acquirers of $SPCB.

https://t.co/6Rcbre28sA

[Neal Arnold @_SilverFox1]

@j16425646 $SPCB is possibly the most undervalued AI stock . Time to load up before next quarterly earnings release confirms what we understand

[Neal Arnold @_SilverFox1]

@j16425646 $SPCB is possibly the most undervalued AI stock . Time to load up before next quarterly earnings release confirms what we understand