MJ-AI-First-Builder

Lifelong learner. AI-builder who runs a 25+ global finance team as a day job. Love reading, outdoor walks & weight lifting. Passion - investing & building.

[MJ-AI-First-Builder @ManitJain3]

Millennials: we’re staring across a chasm. The next decade may be the most disruptive in our lifetimes —AI, robotics, crypto, genomics, space exploration & new world order. Tariffs - Just noise! Stay rational, avoid hysteria, & let compounding quietly do the heavy lifting. 🚀

[MJ-AI-First-Builder @ManitJain3]

Our superpower as humans is the context graph that naturally builds in our heads. It connects topics from multiple fields, experiences, conversations, and more, producing unique insights and eureka moments.

You could link learnings from lifting heavy weights (doing sets faster or slowing down to maintain form) to grasping new concepts while reading and re-reading. Or connect product principles learned on the job to evaluating an investment opportunity—and so on, in infinite permutations.

This is hard for AI systems: how do you get them to establish such interlinkages on their own and turn them into actionable insights? But a human mind with an AI system as an extension of its abilities could take this human capability to the next level.

[MJ-AI-First-Builder @ManitJain3]

Great read!

A gearbox is the most common example. By linking gears of different sizes, it can make a motor turn roughly ten times slower while pushing about ten times harder, trading away speed to gain strength . A gearbox does not create energy, rather it transforms it. For example, at a 100:1 reduction, torque multiplies by about 100 while output speed divides by 100 , which is why high-precision arms often usestrain-wave (harmonic) or planetary drives.

[Interlatent @interlatent]

Hardware is the bridge between AI and the physical world

Atoms and bits must work together to create future systems embedded with physical intelligence

We wrote a guide for those curious about the atoms.

[MJ-AI-First-Builder @ManitJain3]

@suekhim Maths, English, Coding

[MJ-AI-First-Builder @ManitJain3]

https://t.co/fQssOkfpDC

[Henry @HenryInvests]

Upstart CEO Paul Gu @paulxgu buys $1,375,000 of $UPST on the open market.

This follows his $3.9M open market purchase in November.

[Lune @forgotxuser]

$UPST gamma squeeze appears to be catching some attention on Wallstreetbets.

$50 intrinsic at $28?
35% short, 75% utilization on IB, only 100 million shares and the ceos keep doubling down to buy millions more, latest being yesterday.

Seems perfect for a short squeeze.

https://t.co/5zBClZ1RmW

[MJ-AI-First-Builder @ManitJain3]

Misunderstood companies - the ones that I’ve added to recently based on the latest earnings and the pricing action thereafter:

$MELI , $UPST , $NOW

[Roman Chernin @romanchernin]

@nebiusai could probably help SpaceX to become the cloud by partnering and providing the full platform to serve diversity of customers… not just rent out a full cluster to one customer :)

[MJ-AI-First-Builder @ManitJain3]

While the Q1 cost step-up was real, but I'm not sure we can conclude that $UPST has poor cost discipline. The question is whether it's waste (Meta-style bloat to be cut) or investment (funding new verticals that compound the data moat - 3 new S-curves simultaneously). The evidence points to the latter.

The total sequential (Q4→Q1) OpEx step-up was ~$30M on an adjusted basis. The company attributed this on the call to three factors — seasonal comp, annual gathering, and growth investment.

The management should have been explicit about the investment cadence when they set the 21% full-year EBITDA target. If you need to spend $15-20M extra in Q1 to set up the back-half ramp, say so upfront. The surprise, not the spending, was the problem.

This is my recommendation to the management: https://t.co/NQmq8Sroqu

[MJ-AI-First-Builder @ManitJain3]

https://t.co/UMgZII8SXC

[MJ-AI-First-Builder @ManitJain3]

There’s been a ton of noise around $UPST after yesterday’s earnings and the after-hours crash.
Let’s take a step back.

I’ve laid out:
• Bull case (power-law upside)
• Bear case (short thesis combined)
• Plus an open letter to the new management @paulxgu & Andrea Blankmeyer

Full breakdown here: https://t.co/DhYj1NGQbm

cc: @HenryInvests @RonAMacro @CaptTweetz @MacNeillPatrick @goddreng10 @HarryKay_ @zhangtaoSG @Futurenvesting @bradsferguson

[MJ-AI-First-Builder @ManitJain3]

https://t.co/UMgZII8SXC

[MJ-AI-First-Builder @ManitJain3]

https://t.co/UMgZII8SXC

[Tao @zhangtaoSG]

$UPST management should learn from $APP AppLovin:

1) Bring most employees back to office

2) Give SBC only to the top 10%

3) Stop wasting money on year-end parties

UPST needs discipline, not Google-style comfort culture.

[MJ-AI-First-Builder @ManitJain3]

Finally, three things that would change the rating:
1. National exchange listing (Nasdaq/NYSE) — unlocks institutional buyers. Without it, the capital that drives sustained re-ratings can't own this name.
2. Q4 FY2026 gross margin ≥ 38-40% — confirms the energy-services mix shift is real and not one-quarter noise.
3. A second confirmed data center contract with a named counterparty — converts the thesis from narrative-grade to investment-grade. (10/10).

[MJ-AI-First-Builder @ManitJain3]

@daniel_koss The foundation of this thesis is solid. The AI power crisis and multi-year grid interconnection queues create genuine demand for on-site power generation. But I disagree with the bull case at the $11.6 entry price, given the facts at hand right now. Thread. 🧵 (1/10)

[MJ-AI-First-Builder @ManitJain3]

@daniel_koss At $11.59:
Upside if everything goes right over 18 months: ~$18–22 (+55–90%)
Downside if Q4 disappoints or Monarch converts and sells: ~$6–8 (-31 to -48%)
That's roughly 1.5:1. Not asymmetric enough to justify concentration. (9/10)

[MJ-AI-First-Builder @ManitJain3]

@Longviewres acknowledges but doesn't fully model:
Monarch holds ~$80M of convertible preferred. Conversion price: $5.00/share. Forced conversion trigger: $15.00.
At $11.6, every buyer sits 29% below that trigger with ~16M shares of potential new stock overhead. That's a structural ceiling — not a risk footnote. (8/10)

[MJ-AI-First-Builder @ManitJain3]

$CGEH vs. $BE :

Start with the technology: Bloom uses solid oxide fuel cells — an electrochemical process running at ~900°C. Capstone uses combustion microturbines. Different physics, different manufacturing, different economics. Not comparable. But even on pure business metrics:
$BE Q1 2026 revenue growth: +130% YoY. $CGEH: +33%.
Bloom's anchor customers: Oracle (2.8 GW), AEP (1 GW), Brookfield ($5B). Bloom deploys up to 1+ GW per site. CGEH maxes at 20 MW. (7/10)

[MJ-AI-First-Builder @ManitJain3]

@daniel_koss Section 11 builds a $350M revenue scenario on a single 100 MW contract. If that contract isn't in the public record, the model is pricing a deal that may not exist in the described form, at pricing that can't be confirmed. (6/10)

[MJ-AI-First-Builder @ManitJain3]

The report says Capstone has "no direct competitors."
But we do have: Bloom Energy, Caterpillar diesel generators, Cummins engines, GE, and natural gas reciprocating engines. A data center CFO choosing between Capstone and a Cat diesel + separate chiller is making a competitive choice. "Differentiated technology" ≠ "no competition." (5/10)

[MJ-AI-First-Builder @ManitJain3]

Where the report breaks down: The $265B TAM uses total global data center capital expenditure. AWS, Google, Microsoft, and Meta will not run on microturbines. They use utility contracts, large gas turbines, and are evaluating nuclear. The actual addressable segment is edge computing and mid-tier colocation. This is off by an order of magnitude. (4/10)

[MJ-AI-First-Builder @ManitJain3]

Agree: Sub-5ppm NOx emissions may give Capstone a structural permitting advantage in dense urban markets where competitors trigger air quality review thresholds. That's a real regulatory moat. Seven consecutive quarters of positive EBITDA post-bankruptcy is ahead of typical turnaround timelines. (3/10)