N.K

L/S.

[N.K @Nnoahkollerr]

@MeghanBobrowsky Alpha is generated by being able to see the world clearly before empirical evidence emerges.

Implicit with this is volatility.

$META

[N.K @Nnoahkollerr]

There seems to be a prevailing view that, even as $GOOG $META $SPCX etc, are able to directly compete at the frontier, the majority of value accrual will fall to OAI/ANTH, in some kind of "quasi winner take all dynamic" . The 2nd order inference is therefore that if $GOOG / $META severely undercut, they capture that 0 sum model layer value.

I believe this to be wildly wrong, and that 1st order model layer monetisation (selling tokens) is, by definition, a low return lever that $GOOG / $META don't need to enter, directly.

Instead, possessing SOTA models and monetising through their respective "monopoly channels" i.e FoA's (consumer & enterprise sides) enables value capture, many multiples larger that of selling tokens, at layers which the Labs can't operate.

[N.K @Nnoahkollerr]

$TDG

- Seems highly likely that $BA $AIR.PA prod. ramps remained constrained into foreseeable future.
- Even if we assume sell-side air-framer build rates to be true, the delta in implied OEM volumes vs today is insufficient to drive margin compression (when factoring in AM pricing lever)
- Bear M&A cases approach a small error term when factoring in horizontal expansion + pragmatic YTD activity .

What am I missing ?

[N.K @Nnoahkollerr]

Somewhat surprised about how many treat being at the SOTA in LLMs as black/white for the big-tech labs ($META $GOOG )

What i agree that it’s vital, especially on true agentic ability whicih is bound to model quality. But some 2nd-order thinking:

- SOTA across emerging technologies > SOTA on a single technology
- Today, the optionality a SOTA trajectory/team provides > the amortisable value of any given model
- On the consumer side, the next battle is likely one of use-case fit and ability per parameter

[N.K @Nnoahkollerr]

Whilst on $TDG - I struggle to see Sell-Side air-framer build rate fig's materialising in the respective time frames. I lean the opposite direction.

Even assuming Air-Framers exceed stated targets, @ current valuation, the margin headwinds via Mix are overly discounted. Also have strong opinions on M&A fears. Market seems to be applying an overly "Narrow lens" & assigns de-facto negative bias on any form of target set expansion. (Though have written about this on here)

[N.K @Nnoahkollerr]

If one isolates their Long Book to some function of Business Quality & Valuation, there comes a compelling case that the consensus concept of "No Free-Lunch" in Finance is simply wrong.

$TDG has always been a difficult company to value as if one factors in the TGR the business model suggests most accurate, the intrinsic number is stupid. That said, the excess return is only realised over a very long duration, so the claim should be taken FWIW.

[N.K @Nnoahkollerr]

$META

[Alexandr Wang @alexandr_wang]

vibes kinda good 🍉

[N.K @Nnoahkollerr]

Many takes framing $GOOGL Gemini Agent as a threat to $META are fundamentally flawed & extremely surface level. Muse, especially on the Enterprise side, is so valuable due to the scope of application $META has through the FOAs’ (near, and growing) Social Platform monopoly. I’m confident that the steady state of Model/Agentic capability is one near parity. That said, $GOOG scope of application on its own relevant enterprise side is tremendous. This continuous Zero sum thinking clearly misses a lot.

[Ravid Shwartz Ziv @ziv_ravid]

Everyone, repeat after me.
The ARR ("annual recurring revenue") of Anthropic and OpenAI is:
NOT annual
NOT recurring
and
NOT revenue

Thank you for your attention to this matter.

[N.K @Nnoahkollerr]

I would add that assuming a Spin of MI, the stocks discount to IV widens. Multiple is too low.

[N.K @Nnoahkollerr]

Im constructive of $SPGI management and the demonstrated willingness to improve the business via Spins.

It's inherently difficult to see the terminal state of Software. Whilst MI developments re: becoming "AI Native" have been positive, it's likely naive to believe that no risk exists.

That said, MI bottom line contribution is de-minimis & over discounted today. I don't see an MI spin as critical however I am supportive of the idea. The business has absolutely become "Diworseified" over time.

[N.K @Nnoahkollerr]

$AAPL consumer hardware moat has, to a large degree, been that the combination of an integrated stack of Superior hardware + superior OS functionality has been extraordinarily difficult to develop.

Historically, my underwriting of $META captured this. Whilst it has been persistently apparent that Phones were a fundamentally inferior hardware package, a better hardware form remains insufficient in disrupting the phone. In order to disrupt the segment, the utility of the OS has to create a step function change improvement in Consumer lives.

Whether OAI, META, etc wins is up to interpretation. My view is that $META possess a substantially larger right to win relative to the rest (have written extensively about what consumer agent success is dependant on). What, IMO, is less interpretative however, is that for the first time we may have an integrated version of both a superior hardware form factor & "OS" form.

[N.K @Nnoahkollerr]

@WaterportCap Structurally v different.
- Vastly different dynamics re: Centralisation of regulatory power
- Selling to Global Debt markets =/ selling to US mortgage markets (Primarily consumer vs Enterprise implications)

[N.K @Nnoahkollerr]

@mukund This is a complete nuance lacking argument

https://t.co/EAA47GI3sK

[N.K @Nnoahkollerr]

I think this argument is highly surface level & misses underlying structural components . Firstly, $Meta networks are definitionally two sided. In this case, Enterprises & Consumers.

Agentic systems are a tremendous threat to the Intent based, productivity layer ($GOOGL $AMZN ads etc). Simultaneously, Non-Intent based mediums are fundamentally different in that Users engage here for the opposite of maximising utility. Discovery surfaces are, by definition, human.

If one holds these thoughts to be true, Discovery Platforms become structurally better business models over time and continue to take share over a prolonged period. $META holds a tremendous right to win via both

1. the breadth of Enterprise customers, which, currently only spend with $META across one (small) point of the funnel (advertising).

2. The ability to create integrated Value-added services to these enterprise customers by leveraging "proprietary assets" (i.e Consumer Network, Models, Contextual understanding etc )

[N.K @Nnoahkollerr]

$Meta going after & capturing incremental Opex budgets (what are structurally multiples the TAM of Advertising) was clear

https://t.co/AqiZmJBxAc

[Alexandr Wang @alexandr_wang]

1/ muse for small businesses 📊

we found lots of people using muse to run their small business. plumbing businesses, grocery stores, farms, restaurants, and shops.

today we're launching a bunch of connectors to make that easier!

[N.K @Nnoahkollerr]

I think this argument is highly surface level & misses underlying structural components . Firstly, $Meta networks are definitionally two sided. In this case, Enterprises & Consumers.

Agentic systems are a tremendous threat to the Intent based, productivity layer ($GOOGL $AMZN ads etc). Simultaneously, Non-Intent based mediums are fundamentally different in that Users engage here for the opposite of maximising utility. Discovery surfaces are, by definition, human.

If one holds these thoughts to be true, Discovery Platforms become structurally better business models over time and continue to take share over a prolonged period. $META holds a tremendous right to win via both

1. the breadth of Enterprise customers, which, currently only spend with $META across one (small) point of the funnel (advertising).

2. The ability to create integrated Value-added services to these enterprise customers by leveraging "proprietary assets" (i.e Consumer Network, Models, Contextual understanding etc )

[N.K @Nnoahkollerr]

Id place emphasis on the vitality of such a vast enterprise base on one side of the network. Ad spend, to which $META current "TAM" is somewhat bound to, is a pretty small fraction of the net spend of the funnel. An incremental capture of this Opex bucket is therefore a TAM in the oder of multiples of current rev base.

It therefore comes down to execution on the enterprise side, which as I have written about, can be focused down to some combination of Consumer Contextual understanding, and the ability to translate that into integrated, Value added tools for enterprises.

Would also add here that agentic systems fundamentally disrupt the intent based, productivity layer - Likely most explicitly seen via $GOOG search (not that $GOOGL wont/cant innovate on the product side) .

[N.K @Nnoahkollerr]

@buccocapital Why would $META Enterprise be less clear ? Starting point, w.r.t structural “right to win” is stronger than what $Googl had.

I think it’s rather easy to underestimate the nature of the underlying business model & prod. Integration/ delivery pipeline in enterprise.

[N.K @Nnoahkollerr]

!!!!!

[N.K @Nnoahkollerr]

@MeghanBobrowsky Alpha is generated by being able to see the world clearly before empirical evidence emerges.

Implicit with this is volatility.

$META