Markus

Ex investment banker & fund manager

[Markus @markusseneca]

@alojohhardcore should be a good short opportunity soon if it trades up a bit more - given the strong market over the last couple of weeks, desperately looking for a good short to provide a hedge in case of some profit taking and market slowing down

[Markus @markusseneca]

@shorttimelines @GergelyOrosz User agent would need to be able to buy - simple example - 50 tickets off the shelf from the airline. I assume a single user wouldn’t like to assume the risk of paying the airline upfront and reselling the tickets. Not sure OpenAI/Antrophic et all would either as not core busines

[Markus @markusseneca]

@shorttimelines @GergelyOrosz Think the point from gergely is a different one - why are travel agencies/OTAs able to offer so cheaply? Because they get contingents of discounted tickets from airlines which they pass on to consumers. Consumers will never get these contingents directly from the airline

[Markus @markusseneca]

@RihardJarc I am looking at the primary source which doesn’t seem credible. In fact, he doesn’t even confirm what the Redditor posted, he just says the future of GPU programming is agentic without any time context could be tomorrow, 3 yrs, 10 etc. you see my point?

[Markus @markusseneca]

@A_May_MD All I am saying is people shouldnt overinterpret these statements. There are likely negotiations in the background. Unless an agreement is reached, there is nothing to do for the competition authority. It will only engage officially with a bidder after an agreement

[Markus @markusseneca]

@A_May_MD Market is acting stupid - stock went to $150 pre market only to drop to $120 when French competition authority said "no talks with eli lilly". was recently involed in a french M&A deal and competition authority was only informed AFTER agreement on price was reached

[Markus @markusseneca]

@sonusvarghese @DannyDayan5 Yes the original post is about rate of growth. My point is more about capacity, dry powder and avenues to tap for financing. Large Hyperscalers (Microsoft, Amazon, Google) are basically debt free+ample dry powder in private credit (despite strong growth rates but from low base)

[Markus @markusseneca]

@sonusvarghese @DannyDayan5 So according to their definition 200bn out of 1700bn? 12%, not huge. Number probably inflated as well as definition is wide (includes„big data“ and „cloud tech“). Can tell you as of today, European Direct Lending funds have close to 0% exposure to true AI theme / datacenters

[Markus @markusseneca]

@sonusvarghese @DannyDayan5 Headline figure looks big but exposure to AI datacenters / theme is minimal as of today - majority of private credit exposure is related Sponsor backed LBOs

[Markus @markusseneca]

@rev_cap German equities, if you are referring to the DAX, are hardly representative for German businesses as a whole - DAX are mostly international companies with small exposure to Germany based on revenue take SAP as example.

[Markus @markusseneca]

Incredibly amusing how people outside private equity / credit get things wrong constantly. First of all, private credit is most relevant to hyperscalers, not private equity. The hyperscalers won’t give up their equity stake in the data Center and if would, we would be looking at infra funds rather than private equity. Difference is return expectations are significantly lower for infra funds vs PE. Current state of private debt is that deployment opportunities are scarce and dry powder high, resulting in gradual decline of spreads. Pretty much decreasing by 50-75bps pa since Ukraine invasion. Is there stress in private credit portfolios ? Yes. Is it out of norm ? No. You will have 1-2 blowups in the portfolio each year but Funds size positions so that it doesn’t blow up the fund. Agreements with LPs usually prohibit taking large positions which could harm the overall fund. Hence all the noise about private credit blowing up is quite exaggerated.

[Markus @markusseneca]

@JackFarley96 Try to calculate days inventory outstanding and tell me whether you see an increase - huge nothingburger from financial illiterate people

[Markus @markusseneca]

@alojoh @DavidSacks 😂

[Markus @markusseneca]

@MartinShkreli Spoiler: they are all shit

[Markus @markusseneca]

@giorgino1998 @alojoh Of course it would affect NVIDIA. But so far there is no evidence showing this. To the contrary - inference is becoming an even bigger thing than training and NVIDIA delivers the most cost effective inference