Surinder Capital (@SurinderCap1tal) profile picture

Sr Analyst at very fancy hedge fund (TMT L/S). Standup comedian. ex-McK, ex-Taco Bell. Views my own.

Joined December 2022
Photo by Surinder Capital (@SurinderCap1tal)Photo by Surinder Capital (@SurinderCap1tal)Photo by Surinder Capital (@SurinderCap1tal)
Surinder Capital retweeted
Is pod shop style the only thing that works? Are long-only value funds dead? Pod shops play an intellectual or informational advantage. They aim to create alpha over short time horizons: someone gives you money, you find more information than the market, and you make money with low correlation. Fundamental or "value" investing aims to buy businesses well below intrinsic value and hold them for a long time, often without the informational advantage that would support a short-term edge. It is clear why people argue this does not work. But they miss where the real edge comes from, if you can even call it an edge, and why this strategy can create excess returns. It is the edge of patient capital. If a few institutions and wealthy individuals give you money for the long term, without the risk of losing it when you underperform, you have a massive edge over the market. You can make stock and sector bets that many of your smarter peers cannot, because they simply cannot wait that long. The obvious objection is that maybe this is not alpha at all. It is just playing a different game. For exactly these reasons, value-only firms are shrinking. You cannot run this strategy if you want to maximise AUM. Patient capital is scarce and highly competitive, so you end up playing a game you cannot win. And if you are small and new, you have two options. Either go for multibaggers to build a multi-year track record whose returns would excite any patient investor, or stay small enough that you can still win an intellectual or informational edge. Either way its slow, and probably not the best way to make the most money if you are that smart
Reason value strategy has died and the rise in pod investing happened is bec over time markets become more efficient and just having valuation as “edge” got competed away. Markets evolve/change and good investors can adapt to markets vs yelling that markets should adapt to them
5
4
1
91
9,414
Impatience killed the cat. I covered 60% of my DC/infra shorts this week. Cat be nimble, cat get killed.
Incrementally the datacenter trade has to underperform near term. The messaging from Democratic governors in particular will be one of “we need to review how these systems are impacting our communities”. At minimum they are not signing up for the status quo, why would they? Longer term I don’t think the outcome changes much, so they you’re a buyer on weakness … the risk is just how early you are to the bottom.
1
484
After much consideration: Long $NVDA
20% downside and 80% upside (new ATH) feels pretty good? Initial debate is the 2027 growth rate (I think bears now believe flat, with down margins, and bulls think 20-30% with flat margins). Difference is ~$9 to ~$13 in EPS. But the real question (very hard) is whether 2028 grows at all (has capex truly peaked?). Right now I see likely $13.50 EPS in ‘27 (probably above new buyside), with flattish outlook to ‘28 — but I don’t have enough of a good industry level demand framework to be confident in that. If you feel great there is some/any growth in ‘28 this likely reflates to ~20x = $250-300 stock. Downside I can can get to ~$8 of ‘28 EPS assuming we’re flat YoY and down 20% the year after. This is preliminary and the inputs are fuzzy. Ofc Cyclicals should/tend to see multiple expansion at the bottom, but no one will know yet whether it’s a real bottom, or there’s capex declines and share loss behind door #3. I still think ~15x that ~$8 (or $120) is the intermediate downside case. Pretty draconian, but even more nonlinear outcomes are possible given China/Singapore (i think this isn’t as bad as recent fears), circular financing fears, and Jensen overhuang. My instinct is price is over-driving the narrative right now on how real the capex slowdown/reversal is. Hypers told you 3 weeks ago “it’s going up” and then going up again. Anthropic telling people they’re great at compute efficiency is cool and all — but is capex really going from +45% (conensus) to 0% (fresh consensus) and then -20%? That’s really sudden and all because NVDA isn’t writing checks to neos? (Simplifying but you get it - what changed so fast? How sure are you?) My conversations don’t suggest it yet, but I’m talking to guys that can only see about 12 months out and you need a demand forecast, not the supply forecast everyone prefers (because it’s easier to count). I don’t have that at current, so at $155 — I’m long some, waiting for the risk reward to get to infinity or more certainty that capex isnt shifting as hard as new view suggests.
1
2
976
Bill Clinton lied under oath about an affair, was impeached, and kept the most important job on the planet. If you think NVDA’s Board has some independent obligation to “set a precedent” or punish Jensen for his conduct, you’re not very familiar with Delaware law. Their job is to protect the interests of the corporation and its shareholders. If you think shareholders are better off without Jensen, make that argument. And if Jensen launched a proxy fight to replace the Board (he owns >3% of the company, check the Bylaws), I suspect he’ll win in a landslide.
1
6
725
Surinder Capital retweeted
Said this elsewhere but: accuracy is more important than precision. If you’re very right about the qualitative, that should translate directly into a view on the quantitative shape of the business and how it changes over time (i.e., a model with a forecast). Ideally you can drive a truck through the gap between where you see the business in 24 months and what estimates imply. Along the way you need a clear view of which quantitative signals would start to prove you wrong. You don’t throw the model away, at all. But it becomes a tool for understanding the destination, with important mile markers along the way. If the business decels 100bps faster than you expected, a quarter sooner, that should be treated as such.
1
10
831
Buying the dip. The overhuang is gone.
4
1
1
15
4,142
Incrementally the datacenter trade has to underperform near term. The messaging from Democratic governors in particular will be one of “we need to review how these systems are impacting our communities”. At minimum they are not signing up for the status quo, why would they? Longer term I don’t think the outcome changes much, so they you’re a buyer on weakness … the risk is just how early you are to the bottom.
2
1
7
4,303
Let’s say it together: Renewals today reflect a market short on compute. You have to explicitly underwrite the same dynamic in 2028. I think there is increasing unease with that for obvious reasons (CRWV demand and oversupply are reflexive). That said, I think Nvidia has no choice but to come to the rescue here. $CRWV can’t afford to borrow incremental dollars in the 11-15% range and the unsecured market may close to them fully, even at those prices. Nvidia stepping in with a GPU backed plan (ie, vendor financing). CRWV rockets briefly. But then doom sets into the broader implications. Even nvdia can’t backstop the entire ex-hyperscaler build out (ORCL, OAI, ANT). Up and then way down.
5
2,460
Open weights could be a massive positive for the duration of the AI infrastructure investment cycle without being a positive for its ultimate returns on capital
1
1,804
Ironic the most degen GOATs are the golfers. 🐅 🫲
1,498
Surinder Capital retweeted
How ridiculous was Michigan in Las Vegas? This ridiculous: Michigan is the first team — ever — to beat ranked teams by 30-plus points in back-to-back games, per CBS Sports Research. Had never been done in men's D-I history. Until tonight.
7
224
20
2,054
154,064
Surinder Capital retweeted
I’ve been covering college basketball for 20-plus years and have never, ever seen a team absolutely dominate 3 teams in 3 consecutive days like this. This is absolutely incredible.
143
317
38
5,597
1,060,165
Surinder Capital retweeted
Bet 〽️
1,447
3,430
910
18,290
3,276,924
This will change every finance professional's career in the next 24 months. For every AI model, there are imperfections. But, this is just a tool to do your job better. Retweet and follow for access ahead of launch. Trying my best to get you it before Thursday if we can. 💪
25
193
1
504
233,100