Daniel Koss

5.6K posts

Daniel Koss banner
Daniel Koss

Daniel Koss

@daniel_koss

Founder of Edelbridge Capital. Concentrated public equities fund focused on AI infrastructure. Former pro gamer, YouTuber, and serial entrepreneur.

Switzerland Katılım Mart 2013
317 Takip Edilen32.6K Takipçiler
Sabitlenmiş Tweet
Daniel Koss
Daniel Koss@daniel_koss·
My April Performance Report: +72.9% -> everything 3rd party verified and publicly tracked. 📈 April return: +72.9% (Autopilot) 📈 YTD: +113.6% (personal portfolio) My benchmark, the Nasdaq, returned +8.9% over the same period. That means my portfolio outperformed the benchmark by more than 12x. My thoughts: April represented an amazing opportunity to “buy the dip.” But it is important to understand that the foundation for the performance was built in March. March was a rough month for most AI infrastructure, tech and growth investors. The Iran conflict spooked investors out of growth stocks. Rising oil prices, inflation fears, uncertainty around future rate cuts, and even the risk of rate hikes all presented real risks. On top of that, the conflict became ugly and destructive. There was real damage done to valuable infrastructure. The best decision I made in March was to actively manage risk. When I felt the situation was escalating, I increased my cash position to 10-20%. When I felt it was de-escalating, I bought back in. I also made a couple of good trades, like buying $KRKNF into the mining headlines. That was mentally taxing and annoying. Constantly buying, selling, buying back, trimming again, and re-entering is not fun. But it was worth it. While many investors were down a lot in March, my portfolio was mostly flat or slightly up. Instead of just reacting to headlines, I constantly monitored the real situation on the ground (or on the water). If you look at my posts and articles from March, you can see that I mainly focused on drones, missiles and ships in the Strait, and whether the situation was escalating or de-escalating. Trump headlines are bad signals and very dangerous. Then, in April, it became obvious that many stocks were completely oversold. Some of my highest expected return stocks were down 30-50%. The market was deep in extreme fear. At the same time, I expected Trump to find a way to de-escalate, and many of the metrics I tracked started to improve. So I reallocated into oversold growth and AI infrastructure stocks like $NBIS, $AEHR, $OUST, and others. That created the setup for the 70%+ return in April. Investing under Trump is both the gift of a lifetime and the most stressful thing ever. It does feel like I am aging at three times the normal speed, if I am honest😅 Since many people noticed that I timed many of the events and stock picks quite well, I started getting a lot of requests to share my portfolio, research and trades. That's something I planned to do anyway, but in the last month, I have been thinking about it, and here is what I came up with: My current plan is to split it into two products: 1) X Premium subscriptions I am currently testing X Premium subscriptions at $5/month. The goal is not to make meaningful money from X subscriptions. The goal is to create a more exclusive community, filter out bots, trolls and spam accounts, and give serious followers time to look into my trades and portfolio updates before they get copied all over X. I have already seen many accounts copy-paste my stock picks, valuation numbers and ideas, then sell them as their own. Nobody owns a stock or thesis, but sometimes I feel like it would be nice to at least mention the person who made you aware of a pick and who got you excited about a stock? The funniest part is that in some of my models I have made calculation mistakes before, and some creators copied the mistakes too 😆 2) Premium deep research on Substack The second product would be a premium deep research newsletter on Substack. For that, I want to hire full-time researchers who are true industry experts in their fields: AI infrastructure, semis, power, memory, data centers, and so on. This would be a serious, full-time effort for me. It would likely cost around $100/month, because producing this kind of research properly takes a lot of time, and I would need to pay a strong team of expert researchers. This is something I want to do long-term because I do like investing, making great returns and sharing the best ideas with you, but to be completely honest, I often get bored of it. My true passion is doing the research and learning about business strategy, technology and the future! For example, Nebius is a stock that fascinates me because I am insanely excited about Tokenomics, Jevons Paradox, platform businesses, network effects and game theory. The stock price going up a lot more is why I buy, hold and share the thesis, but it's not why I love the company. So the idea is: 🔹 Professional investors and wealthy individuals can pay for premium deep industry research. They will get high-quality research. While there are many fantastic Semi analysts on X and substack, I believe my angle of identifying where most value accrues (= stock will go up most), is quite popular among professional investors. At least that's my experience interacting with PMs so far. 🔹 X Premium would remain intentionally low-priced, most likely below $1-10/month. I want to create a more exclusive community, filter out low-quality accounts, and give serious followers time to study my trades and portfolio updates before everyone starts chasing the same stocks. I will obviously get some picks wrong. That's why I always want you to look at my ideas, picks, etc. and then make your own conclusions. What do you guys think? Would love to hear your thoughts.
Daniel Koss tweet media
English
32
2
172
90.6K
Daniel Koss
Daniel Koss@daniel_koss·
@wwrdbruh much more important than MW are capital efficiency and how many tokens you get out per MW. Nebius is leading there which means higher ROI on new MW. flywheel
English
0
0
1
43
Ww Rd
Ww Rd@wwrdbruh·
@daniel_koss It's massively overvalued vs peers on an EV/MW basis. If you understand that you need MWs, and MWs are really hard, then you understand why its not silly.
English
2
0
0
166
Daniel Koss
Daniel Koss@daniel_koss·
It's pretty silly that $NBIS is flat today. All the Physical AI stocks are pumping like crazy. Market probably doesn't understand yet that Nebius will make more profits from Physical AI than many of the hot names? 🤐 Nebius now had enough time to digest the great earnings with some mixed news markets might not like (more capex, fuel cells, etc.) but considering they just demonstrated their pricing power, I feel like we should be ready to run again soon? Market currently thinks you should only own the shortage sellers. But Nebius controls the full stack. Buying more costly components is clearly bad, but I think people don't understand they can probably more than offset that by increasing prices and adding services.
English
20
9
170
16K
Daniel Koss
Daniel Koss@daniel_koss·
@hiteshkar skimmed it and saw the BOM part matches what others are saying
English
0
0
0
59
HK
HK@hiteshkar·
@daniel_koss The reason I didnt take a big position was because I dont see their order book growing much yet. But they are an absolute bottleneck for high end joints. Did you read my article? blog.stockdrifts.io/p/robotics-bot…
English
1
0
1
153
Daniel Koss
Daniel Koss@daniel_koss·
For Physical AI, Harmonic Drive Systems actually looks like one of the best investments to me. It gets hype here on X, but totally deserves it! This is not a speculative low quality shitco. Ticker: 6324 Exchange: Tokyo Stock Exchange Company: Harmonic Drive Systems Inc. For most international investors, the cleanest way to buy it is probably directly on the Tokyo Stock Exchange through a broker like IBKR. I would personally avoid random low-liquidity OTC tickers if you can access the actual Japanese listing. What stands out is that literally all the smartest people I know who understand robotics seem to love this stock. The main point they keep making: Harmonic Drive makes best-in-class strain wave gears. These are essential for co-bots and still essential for humanoids, especially at the elbow and wrist. Even if bigger joints use QDD or cycloidal gearing, and even if smaller joints end up using something else, the field has not fully settled on one winning architecture yet. But Harmonic Drive is clearly in the middle of the serious robotics supply chain. Harmonic Drive gets a huge part of the BOM! The moat is very strong. Producing high-quality strain wave gears is technically complex. Chinese competitors are coming up, but supply-chain risk and dual sourcing make Harmonic Drive hard to replace. The US and its allies need supply-chain independence from China in humanoids, robotics, and defense. Even Tesla and others likely cannot just make these themselves anytime soon. The bull case, which I think is straight up what will happen, is that Harmonic Drive has to grow enormously to meet demand, and this can run for many years. I think humanoids are a classic example where short-term expectations are overhyped, but the long-term potential is insanely underhyped. Not financial advice. I am probably going to add it as a 20% position to my portfolio. Another name I need to make a Substack post for 🥲😂
Daniel Koss tweet media
English
8
2
44
6.5K
synapsee
synapsee@DocSynapsee·
@daniel_koss I’ve been looking to get in too. Are you a buyer at these levels?
English
2
0
0
162
Daniel Koss
Daniel Koss@daniel_koss·
In case anyone wonders why I record so many interviews with $NBIS executives (spoiler: more coming)... It's really simple. I'm 100% convinced $NBIS is the AWS of AI and the next hyperscaler. If you could interview the leaders of AWS 10 years ago, wouldn't you want to learn as much as you could from them? They're brilliant, really kind people and what might be underrated: they work freaking hard! All these headlines we get to enjoy as investors, all of that is only because they spend months grinding and making shit happen. I promise you guys: $NBIS x Physical AI will be massive. Soon... People once again sleeping on what they are cooking. While I do publish some interviews (with their permission), I also talk to many leaders and smart people that work in AI at many of the world's leading companies in my free time (honestly where's the difference). I'm insanely happy with how well investing in AI pays, but at the end of the day, enough money is enough money. My true addiction is and has always been learning. Knowing what's going on in the world and where we're going is so freaking exciting. Everything I do is really just a lazy excuse to learn every day and somehow have people not realize I'm still just a little kid following his curiousity every day.
English
10
4
107
7.1K
PhotonBull
PhotonBull@PhotonBull·
@daniel_koss I don't think people sold Kraken to buy other defense stocks, everyone probably moved to AI stocks😂
English
1
0
1
451
Daniel Koss
Daniel Koss@daniel_koss·
Pretty funny how $PNG.V is an easy double from here but FinX only liked the stock at all-time high. Now everyone just... forgot about Kraken to buy much more expensive, more speculative defense stocks at significantly higher valuation with worse dilution that are not profitable and grow slower. Ok, I'll stick with Kraken ✌🏻😂
English
10
2
53
5.8K
Daniel Koss
Daniel Koss@daniel_koss·
that is what i'm asking myself. i think it does come down to 2 things: 1) taste 2) judgement ai writing still sucks, no matter what prompt you use. it still hallucinates on obvious issues. it still makes very unfunny jokes. it just often misses the mark in a way a human wouldn't.
English
1
0
8
456
Mr. Big Bits
Mr. Big Bits@jpizzzal·
@daniel_koss What will an intern do that an agent can’t already do? (Specifically)
English
1
0
2
416
Daniel Koss
Daniel Koss@daniel_koss·
I'm considering hiring an intern. The list of stocks I want to research just grows to damn fast. What I'm seriously thinking about right now: hire a human or build better agents? Funny time we're in.
English
21
0
66
6.3K
Michael Sikand 🦑
Michael Sikand 🦑@michaelsikand·
Congrats to the 730+ investors following my "Memory Supercycle" portfolio on @joinautopilot with $5.3M. $MU surpassed $1T in market cap today. The port is up ~11% this morning. $MU $SNDK $EWY $WDC $STX
Michael Sikand 🦑 tweet media
English
21
4
82
14.1K
Daniel Koss
Daniel Koss@daniel_koss·
@michaelsikand 👌🏻 Let's also be real for a second, the media part of our businesses works better with collaboration 😂📈 I've seen so many Creators get left behind over the years because they didn't want to push smaller accounts lmao
English
0
0
6
805
Michael Sikand 🦑
Michael Sikand 🦑@michaelsikand·
I get annoyed when FinX creators disparage one another. Y'all are thinking too small. I see infinite TAM for the stock market media product. The next Cramers, Sorkins, Buffets, Ackmans will be born here. The old guard of stock media is aging out. Retail is now $0.30 of every trade made. Most people on this planet do not own stocks. There are billions of followers and billions of subscription dollars to go around for everyone. Hell, stock influencers might be the next rap stars and ballplayers once AI takes over and all anyone cares about is trading 🤣
English
34
2
189
24.4K
Daniel Koss
Daniel Koss@daniel_koss·
Anthropic only got H100 from Elon apparently. $NBIS will give them that new hot shit. "Nebius will take care of you." When are we getting a nice deal announced? 🤔 GPUs are still king for training, in case all the inference bulls (that's me) forgot. For OpenAI and Anthropic it's mission critical to always have maximum access to the newest GPUs to train the best models and not fall behind. @wienglischmann when will @nebiusai start the next PR marathon and announce who that "visible demand into 2027" is for? 🤭 I think we all agree Nebius stock would look much better at $300.
English
7
3
67
5.6K
Daniel Koss
Daniel Koss@daniel_koss·
Thx for having me on your pod! FYI this was recorded a few weeks ago.
ᗩv𝔦 🥞@AviNMash

Everyone who got burned on lidar SPACs back in 2022 wrote off the entire group, and that fatigue is exactly why almost nobody noticed that $OUST quietly became the revenue leader of the whole category while trading at a fraction of its flashier peers. That gap between the perception and the reality is the setup worth understanding. Props to @daniel_koss for putting this on my Radar (Video below). Ouster builds digital lidar, the laser-based 3D eyes that let cars, robots, drones, and smart-city systems actually see the world, and after merging with Velodyne it now owns the widest sensing and perception platform in Physical AI. The portfolio runs from its OS digital lidar line to solid-state sensors, the Velodyne surround-view units, Gemini perception software, the BlueCity traffic platform, and the ZED cameras and AI compute it picked up buying Stereolabs. Hardware plus perception software plus custom silicon under one roof is a real moat in a market full of single-product startups. The financials are where the crowd reads it backwards. They see a net loss and an EPS miss last quarter and they file it with every other failed lidar name, but they skip the part that matters. Ouster just posted its 13th straight quarter of product revenue growth, revenue jumped 49% to about $49M, GAAP gross margin climbed to 43%, and the adjusted EBITDA loss narrowed to $7M while management holds operating expense growth to single digits. They are guiding toward a sustainable 35% to 40% gross margin and a path to profitability in 2027, and they are doing it with about $175M in cash and no debt, which is the strongest balance sheet in the group by a wide margin. The catalyst that just landed is the one I would not ignore. Two weeks ago NVIDIA cleared Ouster's new Rev8 lidar through its DRIVE Hyperion qualification, which makes the sensors eligible to run on NVIDIA's platform for Level 4 autonomous vehicles, and the stock jumped 26% on the news. That drops Ouster into the same NVIDIA reference lineup as Aeva and Hesai, which matters for the next wave of robotaxi and autonomy programs being designed around NVIDIA right now. The valuation piece is what makes the whole thing click. Aeva, the lidar name everyone chased for a 431% run, carries a market cap near $1.4B on roughly 30 times forward sales while guiding to maybe $36M of revenue this year. Ouster sits at a similar market cap near $1.7B on about 7 times sales, except it is doing more than eight times Aeva's revenue and still growing 49%. The market paid a rich premium for the smaller, earlier, faster-burning peer and left the actual revenue leader sitting cheap. Analysts carry a Strong Buy with a consensus target around $39 and a Street high of $50. I will not pretend the risks are gone, because Ouster still loses money on a GAAP basis, it just put an at-the-market program in place to sell up to $100M of stock that can dilute holders, and lidar pricing competition out of China stays brutal. The flip side is that the balance sheet removes any near-term funding gun to the head, the diversified base across automotive, industrial, robotics, smart infrastructure, and defense means it does not live or die on one design win, and that NVIDIA qualification is the kind of validation that pulls in the next round of customers. So the market is still pricing a busted SPAC that burned people three years ago, when what it actually owns is the broadest perception platform in Physical AI, the most revenue and the best balance sheet in lidar, and a fresh NVIDIA stamp, all at roughly 7 times sales. Every robot and autonomous machine being built needs eyes, and that perception layer is scarce the same way compute is, the same logic that pulls me toward Bitcoin and the rest of the AI buildout. A roughly $1.7B company that became the category leader while everyone was looking the other way. NFA. Do your own due diligence.

English
8
1
21
6.5K
Daniel Koss
Daniel Koss@daniel_koss·
Which one of you degens just dropped 300k into my way too concentrated Autopilot portfolio? 😆 Investing other people's money is way more pressure than investing my own money, not going to lie. I hope all of you who invest in my Autopilot portfolio know that I'm going extremely aggressively and with very high concentration into very few names. So please, please, please do not invest money that you can't afford to lose!
Autopilot@joinautopilot

Big Whale Alert  Someone just deposited over $300K into @daniel_koss "AI Infrastructure" portfolio The portfolio is a concentrated bet on AI infrastructure & is currently up +117% in two months

English
10
2
44
15.6K
Autopilot
Autopilot@joinautopilot·
Big Whale Alert  Someone just deposited over $300K into @daniel_koss "AI Infrastructure" portfolio The portfolio is a concentrated bet on AI infrastructure & is currently up +117% in two months
Autopilot tweet media
English
4
1
44
21.9K