Martin Bradstreet

How to use my X account Live Portfolio: auto updated daily: app.alphaport.ai/shares/74f3…. ALPHAPORT.AI share links let you subscribe to portfolios (if the user opts in). Mine is opt in, so you can subscribe with your email. You will get an update of my moves daily before the market opens. Substack: I write 3 substack posts a month. 1 end of month update on my thinking, and 2 pieces on themes or specific companies I am interested in. The most recent one is linked in my profile. ALPHAPORT.AI - This is my portfolio tracking and investment research platform. I work a tonne on it. It is very affordable with a 1 month trial, try it out! Or join the discord here: discord.gg/Nwh3RtMesJ If you have any questions or suggestions about these topics, feel free to post in this thread, which will be pinned.

Was pretty consistent selling in $BDSX Friday, i increased to 12.2% which is about as high as i can make a position (actually this is irresponsibly long, lets be honest, though I am also mentally prepared for a 0 on the entire position, so maybe not). I guess we are just going to see who is right and who is not right on this one but momentum clearly in control atm. That said, given the distribution of outcomes here and my thesis, i find it hard not to believe Kelly Criterion > Sharpe. -We are talking 2.63x gross profit -i think we see 23-24% growth when they report November 5th. -There is a very credible acquisition thesis (strategically important to multiple buyers as BDSX's tech is increasingly important as multicancer screening increases (more multicancer screening = more lung nodules detected = more stratifying the risk of those nodules). -We're also well below where Schuler (huge insider) was buying in August now ($24.50). The next interesting information to come up will be in November when Gary Pestano, Chief Scientific Officer, speaks about MRD November 11 “Listening for the Whisper - Monitoring Residual Disease and Emerging Resistance in the Blood”. Pestano worked at Ventana which was another huge Schuler investment (flipped to Roche Diagnostics for 3b, Schuler made 300m). Now he’s Chief Science Officer at BDSX? I dunno, seems like the team is working to get the best possible result here, and over a very very long period of time. The price action says something is up - but we're not talking about a big company here (223m market cap with half the float owned by insiders) so the PA is less trustworthy (there’s no fund with 50m of this and a full time analyst sweating it non stop). This ticks all my boxes for an investment, so 🤷‍♂️... my fate awaits. It's really tempting for me to be defensive since my year has still been very good, even after Thursday (i entered with 20% of my portfolio as compute which pretty aggressively nuked though Friday bounced at least a little), but i guess i just have to continuously tell myself to make decisions that lead to maximizing the amount of EV / money i make. There are a lot of incentives to not do that, but ultimately it is the only thing that matters for me personally. Nothing in this post is financial advice or a suggestion to buy $BDSX, i am just clarifying my own thought process around a large holding of mine and thought others might find this entertaining - as i mentioned, i'm also extremely aware that this company could cause permanent capital loss.

Yesterday $ABT released earnings, and once again mentioned 'tuck ins' on the call. As a $BDSX investor, i've been waiting to be tucked in for a little while now, so i thought i'd revisit my thesis. The new writeup on Biodesix is in my bio.

Portfolio Update Sunday October 18 2026 L/S account: +83.15% (55% of net at year start) LO account: +45.42% (45% of net at year start) Weighted Performance = +66.35% Nothing below is financial advice or a suggestion to buy specific securities and is shared for entertainment purposes Well, my attempts to ‘stick the landing’ on the AI trade have been mixed, I think largely because I am still fairly bullish about the long term prospects of the technology and view multiples across many companies as exceedingly reasonable. That said, I was pleased to see that when I filtered for the AI trade in my ALPHAPORT portfolio, perhaps with generalized fear toward a growing call for “Black Monday” I see that I am now at adjusted net exposure of 23.4%, down from a fairly steady 30-35% earlier in the year. This is on a book that is 88% net long. So, not ideal, and certainly not net-short going into Monday, but it could be a lot worse. In the more defensive sectors I am at 38.5% long Healthcare as well as some other positions I consider fairly defensive. In fact, I did also long $SKY, $RKT and $CVCO - normally these might be more aggressive low rates plays, but in the kind of environment we have low rates because the AI bubble explodes, these would do well i think. So, not defensive but handle a particular regime that other parts of my portfolio might underperform in. I feel especially good across the top 10 positions which are well prepared for a nuking of the trade 1. $BDSX - Biodesix sells a blood based diagnostic that tells patients how to stratify risk of lung nodules (is the nodule cancer or not). I have continued to add to this position in sell offs and it is probably my best idea. The company sells at 2.43x revenue with 82% gross margins and 30% growth, in a space that often sells up to 6x revenue. I also believe it is strategically important to both Abbott Diagnostics and Guardant Health and that ABT will try to buy it. 2. $HY9H - I am tied to the mast on this one. With a clear goal of buying back shares opportunistically Hynix has enough visibility into earnings through 2027 and 2028 that I think the pay back is simply too fast here to do anything except wait it out on the LTCG point and the future 2027 earnings which will be the highest. 3. $7974 - I put this position on initially near the bottom, but smallish, but recently ramped it very large. My logic is that the 5 year yield of Nintendo is going to be higher than 5%, but it trades inverse to memory (rationally - Nintendo’s costs went up significantly from higher memory prices and they definitely cannot pass all of it on). If you could buy a 5% treasury that traded exactly inverse to korean memory, you would go extremely huge on it. I think Nintendo is going to be better than that. 4. $DAVE - Payday lender moving into banking relationships with its customers (going well). I trade around a core position on this one opportunistically as it is high volatility - but typically trades between 12-17 aEBITDA w/ 30+% growth. 12/30 is too good so i usually buy whenever it gets to that point. 5. $LOCO - Relatively new position on an extremely value focused flame grilled chicken business in California that is popular with SNAP recipients. Attempting to branch out of the California market into New York. 6. $FTAI - I view this as primarily an aerospace play with a little bit of a ‘behind the meter power’ call option, as opposed to the AI trade. But since it sells off when AI power sells off, i tend to buy at that point. 7. $9984 - this isn’t really a top 10 position as i have totally hedged out the $ARM exposure, so its net weight is quite low. Since my lots are up a lot and Masa has largely done well so far on the AI trade, I am just going to let this ride. The basic argument if you short out $ARM is that you are getting the OpenAI shares for free. Of course, things are not that simple, but that’s roughly how I view the spot. 8. $005935 - Samsung Preferred is incredibly cheap and trades at a major discount to Samsung (25-30% usually). This gives it a market cap around 700 billion. The preferred has the same economics as the common. Samsung made over 80 billion in operating profit last quarter. Even if memory revenue do not go up further from here (expectations are there’s a little more left) this is extremely high earning power relative to market cap. Will this continue for 8 years and we make massive amounts of money? Probably not. Will it continue long enough that a combination of dividends (goal of returning 50% of FCF) and internal compounding (building out the samsung foundry business, phones, and future businesses, such as robotics - they own 35% of Rainbow Robotics for example) grow the company further? Sure, why not. Samsung is good at making money - from 2016-2021, tangible equity grew 65%. This had nothing to do with ‘the AI trade’. From 2016-2026, tangible equity is up >250%. So, money that is internally invested has generally had an ok ROI. 9. $AMZN - This is a long term hold for me and one i plan to defer as long as possible. AWS will be the biggest winner of the AI trade, both from selling AI compute directly (sold out forever with rising rental hour prices) and from selling compute that is more in demand as a result of AI (Athena, OpenSearch and S3 all end up being used a lot more with agents). 10. $XGN - Another blood based diagnostic, this one has a really killer team in place and steady growth. It’s still cheap even after the run up, and should be fairly uncorrelated to the AI trade from a business perspective (as the government is a payer), but maybe not from the perspective of who is buying the stock (certainly has some retail interest). So, overall I am happy with the top 10. One major sale over the last week was $MKSI, which I think will do well in the future. Perhaps I will re-enter it in the future. I simply closed it out to buy higher beta AI stocks which had sold heavily on Friday. It looks like by Monday this may be a mistake in the short term, but i am comfortable with the decision. It went like SELL $MKSI BUY $AAOI / $LITE / $PENG Essentially jamming oversold higher beta names that have great upcoming results. $PENG especially is down since their stellar earnings. And i think we have a lot of visibility into the optical earnings, I don't think this part of the trade is going to just dissapear overnight from pricing (like people's memory fears). On the short side, I was reasonably efficient catching the top of the AI Bio bubble, though I am keeping some of that trade on for now. But I did close out the higher quality stuff like $TWST and $TXG, both names i’d love to own at cheaper prices. I also increased the higher quality clinical names a bit here - I have been long $CELC for a while now, probably the cheapest on a price to peak sales basis, but I increased it a little more on the continued selling. I also added $ABVX, $CYTK and $INSM. I need something closer to $90 a share for $INSM to really increase that though. Prospects look extremely good but for now its still a little expensive. The cybersecurity bubble is also finally starting to pop, but perhaps i was just too early there and there will not be continued selling. I don’t know why, valuations are still incredibly rich, all it takes is selling to beget more selling. So i will keep those on. But we have a lot of ground to make up in $CRWD :) So, that’s where we’re at going into “Black Monday”. I expect the portfolio to be down in the asia session (simply follow through from Friday and then continued weekend pressure on Hyperliquid in memory names etc). But overall I think i could have been positioned worse. I know there is a crowd out there who is mega long $NBIS going into Monday - we will see what happens there. It’s not a position I am averse to owning longer term, but I do also think this will be the most likely quality name to have a massive washout as it is the AI stock with the most optimism built into its price, even at $160 a share. With Samsung Preferred you really only need 2027 and 2028 to be good, half the money to be return as dividends, and for Samsung Foundry and any future projects to not be total money pits. With $NBIS you really need the whole trade to work really well. In some ways it is really just $SOXL. So, we shall see. Good luck this week.

"Semicap September" didn't really go according to plan but the businesses still look pretty good to me. I own $MKSI which I think has a lot of torque - but less of the service revenue that makes $AMAT / $LRCX slightly higher quality. Not a suggestion to buy any of these of course.

Narrator: it was not enough.

Really torn on MedTech lately - valuations are starting to look more reasonable - you have the insider buying on $BSX at a mid to high teens multiple, for example. But with the 10 year at 5.24% and key franchises facing competition, as well as the general vibe that maybe biotech innovation will reduce procedures overall over time (GLP1s etc), one does wonder. That said, I did some thumb sucking on $ISRG at the absolute bottom and couldn't get there on valuation, would have been a nice trade.

10/10/26 Portfolio Update: 73.89% YTD Nothing below is financial advice, and is shared for entertainment purposes. Longs (184.8%) $HY9H.DE 11.7% $BDSX 9.5% $DAVE 9.4% $7974 8.6% $LOCO 7.1% $9984 5.9% $MKSI 5.9% $FTAI 5.2% $XGN 5.0% $2454.TW 4.9% Short positions (-91.2%) $MTUM -28.4% $EWT -11.4% $ARM -5.9% $ARKG -2.8% $9434 -1.9% $CRWD -1.7% $KORU -1.6% $ARKK -1.5% October has been a pretty intense period of work as I try to fill non US/Asia markets on ALPHAPORT (Europe and Canada atm) and bring translations online for all Korean and Japanese filings (mostly covered for 2026 now, and all filings going forward, doing my best)... While also juggling a highly dynamic trading environment. Being in California has also made responding to market moves quickly a bit trickier. But i have managed through the stress ok I think. It just sometimes feels like real cowboy shit smashing an $ARKG short on my phone from the bathroom at 6:30 am. But hey, i’m outperforming most the BBG bros. Maybe the additional time for research and study is more important than time spent looking at charts? And I am always on top of the increasingly important Asia opens. Three big things happened this month wrt the portfolio: 1. The AI Bio Trade hit bubble territory Previously I had written that i had hedged out some ai/bio risk on my $TEM / $CAI longs with a little short $MRNA / $TWST. This was prudent although neither worked in the short term, but it got me in the right headspace to respond quickly as $ARKG hit a fairly bubbly peak and I got some nice short trades in $ARKG, $TXG, $RXRX, $DNA etc, which also hedged out some risk. But maybe could have gone harder and potentially waited slightly longer. I was too early selling $TWST as Emily Laproust dumped huge amounts into the market around 155-160. Basically I just copy traded her. After all, she was selling at 20x sales. But it did sting a bit to have it work out and end up a couple % so far. So it goes. Unfortunately, although my shorts on $TWST, $ARKG, $DNA, $RXRX and even $TXG from the top were all profitable and well executed, $MRNA ripping here was painful. When your longs all drop, but then one of your correlated shorts goes up, well, that is not a great outcome, and why even highly correlated L/S positions do still add risk as gross exposure rises. That said, we are talking about small things. Overall performance has been ok on this pull back. I feel like a lot of the complaining from $MRNA shorts must be from people going WAY too big because they aren’t used to running a short book and can’t find that many credible shorts in biotech. Like, are you putting on a 10% short or somthing here? That is not a good plan even if you have very high conviction. I still like the longs like $TEM and $CAI and especially $BDSX here which is back to my entry point and objectively cheap relative to the rest of the bio-diagnostics space. I think to add $TEM we would need $60 or below at this point though. 2. The AI trade Well.. i sold down a lot of $INTC having hit LTCG. I didn’t really like this, but i just felt like the spread at $120 INTC had gotten too wide against other AI longs to justify a much bigger position than normal. I did keep a modest position. My timing here was good. That said, I also took off an SMH short and added to Samsung Preferred after the Samsung results - only to have the trade nuke the next day on fears that OAI’s earnings were exaggerated. I think it was a combination of the two things - Samsung down on good earnings (the most money ever made by a company in a quarter!) and then the OAI stuff. I was pretty frustrated by my timing here but honestly writing this out now, it really was just variance - I reduced exposure with the INTC sales. I didn’t love that i was longing Broadcom while still being short SMH. And I continue to believe Samsung Preferred is way too cheap here. We will see what happens on Monday. Kind of two wolves in me on this one, part of me wants to just ride into the sunset and stay 15-20% long AI, the other half wants to go 10% long Samsung Preferred if it sells off further, given the strength of their memory business and the growing importance of Foundry (plans to work on an Anthropic chip seem incredibly bullish to me). People say there’s no catalyst, but if they dedicate 50% of their earnings to capital returns in the form of dividends, the Samsung Preferred dividend will be very large. And unlike SK Hynix, where all capex is a sign of a potential overbuild, with Samsung you actually do want them investing money in the non-memory foundry business, which is a good business long term and will also absolutely rip if China ever makes a move on Taiwan. 3. Short cybersecurity and software I have been short cybersecurity for a while (a quick check of my portfolio shows a $CRWD short down 37.5%, oops) and I remain confident that it has hit bubble territory here and trades purely on vibes. I doubt most investors in the company even understand what CRWD does in the cyber stack. Go on, explain how their business works. With that in mind, i have shorted across the theme further, with even the super high quality stuff like $PANW and $NET looking too expensive to justify even dominant market positions (seems like a lot of different companies in this space if these are monopolies?). $OKTA is also indefensibly expensive at this point in my view. In my view, there's a few issues with cybersecurity as a long even though I think it is hugely important. 1. Google bought Wiz for 32b so they could integrate cyber capabilities into the cloud directly - this will be an amazing acquisition long term. But Google will use the Amazon model to price it, eg every slice of their business will be cheaper than their competitors, but they will still end up taking 50% of your revenue somehow. All the hyperscalers will do something like this. 2. The companies making the AI badly need to make more revenue, and cyber capabilities are obviously something they can do. I mean, they already are. If you want Mythos to review your codebase you can ship Anthropic 10k or so and Mythos will go over it with a fine tooth comb and make sure you’re not going to get hacked. Again, like Google’s positioning of Wiz, Anthropic and OpenAI have an extremely strong entrypoint into the Cyber business. You just set the dropdown menu to the cybersecurity model and say 'let it rip!' and it'll go over your codebase and help you remediate immediately. 3. The money to pay the cyber companies has to come from somewhere. Companies are already having to do significant layoffs to increase token spend budgets (fire 20% of your staff and use half the money saved on AI tokens and you have a way better business). Ramping up cyber spend to the level these companies are underwriting is a big stretch. 4. The AI is basically already taking over your pc at this point. Two issues 1. The frontier labs want you to use a cloud environment long term, not your local pc - but CRWD is getting paid using your laptop as a subscriber. They do cloud security too, but that is very competitive. 2. Having the frontier lab AI secure your physical laptop might make a hell of a lot more sense than having CRWD do it - after all, its OpenAI you are giving Full Access to, not CRWD. The bull case is that this theme obviously has legs. A young chinese individual was reported to have hacked a string of Korean financial institutions using AI just a few days ago. But I tend to think that the frontier labs will quickly figure out how to stop this sort of thing and end up taking a fair bit of the economics around the theme. Between $PANW, $CRWD and $NET you have almost 1 trillion of market cap with less than 20 billion of revenue. I don't see it. It feels a lot like $SHOP at 50-80x rev over COVID - incredibly well positioned, but also very overvalued in the short to medium term. We shall see. As always, if you have any questions, feel free to ask here. And my full portfolio is available to check out with the link in my bio.

Do these guys read Investing with Martin or something? My $LOCO writeup here: martinbradstreet.substack.co… I own shares, not a suggestion to buy them yourself. If you're looking for fresh flame grilled chicken at value prices in California however, I'm a fan. And if you're totally broke, their $5.50 Pollo Bowl has 41 grams of protein and is 600 calories. Tough to beat in this economy! So i guess that part is budget conscious financial advice - don't spend $20 on lunch if you have credit card debt, buy the Pollo Bowl.

People have been posting bearish $SMH charts but I have my own chart inspired by the great man himself

Hard to say what the market will do here when the narrative is 'all good developments are well known thus the price should not go up', but i am a troglodyte so i am simply satisfied with the 370.3% YoY growth, at least for the time being.

the most profitable quarter ever by any company, but will it be enough to justify Samsung's ($005395.KR / $005930.KR) forward PE of 4?

What does it take for luxury to rebound here? I look at Hermes ($HESAY) and $LVMH ...i'd love to buy this stuff when it is distressed. But I don't know why 30 PE is acceptable here?

It's a small thing but since all Japanese filings now come into ALPHAPORT.AI in English i can put the filing titles in the Feed and surface them for stocks relevant to your portfolio. Just click to read the report in English. In the screenshot here, there's some important stuff that I would not have been on top of normally. - Monthly recurring revenue was released for Robot Payment ($4374.JP) - standard YoY was good but sequential down very small. - Softbank finishing the last tranche of investment in openAI - A quick look at a filing from $4377.JP shows their business strategy, embrace of AI and ROE targets. These are small things, not worth laboriously translating documents one at a time, but it helps you tend to the garden. Try it out at alphaport.ai/subscribe with a one month free trial, your feed will start populating straight away.