Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

This post is for those of you who read the Cerebras deep dive and are curious to continue following the story in the months and years ahead. A couple of evenings ago CBRS held its first earnings conference call. There are already plenty of ‘hot takes’ out there following the results - this is, emphatically, not intended to be another one of those.

Everything about CBRS currently is very very noisy: the financials, the punditry, the share price… Just to update you since the deep dive, the IPO was priced at $185 but the shares opened on 14 May at $350 and hit an intra-day high of around $400. As I write, the shares are now below the IPO price having plunged 20% yesterday in the earnings aftermath:

As I suggested at the time of my deep dive, the post-IPO reaction is all turning into a bit of a circus. So the point of this post is to take stock of where we are following the IPO and what relevant signals there might be for the future - on a multi-year view.

One important recent development is the commencement of sell-side coverage just a few weeks ago. And what’s really notable about the Q1 earnings is that it kind of looks as though CBRS management were, in their first earnings event, trying to engineer a big beat and raise on both revenue and margin relative to expectations previously shared with analysts - and this is in fact what they did. That the shares reacted so negatively appears to have been a surprise to them (e.g. see here and here). But it might be less surprising to observers of the Beauty Contest. Along with the beat and raise, they signalled a willingness to take some short term margin pain for longer-term gain - the type of act generally not looked upon favourably in present market conditions. Neither does it help that one of CBRS’s greater perceived fragilities is around the margin question - room for doubt. Throw in some second-, third- and fourth- orders of guessing along with a thin float, and a single-day correction of 20% seems fairly run-of-the-mill.

One interesting question worth grappling with is the CBRS valuation at this stage. If the consensus 2028 revenue numbers are in the right ballpark (and the CFO has suggested they are), then the stock is perhaps in the region of 5-6.5x EV/S on a 2028 timeframe (depending on what share count and net cash position you use) - putting it at a significant discount to many other semiconductor stocks (and practically all of those with links to AI). And it’s also growing much faster. But it’s not necessarily to say it’s low risk - the burden very much appears to be on management at the moment to prove that they can execute smoothly and simultaneously on a massive manufacturing and cloud capacity ramp. That said, once you can get past all the hand-waving about high multiples on trailing revenue and such things, CBRS may even be in the process of emerging as a cheap stock following the not-entirely-unexpected post-IPO blowup. For my part, I still own my pre-IPO shares. I will, of course, let full subscribers know if I change my holding in any way.

Now onto the meat and potatoes…

[NFTBC does not give investment advice. Please do your own research.]

Subscribe now

Share

Read more

Friends of NFTBC

About a week ago a sell-side analyst downgraded PRCT on concerns around competition from PAE:

The stock has gone down in a straight line ever since:

So the plan here is to briefly revisit the PRCT thesis and talk a little about how PAE might impact it, if at all.

First of all, I’m working on the assumption here that the Leerink downgrade is in fact, the catalyst for whatever is going on with PRCT currently. It does coincide perfectly with the start of the slide and the threat of PAE is something that analysts often like to raise on PRCT’s conference calls - so it’s been something of a latent fear residing in the background. Otherwise, I would just assume that some investors are taking the opportunity to position for a negative guidance revision. It’s fair to say there’s a degree of scepticism about PRCT’s guidance for the rest of the year, which entails an acceleration in H2. At NFTBC I don’t work with alt data or such things, so I have no special insight into that. My strategy is more to give Larry the time and space to carry out his strategy, all the while trying to make a sensible ongoing assessment of whether it’s going to work - in all likelihood this is going to be more of a multi-year process rather than making strong conclusions from one quarter to the next. Having said all this, PRCT management were supposed to appear for a fireside chat at a Truist medtech conference this week and I was looking forward to hearing what they had to say. But this event seems to have been cancelled. Some may interpret this as incrementally suspicious but equally these things do get cancelled from time to time for innocent reasons. At the time of writing, there has been no 8-K or suggestion of a material development.

Subscribe now

Share

[NFTBC does not give advice - please do your own research. I currently own shares in PRCT].

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

It was exactly a year ago when I wrote to you about the outright despair that had taken over REGN sentiment at the time. Unwittingly, I caught the exact bottom with that post, with the shares going on to appreciate by around 70% over the following 6 months.

…Yet here we are a year later and we’re back in much the same position from a sentiment and valuation perspective:

Once you factor in REGN’s massive $16bn net cash pile, the effective P/E is 8-point-something on consensus numbers for the next 12 months. 8x earnings is so low, that you’ll be hard pressed to find anything much lower in this sector - not even in the stodgiest dregs of old big pharma. And don’t forget, REGN invests in R&D at twice the rate - on a like-for-like basis the P/E drops to just 6x. Moreover, on consensus numbers (which I think are too low by the way), REGN is set to grow its earnings by 70% out to the end of the decade. This is why, at various times over the last 12 months, co-founder Len Schleifer and others have claimed that you can arrive at the company’s entire market value just from the cash flows of Dupixent.

Here’s a similar and very simple way of thinking about it:

Note I’ve picked consensus net income here for simplicity (and because it’s the lower number than free cash flow). And you can pick your own discount rate, but the point is that earnings between now and the early/mid 2030s account for 100% of REGN’s current unlevered market cap or enterprise value.1 Or in other words, the implication is that REGN has zero terminal value beyond the mid 2030s.

Now, of course, I don’t think that anyone really believes this - that would be silly. I think one of two things is going on. Maybe investors have cleverly sussed out that REGN will very shortly go into steady decline and will be much smaller in the 2030s than sell-side analysts collectively think. For example, REGN would be approximately fairly valued today at a terminal multiple of 10x if you assume it makes 40-50% less money than consensus over the coming 10-year period. Even if someone only had a moderate understanding of REGN’s core products, collaborations, IP and competitive landscape - this wouldn’t be much less silly than the notion of zero terminal value in my view. So here’s the more likely explanation: investors are collectively just playing their own games, and the market cap is merely the outcome of those games at this snapshot in time. I’ve written about these games before, and in fact the very name and premise of this blog comes from the Keynesian Beauty Contest. There’s no shortage of reasons to positively downvote REGN’s in the Beauty Contest right now:

• Two phase 3 misses in the last year (and extrapolation therefore / Gambler’s Fallacy)

• Ongoing Eylea approval delays in the US (Eylea fears being the traditional bête noire with this company)

• A handful of CRLs in recent years (and the assumption the problem is structural)

• Dupixent loss of exclusivity creeping into the distant horizon

• Monkish abstinence from big M&A

• Investing more than twice the industry average in internal R&D2

The last two in particular render REGN something of a total weirdo in the minds of many players - and surely no one’s going to vote for the weirdo in a Keynesian Beauty Contest. If only REGN would behave more conventionally - throw on a little Instaglam if you like - then perhaps the players would take more notice. Or so it might seem. But we all know about unconventional beauty - and it’s arguably the more profound and enduring kind. If there’s a purpose to this blog, it’s in seeking out and exploring some of these cases. Naturally, I make no promise that I’ll always be right about them. That said, I suspect I am right about REGN - only time will time. Time being the operative word here. As I’ve said a few times at NFTBC before, I don’t even think the bears are wrong at all - we’re just playing different games. Games with very different time horizons for the most part. Which brings me to…

…the more serious part of this post. So why do I believe that REGN deserves a terminal value much higher than zero? For one thing, I was struck by something co-founder George Yancopoulos said earlier this year:

A lot of people have a mistaken belief that it's easy to make antibodies and all antibodies are created the same. More fully human antibodies have been approved from our platform than any other platform, 10% to 20% of all approved antibodies come from our platform. And we repeatedly are producing the best agents with the least number of problems such as antidrug antibodies and so forth. And so when you have 2 antibodies against the same target, they’re not necessarily the same. And across many programs. Remember, other people have tried to make DUPIXENT-like antibodies that all failed, including Amgen. So it’s not that easy to always make the best antibodies. Our platform and our people have perfected that ability. And that’s why our antibodies seem to work in places where other antibodies against the same targets don’t work.

I had never thought about it in those terms before, but it’s true. Depending how exactly you count it, somewhere in the mid- to high-teens percent of all fully human antibodies currently approved came out of REGN’s VelociSuite platform. The odds that this platform continues to turn out great therapeutic antibodies therefore seem very favourable - VelociSuite’s not worth nothing anyway. I have written about it before, so I won’t repeat myself here. But I thought it might be more interesting to write about another important and overlooked contributor to REGN’s terminal value: platform partnerships. And this is where we come back to time horizons.

The central platform partnership in REGN’s commercial operation is, of course, the one originating in the Sanofi antibody collaboration - established back in 2007. VelociSuite was the platform and Sanofi were paying to access it, on pretty favourable terms to REGN I might add. There were various additions and amendments made in the ensuing years. Here we are almost 20 years later and antibody therapeutics born of this alliance are currently annualising at around $23bn and together grew more than 34% in the latest quarter. A great pharmaceutical collaboration can take years and years to really pay off. But who in this industry has a 20 year time horizon? Who even has a 5 year horizon? Certainly not your typical investor anyway. As Len recently noted, he’s worked with 8 different Sanofi CEOs during the collaboration. It’s almost impossible to imagine that the collaboration would have been as successful if the same turnover and constant rearranging of the deck chairs had applied on both sides. Despite Sanofi’s numerous unforced errors3 this is arguably the most successful pharma partnership of all time. But what if instead of Sanofi, both parties were more REGN-like in time horizon and innovative by nature, and the resulting value creation was more multiplicative rather than one sided? And what if instead of one such partnership REGN had 10 or so of these with new ones landing every year? These are the sorts of questions that I believe should feed into the conversation about REGN’s terminal value. Why? Because it’s literally their strategy, or a major part of it anyway.

As a reminder, via its sequencing operation RGC, REGN plans to sequence the genomes (and increasingly, proteomes) of tens of millions of people in the years ahead, with the specific intention of using this unique resource to generate insights and invent innovative therapies. But only a portion of this tidal wave of genetically validated drug targets will be optimally addressable with plain vanilla VelociSuite-derived antibodies. REGN is well aware of this hence their desire to expand into new modalities. While some of this can be done in-house (e.g. bispecifics, ADCs), much of it is better achieved through platform partnerships where both parties can bring something valuable to the table: world-leading target discovery and antibody technologies on REGN’s part and world-leading capabilities in new modalities and related technologies on the part of partners.

By the mid-2030s I would expect REGN’s revenue from partnered products to be one of the fastest growing parts of the business, albeit still a minority share relative to pure VelociSuite products. The evidence for this is the fact that ~60% of REGN’s phase I pipeline is constituted of partnered molecules, most of which from the Alnylam partnership. Including also the partnered assets already in phase 2, one could well envisage waves of new approvals during the early- to mid-2030s period - right at the time when REGN is implicitly supposed to be worth nothing at all. And we can only imagine what the pre-commercial partnered pipeline might look like at that stage. But outside of REGN and its partners, no one is thinking that far ahead. We’re talking here about the event horizon, beyond which the analysts’ models cannot possibly see. And they’ll give you all sorts of very good reasons why: this industry is far too unpredictable to model that far into the future - fair enough, I don’t disagree. But I ask: do you even need to model it? At today’s prices you’re getting the 2033 instantiation of REGN essentially for free. The pharmaceutical industry being inherently fat-tailed, this makes the option-like nature of REGN kind of interesting to me. REGN is built from the ground-up to exploit these fat tails and stay in the game long enough to reap the rewards:

1. Platforms, platforms, platforms: lots of low-cost bets through repeatable constantly-improving platforms, rather than one-off assets and singular bets acquired through competitive bidding

2. Skewing the odds of success: it’s not only about shots on goal. The centrepiece of George’s philosophy is removing bottlenecks in the R&D process. This is what VelociSuite/platforms and RGC are for: making the shots on goal faster, cheaper, much more precise - more option-like

3. Pipelines in products: one product, multiple indications, massive payoffs

4. Unconventional behaviours: time horizons measured in decades rather than quarters, optimising for fat tails rather than smooth earnings, avoiding the average, being unafraid to look weird/wrong etc

5. Massive pile of cash: safety + optionality.

I think there are two ways to play REGN. If you’re skilled at reading the tea leaves - then go for it and play the short term game. This stock moves around a lot (actually for this reason) and that creates opportunity. Alternatively, if you’re as fascinated as I am by the things going on at this company you can use such occasional opportunities to build yourself a stake and settle in for the journey whenever that may go. One way or another, it’ll be fun.

Now lets talk some more about partnerships.

[NFTBC does not give advice - please do your own research. I currently own shares in REGN]

Subscribe now

Share

The idea of doing some version of this post has been percolating for a while. I have written about the Decibel platform before and the increasing importance of the Alnylam partnership has come up a few times too. But REGN’s recently announced collaboration with Parabilis Medicines was the eventual catalyst - it looks like a really interesting partnership.

Overview

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, updates, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Friends of NFTBC

Welcome to The NFTBC Digest - a regular email that normally goes out to full subscribers only. Since I started it last year, The Digest has essentially developed into a repository for all the material that I have decided not to make into standalone posts - generally news, earnings, links and occasional updates around my own portfolio activities. Every once in a while I’ll do a wider-circulation email, principally to ensure that newer joiners are aware that The Digest exists. This week’s will be slightly longer than usual as there were plenty of things worth writing about with regard to NFTBC stocks.

The week certainly had its ups and downs. I’ll dispense with my usual alphabetical ordering and get straight to the bad news.

Regeneron

On Friday evening Regeneron provided an update on their long-anticipated fianlimab-cemiplimab combination trial in metastatic melanoma. The trial did not reach statistical significance for the primary endpoint of improvement in progression-free survival (PFS). Regeneron has inevitably taken some heat for releasing the results on a Friday post-market close, but per the press release they only reviewed the results themselves on Friday. The press release itself contains not a single word of commentary, only headlines and data - which seems rather unusual. My guess is that Regeneron probably felt the need to get this news out as quickly as possible before having had sufficient time to process it in any helpful manner, but no doubt others will have different takes.

It’s clear that Regeneron was expecting a positive result in this trial, as was I. However, the timing of the readout was previously deferred a couple of times due to the unexpectedly slow accrual of events in the study - this has been well communicated and discussed and consequently there was already some degree of building scepticism prior to Friday. Additionally, some months ago they amended the trial protocol to include additional patients from the overall survival cohort in order to be able to accumulate more events.

Last year they provided two possible explanations for the slow rate of events - either the pembrolizumab control arm was performing better than expected or the study drug was delivering long and durable responses. They also mentioned they were looking for a PFS in the low- to mid-teens of months in order to be differentiated against the existing field - and that’s against the context of the 24 month PFS they reported from their pooled phase I cohorts.

Here are the headlines:

You can see the high-dose combination (the principal dose of interest) failed on statistical significance - and it seems to be a particularly narrow fail as indicated by the upper bound of the hazard ratio confidence interval only very marginally crossing the 1.0 threshold. In any case the 15.5% relative risk reduction likely falls short of the company’s expectations. Median PFS at 11.5 months arguably just makes it into the low- to mid-teens they were hoping for at a minimum (if we take that to mean 11 to 15).

Notably, they were only expecting the pembro arm median PFS to land somewhere in the range of 4 to 5 months when they designed the study, although they previously said the study was conservatively powered to assume 7 months. In the event it was 6.4 months - which suggests that not only did pembro outperform their expectations, but the fianlimab combination underperformed their expectations. I think our working assumption has to be that the fianlimab combination is not as good as Regeneron was hoping.

However. We are told that detailed results will be presented at an upcoming medical meeting and presumably the data will continue to mature (the study officially runs until 2031). One plausible theory is that at the cutoff date there was a high number of censored (i.e. event-free) patients in the fianlimab arm who may go on to have durable responses beyond the 11.5 months - especially in the context of the 360 overall survival cohort who were dropped in later (and enrolled January to late Summer 2025). In such a context, the study could still have been underpowered at the analysis point. With maturing data, it’s not impossible that PFS goes up and hazard ratio improves. How likely this is, I just don’t know. There’s also the overall survival data. While I’m certainly not betting that this trial delivers subsequent positive surprises, it’s probably premature to draw any strong conclusions about fianlimab before there’s more data in hand.

The good news is that fianlimab still has a further shot on goal in metastatic melanoma coming up in around 10 months - the head-to-head against Opdualag, another LAG-3 combination that’s already on the market. A convincing superiority readout against Opdivo would certainly be helpful in reviving hope of approval and a meaningful commercial opportunity. I think there’s a decent chance that the fianlimab combination could still stack up favourably against Opdualag. That’s partly because Regeneron likely has the better anti PD-1, partly due to their being able to administer their anti LAG-3 at a higher dose and partly due to the relative ORRs they’ve previously shown. We’ll just have to wait until Q1/Q2 next year to find out. But what might make the difference in commercial traction will be the eventual overall survival data, if it’s encouraging.

I wouldn’t write off fianlimab in metastatic melanoma just yet, although in the absence of further data we should downgrade our expectations and push the timeline out by another year. In my mental model, fianlimab was not one of the sizeable opportunities that was going to have a major contribution to intrinsic value. Rather, it was going to be incrementally helpful out to the end of the decade in the interim phase before before some of Regeneron’s bigger swings start to come to fruition. For some context, Regeneron recently said they believe the global metastatic melanoma market opportunity to be between $2bn and $3bn, of which Opdualag does ~$1.2bn.

A convincing win in the pembro head-to-head would certainly have been very helpful for sentiment around the stock and the company in general. But, alas, I suspect this failure will be received in a similar way to the itepekimab failure 12 months ago. As such, we should be braced for a severely negative market reaction next week with this readout being construed as another loss of confidence. The market seems to be much more of a mind to focus on Regeneron’s failures rather than the successes at this moment. But it’s important not to overlook the successes - for example the C5 program is the much more interesting commercial opportunity than fianlimab in my view and it’s coming soon.

An outstanding question is whether Regeneron blundered in designing the pembro head-to-head trial in the way they did or if they just got unlucky. I don’t know enough to have a view - if you have one, I’m happy to hear it. I’m not a clinical trials expert, but I know some of you are. I have left the comments section open, so feel free if you would like to add something or otherwise explain where I have erred.

Now back to alphabetical order…

BMW

In the last Digest I suggested the possibility of a standalone post following the AGM this week. In the end I decided against this. Last year’s AGM was unusually eventful (so far as BMW AGMs go) with the first public showing of the (camouflaged) iX3 and i3 Neue Klasse models as well as a rare speech given by controlling shareholder Stefan Quandt. This year was more business as usual, although did mark the last occasion with Oliver Zipse as chair of the executive board. The renewed plan here is to post a few combined notes and interesting quotes from the both the earnings conference calls (analysts/media) and AGM.

The most notable thing that came up was the idea of export credits in the US. Apparently BMW is in discussions with the US administration about implementing an offset agreement whereby BMW is allowed to import one vehicle for each one it exports. Per comments, the administration seems to be very supportive of the idea and “there’s a very good chance to implement it” - but it all hangs on the EU first implementing their side of last year’s trade deal first. This would clearly be tremendously positive for BMW’s margins given that they import as many vehicles into the US as they export (i.e. net imports are roughly zero) - and it would allow for a faster recovery to the 8-10% strategic automotive margin corridor.

Speaking of which, in my last BMW post I modelled a return to the 8-10% corridor in 2029 along with a potential doubling of automotive profits in the 2025-2029 period (via a dip in 2026). The corridor came up on the analyst call:

the 8% to 10% is not unreachable. Don’t forget that it is 3 x 1 EBIT point we have to achieve. One is on the performance side. We love our BMW ecosystem, and we can do even better. Not to forget Alpina to come, which is also positive for the contribution side. Not to forget that the Gen 6 Neue Klasse has a much better contribution margin than the Gen 5, as we all know for a while. So that’s the performance 1 EBT point.

Second one is, of course, we are doing our homework, working on the material cost side, on the manufacturing cost side, on the warranty cost side. If you have a better quality, you have to provision less warranty costs, which is also contributable to the EBIT margin, not to forget and of course, also logistic costs, et cetera, to be optimized with our global footprint. So that’s the second EBIT point.

The third EBIT point is also homework, is our fixed cost levels, which we are also coming down, as we mentioned already, step by step in ‘26 and ‘27 and also in ‘28. Not to forget the BBA purchase price allocation depreciation, which is always accounted for 1.1, 1.2 EBIT points. This is only lasting until mid- ‘28. So you will have already a half year effect in ‘28 and a full effect in ‘29.

So with these elements to come, there is a chance to come back to 8% to 10%. Of course, we are not naive. There is always something coming around the next corner, whether there are tariffs or other crisis points, and we mitigate as they come along. And we have plans already to do so.

“Not unreachable” / “there is a chance”. I’ve been following BMW long enough to have developed an ear for ‘BMW Speak’ which is characterised by wiggle room and conservatism. My translation is something like this: “We’re gunning hard to get back to 8-10% in the next several years and have a well-developed plan to do it - but we’re not giving guidance on it”.

On differentiation versus ‘monoculture’ in the China auto market:

when you walked over the Beijing Auto Show and compared what the BMW Group was representing and what the rest of the show was representing by and large, there was a very focused presentation of most of the players in that industry. The majority was electric. The majority was the same segment of cars, very few sedans, a lot of SAVs of larger size, focus on autonomous driving, it’s all the same. It was almost a monoculture. And of course, markets worldwide, they are not monolithic. They are widespread. They are differentiated by drivetrains. They are differentiated by size. They’re differentiated by price segments.

And then you looked at the BMW Group stand. There was a large portion of MINI with all drivetrains available. Then you look at the BMW stand, you saw individualization. You saw all drivetrains. You saw V8s. You saw, of course, the Neue Klasse, fully electric. You saw smaller cars. You saw the new 7 Series. You saw our focus on hydrogen. So the full breadth of a premium player, you saw that. And that differentiates BMW. The variety, which represents the variety of market segments. And you don’t have to look on a global scale. You see that in China as well. The markets are much more wider than you would see on that fair.

And that is a worrying thing. If the whole industry thinks they can concentrate only on one segment, larger cars, SAVs, electric, then something is wrong. On the other hand, this makes us more resilient, antifragile to offer the whole breadth. Let’s just look at the first 3 months. EV sales in China were plummeting, not only decreasing, they were plummeting. And what happened, our ICE sales went up at the same minute. So especially China shows what it means to be a full segmenter, to be technology open and at the end, also to be a global industry. That was proof that our strategy is right.

Much is being made currently of the entry of the Chinese into the premium market. When I originally wrote the deep dive, I had some of these things in mind. It’s one thing to pack in a load of features and undercut competitors on price (which gets much of the media attention in the moment), but it’s another altogether to have a desirable and successful premium brand with longevity - I think this is a really under-appreciated characteristic of BMW that gets lost in today’s noise. It’s really really hard being a successful premium automotive brand, but BMW is uniquely good at it. Zipse added some useful thoughts on the topic:

on the strategic side, what is premium in the future? Premium of the future is product quality, specifically long-term product quality. It does not make sense to push cars in the market who lose their resale value after 2 or 3 years’ time. This is not premium. So it’s not feature count in a new car, it’s what is the residual value after 1, 2, 3 or even 5 years. I’m not even talking about historic cars of 15 or 20 years. That is premium.

The second thing is premium system integration. It’s not feature count. We discussed what we saw in Beijing, again, feature count over future count. That’s not the point. How does that all play together? How does it fit? And is it for someone who drives the car, not while it’s standing, but while it’s driving, does it fulfill the function they are supposed to?

And the third one is safety. And I think that becomes one of the foremost criteria for premium of the future, safety, no casualties, no risks. Absolute promises to a brand who promises to be premium means no casualties, no risk in driving the car. And I think that will define premium of the future, and I think we are well underway to fulfill that promise.

And, relatedly, on who will be the survivors in the next five years:

I still think there’s too much focus from the media, but also from the capital markets on individual players, individual technologies, singular events. In 5 years’ time, the players who will still survive, they have the competence to system integrate, to bring all technologies into a car, who build cars who have a long-term quality, who are not falling apart after 1 or 2 years, who will still be in service after 5 years. It’s a completely different game which is happening there, and who are able to differentiate between a singular hype and an overall business model.

If you watch the industry today, actually, there are too many bets going on, too many bets on a singular technology that might happen, but this cannot be endured. So much money is lost in this industry on bets. So system integrators who are able to follow regulation and at the end, have still a business model on high-quality products, that’s the key of everything. And in 5 years from now, to follow CO2 regulations specifically and remain a profitable business will be more or less the key to everything.

Some interesting facts about BMW in Germany (that came up during the AGM):

• one in three customers who have ordered the new iX3 are new to the BMW brand

• in 2025 they built more than one million cars in Germany - the most ever. 25% of all cars built in Germany are now BMWs. I have a strong impression that few investors/observers appreciate the contrast with the wider German automotive industry here

Next - conversion of preferred shares to ordinary shares. This was approved at the AGM. One thing I actually hadn’t appreciated was the possibility of a technical benefit to the share price from the conversion and a strengthening of BMW’s weighting in the various indices.

One last thing: BMW has now unveiled its first Vision Alpina model. As a reminder, a couple of years ago BMW licensed the historic BMW-underpinned Alpina brand from the Bovensiepen family with plans to relaunch Alpina has a BMW sub-brand positioned between the highest-end BMWs and Rolls-Royce. This is presumably Bentley and Aston Martin territory with ASPs in the EUR 200,000 to 250,000 range. Alpina production models are due to hit the roads starting in 2027. It’s worth noting that CFO Walter Mertl specifically referenced their expectations of a positive margin contribution (see above). As I’ve noted before, Rolls-Royce ASPs are over EUR 500,000 and they make, I suspect, margins at least in the higher 20s percent range on 6,000 units per year - it’s potentially in the region of 15% of group automotive EBIT. Alpinas were previously very rare with around 2,000 sold per year prior to the license deal, although there’s perhaps no reason in principle why they can’t drive this to 6,000 (~Aston Martin) or even 10,000 (~Bentley) if they’re particularly shrewd in their positioning/strategy. Assuming margins in the low 20s (such as Bentley managed until recently) it becomes possible to envisage a 20-30bps automotive margin contribution by the end of the decade - hence why I assume Mertl brought it up. Here’s BMW’s video of the Vision Alpina:

Cerebras

Following my recent deep dive Cerebras has now gone public. You can watch the five founders ring the opening bell here. I will now start accounting for my shares in my listed stock portfolio and you’ll see further updates at NFTBC. Until the lock-up ends in September, effective free float is only 16% at the moment and according to comments made by the CEO, the IPO was 25x oversubscribed - in other words, some more reasonable measure of price discovery perhaps won’t come until later in the year.

Cerebras’ CFO Bob Komin was interviewed on CNBC and a couple of notable points came up:

• He said Cerebras’ valuation was probably trading in line with Nvidia on 2028 revenue. With Nvidia at ~8x calendar-2028 EV/S this might imply Cerebras is anticipating a much more aggressive revenue ramp than what I modelled in the deep dive. I said in the deep dive that the range of outcomes is particularly wide, and this also includes the possibility I was too conservative

• They are already serving one trillion parameter models, including ChatGPT 5.4 and 5.5 internally for OpenAI. These are not yet released to the public “but that’s coming very soon”. This is one of the key debates around Cerebras. A one trillion parameter model might require 46 or more CS-3s to fit it, at $1 to $1.5m each - so there are serious questions as to whether that is economical relative to the throughput. I’m assuming there are various co-designed elements to it that we can’t see from the outside including things like disaggregation. But I’m also starting to wonder if the relationship could eventually involve an element of sparse pretraining (weight sparsity) which is only possible on Cerebras hardware. It was supposed to be one of Cerebras’ USPs before the pivot inference in 2024 and everyone seems to have assumed it was a dead end - but is it..? In any case, if they start serving 5.5 at high speed soon, I think that’s going to be a really big deal - there won’t be anything else like it

Ottobock

Ottobock announced that it is acquiring Blatchford’s Norwegian patient care business. The price is EUR 110m and comes with EUR 40m in revenue. In effect, this signals Blatchford’s retreat from patient care outside of the UK. This is actually quite interesting. I discussed some of the competitive dynamics in the deep dive - I think Ottobock’s forward integration model is going to be very tough to compete with and Blatchford has apparently dropped its decade-long ambition to try the same.

Procept

Procept spoke at the Bank of America Global Healthcare Conference this week. CEO Larry Wood provided some helpful context around the prospect of a hybrid support model that they started talking about recently:

The going to a hybrid support model is really not about trying to drive leverage out of the P&L. What it’s really about is trying to drive procedure growth. And what I worry about is that we could be losing procedures here and there because a doctor sits there on a Thursday and says, I got 2 more cases tomorrow, I could do Aquablation, but I don’t want to call the rep or maybe he’s always [out of town] on Friday.

And what I don’t want to do is ever lose any cases because we can’t get somebody there. I want our team routinely visiting with our customers. I want them routinely making sure the system setup is being done properly. There’s a lot of turnover of staff in hospitals, and there’s a setup element to the robot, and we want to make sure that people are doing that right and doing it correctly.

So I always want our people in that. There’s never going to be something where we don’t see a customer for 3 months and they’re just written off cases. And I think there’s actually - at some level, there’s a risk to the business that you quit paying attention to your customer and you just open the door for other people to come in. So we’re always going to be there doing that.

I just don’t want somebody dependent on us. I don’t want somebody saying, if a rep can’t get there, if a rep is sick tomorrow, I can’t do cases and thus I’m going to have to convert them to something else. So I just see this as a lever for us to be able to continue to drive case growth more than I see it as an operational lever. Now the byproduct of it is if we can cover - if we can double the number of cases and we can keep our sales footprint the same or grow slightly, obviously, that drives tremendous leverage for us. But it’s a byproduct, it’s not the reason.

That’s it for this edition. Thanks for reading.

Get in touch if you have any comments or questions.

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Friends of NFTBC.

As seasoned NFTBC readers will know, I am particularly drawn to stories of unconventional business success. Naturally, these sorts of stories are rarely obvious in the beginning - so any satisfaction in the pursuit comes from following them over time and living inside the story while it writes itself. All the better, if I can make some useful investment returns along the way. Cerebras Systems is an example where I have been following the story for about five years and an investor for just over two years. With Cerebras having recently filed an S-1 and an IPO looking imminent, now seems like the right time for a post. But I have been of two minds as to whether I should write one. AI, IPOs, hype - these are certainly not the traditional NFTBC territory you’re used to, and to some extent they did put me off from wanting to involve myself. But on the other hand Cerebras is, in my view, an extraordinary business - and a fascinating one. It’s worth writing about. I’m also well aware I’m posting this at a time of especially high enthusiasm for semiconductor stocks - and this unsettles me to some degree. But equally, if ever there was a propitious moment for Cerebras to raise needed capital on the stock mark, this is it.

This post is intended for curious readers with an interest in unconventional businesses doing interesting things. It’s probably less relevant for those who want to go deep into the weeds on engineering questions. But all are welcome. Additionally, the post will be much more concerned with the business itself rather than IPO dynamics.

Let’s begin.

[NFTBC does not give investment advice. Please do your own research.]

Introduction

Since its founding a decade ago, Cerebras has been in the business of inventing, producing and selling a highly unconventional computer system with specific design intent for AI applications. What makes Cerebras’ system so unusual is the way they’ve gone about it - using an entire monolithic 300mm silicon wafer as the basis of their processor, rather than dicing their wafers up into small chips like everyone else does. Doing it this way also necessitates rethinking the entire system as well as the manufacturing process and supply chain: from packaging to power to cooling to software. And there are good reasons why no one else does this - for decades it was supposed to be a dead end. Cerebras’ founders decided to do it anyway, succeeding against all the odds - they’ve now built something unique and very hard to replicate. As to why you might even want a ‘wafer-scale’ processor, I’ll cover some of that later. But suffice to say for now that the founders came to realise in late 2015 that the future computational requirements of AI, would look very different to those that the existing architectures were designed for (e.g. the graphics processing unit). And so, the Cerebras Wafer-Scale Engine (WSE) is what you get when one of the most determined, experienced and innovative systems design teams comes together at the right time, with the right backers, a bit of luck - and with a blank sheet of paper.

Cerebras’ current system, the CS-3, was released two years ago. The CS-3 was primarily intended for model training and provided an elegant solution for simple near-linear scaling out to large clusters. I thought this was the way things were going to go. But despite impressive demonstrated performance the CS-3 has remained very niche in training - training being where Nvidia’s moat is strongest. Within months of the CS-3 launch, in mid-2024 Cerebras quickly pivoted to inference in order to capitalise on the emerging shift to inference-time compute - or smarter ‘thinking’ models. In 2026, we’re all familiar with thinking models - they’ve become mainstream. And we’re all familiar with waiting while the model thinks. To calibrate your expectations here, the CS-3 typically runs models around 10-20 times faster than GPUs - in practical terms this means that chat responses come back more or less instantly without having to wait. Seeing is believing of course, so you can try out Cerebras Chat for free at this link. Or you can just watch Cerebras code an interactive pool game in 12 seconds:

Of all places, I came to be aware of Cerebras some years ago via GSK - yes, the stodgy old big pharma! (see my GSK post here). One of the lesser-known things about GSK is that they have perhaps the leading in-house AI and machine learning capabilities in the pharma industry. They were also Cerebras’ first commercial customer. Intriguingly, it was actually GSK who sought out and approached Cerebras - a complete unknown at the time and only recently out of stealth mode. The GSK AI/ML group had some big problems they were struggling to solve with GPUs. They looked into rightsizing the hardware to their problem, rather than the other way round - Cerebras had the hardware for the job. The two companies remain partners years later.

It took Cerebras around four years to design and launch their first system - a process defined by formidable challenges that they ultimately overcame. It took another six years to get traction with hyperscale customers. But they stuck at it, kept making their case and made a couple of important pivots. And so, after years of having to promote an offbeat roster of reference customers such as GSK, TotalEnergies, Mayo Clinic and G42, Cerebras can finally point to major megawatt-scale contracts with OpenAI and AWS. The foot is now firmly in the door. It’s no coincidence that the list of chip startups with major cloud deployments is particularly short. The barriers to entry are mighty. Turning up with something just a bit better simply will not do - you have to turn up with something an order of magnitude better, and that was the plan from the start. In many respects the WSE was ahead of its time but it’s now coming into its own.

In my case, it took a good couple of years for the potential of wafer-scale computing to dawn on me. By early 2024, I had become convinced enough to buy Cerebras shares in the private market. Any early excitement around Cerebras had now been squeezed out by Nvidia’s tightening grip on the AI narrative. Private market appetite for wafer-scale was minimal - which, of course, is exactly what made it interesting.

In the domain of frontier models and the underlying infrastructure that makes it all possible, we are, of course, talking about a space that is moving with extraordinary velocity. The number of people (or more accurately, teams of people) who are able to synthesise all of the associated technologies at the leading edge and hold a useful idea of where things might be heading, is incredibly limited - and almost all such people are active participants in shaping this future rather than investors passively trying to make sense of things from the sidelines. So we should be honest with ourselves: practically all of us are mere passengers. As I have said a few times before, investing for me is sometimes about striking the right balance between knowing too much and too little. It seems to be the case that some of Cerebras’ first venture backers felt the same way1 - they knew enough to grasp the significance of what the founders were proposing, but perhaps not enough to be scared off from trying. But most importantly, the reason Cerebras got funding at all was because the VCs were backing a team - in this case, a team who had worked together for years and had founded a successful systems venture previously (SeaMicro). They’ve done it twice now, and achieved something even more challenging the second time around. So in its most simplified form, the Cerebras thesis might be stated something like this: whoever had the requisite foresight, ingenuity, experience and persistence, against all the odds, to first create the Wafer-Scale Engine and then drive its adoption into the mainstream - it’s worth betting on those people and whatever they’re going to do next in the field of AI computing.

The AI market is moving extraordinarily quickly, and we are well-positioned to anticipate its twists and turns

- Cerebras Founders Letter

It’s worth stating at the outset that the range of outcomes for Cerebras is particularly wide and, in the context of an IPO and a possible AI infrastructure bubble, this will result in highly polarised views and speculation every which way. Cerebras produced around $500m in revenue for 2025, and per the S-1/A the IPO market cap looks like it could be in the region of $26 to $27bn. As the sceptics will rightly point out, these sorts of starting valuations rarely produce attractive prospective returns. Sceptics might also point to extreme customer concentration, not to mention the constant uncertainty of when the AI capex bubble will eventually pop. But on the other hand, if Cerebras executes to the full extent of the contracts disclosed in its prospectus, it could conceivably ramp to well over $10bn in revenue over the next five years - for context, this might just sneak Cerebras into the top 10 fastest ever growing businesses from product launch (roughly similar to Facebook’s 2004-2014 run). And, let’s say, if Cerebras can make it to just 5% of Nvidia’s current revenue run-rate by the early 2030s (that’s ~$15bn), you could still be looking at a $100bn market cap business even with an extreme de-rating from 50x revenue to 7x. And I want to be clear about this: Cerebras is not an “Nvidia killer” such as you see in some of the sensationalist headlines - the thesis is more about winning a small slice of a very large pie. We’ll look at scenarios later on. For my part, I think Cerebras has a genuine shot at getting big quite quickly. But it’s by no means a dead cert and there are important risks and considerations that we’ll look at.

I have deliberately decided to try to make this post as non-technical as possible - although it does assume some base level knowledge around certain terms and concepts. I’m going to link to some of Cerebras’ technical materials/presentations and I suggest diving in for those who want to go there. From following Cerebras over time, I have encountered a good deal of punditry, typically concerning why the company will remain limited to only a highly specialised niche or why its systems will surely prove economically unviable. Some of this commentary is simply a result of insufficient familiarity with the company, and some of it seems to assume that Cerebras will not continue to adapt or make advances in future system generations. But some observations are fair and worth making note of - which we’ll touch on later. I’ll try to avoid making too many technology-related claims of my own, especially where they might be contentious - to do otherwise, I risk getting in over my head. So I’ll try to stick to claims that seem fairly robust and highlight where I’m choosing to put my faith in others. It’s inevitable I’ll get some things wrong, so I shall endeavour to clear up mistakes later on.

But if there’s only one thing you need to understand about the Cerebras thesis it’s this: speed. It’s generally uncontentious to say that the principal bottleneck in inference output speed is having to constantly move massive amounts of data from memory to compute - move the data too slowly and the processor sits idling while waiting to be fed. Cerebras’ answer is to pack an entire wafer with the very fastest type of memory (SRAM) interspersed among the compute - data travels only tiny distances and through incredibly fat pipes. If you have to move data between individual chips (like everyone else does), the distances get longer and the pipes much thinner. This is why there’s nothing else remotely as fast as the CS-3 at inference:

Moreover, it has become clear that much higher output speed is able to command a price premium. In a similar vein, Anthropic have shown they can charge 6x more for their ‘fast mode’ at 2.5x the speed.

Cerebras is the only player who can produce a wafer-scale system such is required for extreme memory bandwidth. Extreme memory bandwidth enables extreme speed. Extreme speed potentially enables a breakout from the commodity token space through premium pricing and new use cases. Competitors are, of course, lining up to enter this gigantic market. And it’s also true that barriers to entry in chip design may have come down in some respects - there’s far more venture money washing around in semiconductors compared to a decade ago and the hyperscalers have deep pockets and motivation. Thus far, there’s not much evidence of anyone else going wafer-scale and, moreover, Cerebras owns important IP in that domain, having pioneered it. And, as we shall learn, it’s by no means easy. I therefore believe that barriers to entry in wafer-scale remain intact. Most alternative approaches to wafer-scale appear to retain the ‘thin pipe’ problem, go after cost rather than speed or else sacrifice vital programmability (which has obvious limitations if the models keep changing). And perhaps most importantly it’s important to weigh up the following considerations: 1) the CS-3 wasn’t even designed principally with inference in mind and 2) no frontier lab has yet designed a model specifically to take advantage of wafer-scale benefits and to optimise around the constraints. Both of these things are about to change. It’s clear that Cerebras and its partners have big ambitions for major system improvements from here - more on that later.

But who knows? Maybe the wafer-scale critics will have the last laugh in the end. Maybe the economics won’t play out competitively or maybe Cerebras will struggle to keep up with Nvidia’s relentless pace. Maybe they’ll struggle to manufacture these systems in the necessary volumes. Or perhaps they (or their largest customer) will run out of cash. We’ll review some of the risks later on. But what excites me most is that I don’t think we’ve even seen Cerebras’ A game yet. Here’s Cerebras co-founder and Chief Architect, Michael James, speaking at the end of last year:

If you look at the next designs that we’ll come out with, they might make [the WSE-3] seem quaint.

This? Quaint?:

I’m curious to see what comes next.

The Meaning of Speed

I was reading about George Stephenson, the great steam train pioneer, earlier this year. In Great Britain the idea of travelling twice as fast as the fastest horse-drawn mail coach was once considered absurd, unfeasible and dangerous, not to mention unnecessary - a dead end. And the country had previously invested heavily in a comprehensive network of canals for the efficient horse-drawn conveyance of commercial goods - the canal network was itself a paradigm shift and a key accelerant of the industrial revolution in Britain. So when in the 1820s Stephenson proposed the world’s first intercity steam locomotive railway between Liverpool and Manchester, he faced a good deal of scepticism:

George Stephenson’s idea was at that time regarded as but the dream of a chimerical projector. It stood before the public friendless, struggling hard to gain a footing, scarcely daring to lift itself into notice for fear of ridicule. The civil engineers generally rejected the notion of a Locomotive Railway; and when no leading man of the day could be found to stand forward in support of the Killingworth mechanic, its chances of success must indeed have been pronounced but small. But, like all great truths, the time was surely to come when it was to prevail.2

I am, of course, just having a bit of fun here. But I do enjoy these stories where some maverick succeeds with an audacious groundbreaking idea against the odds - and the analogy of moving from the canal age (i.e. GPU) to the steam locomotive age (i.e. Wafer-Scale Engine) is a tempting one. Stephenson got his railway in the end and it changed the meaning of speed for the whole world:

Notions which we have received from our ancestors, and verified by our own experience, are overthrown in a day, and a new standard erected, by which to form our ideas for the future. Speed – despatch – distance – are still relative terms, but their meaning has been totally changed within a few months: what was quick is now slow; what was distant is now near; and this change in our ideas will not be limited to the environs of Liverpool and Manchester – it will pervade society at large.3

The rest is history, as they say. Back in Stephenson’s day, there were those who pointed to the efficiency of the existing canal system to demonstrate why railways were unnecessary. They weren’t wrong. But it showed a lack of imagination. Even today’s canal baron, Jensen Huang, is finally hedging his bets and ponying up for some steam engines of his own - as evidenced by Nvidia’s recent acquisition of Groq (the closest Cerebras has to a fast inference peer). The first sections of intercity track are now being laid.

Staying on the topic of speed, Cerebras co-founder and CEO Andrew Feldman likes to draw the more modern analogy between dial-up/broadband internet and slow/fast inference. I’m old enough to have had my formative internet experiences with dial-up modems. And Feldman has a point - waiting for a model to respond in 2026 conjures up old memories of loading webpages 30 years ago. For any patient youngsters or nostalgic oldies, you can see what this was like here:

What did it mean when broadband arrived and we no longer had to wait? The internet changed: websites could become far more data rich, streaming, cloud computing, you name it - entirely new business models emerged and their effects continue to ripple through our world today. And as the world moves forward, expectations change - Feldman and others like to point to research carried out by Google and Amazon some years ago showing how mere milliseconds of latency can cause a reduction in traffic or even revenue.4

So what of fast inference? Here’s how Cerebras co-founder Sean Lie put it when they announced an experimental ultra-fast coding tool with OpenAI a couple of months ago (Codex-Spark):

What excites us most about GPT-5.3-Codex-Spark is partnering with OpenAI and the developer community to discover what fast inference makes possible—new interaction patterns, new use cases, and a fundamentally different model experience. This preview is just the beginning.

-Sean Lie, CTO and Co-Founder of Cerebras

Readers may have their own views on this, but in the prospectus (p121) Cerebras cites coding as a strong example, as well as real-time insights in deep research and voice applications. We’ll just have to see what emerges in the months and years ahead. But the trend is clear: more thinking time » more inference » more waiting (unless we do something about it):

Here ends the complementary section of this post - thank you for reading! In the remainder, we will cover the following topics:

• The Road to Wafer-Scale

• Brief Thoughts & Speculations on Tokenomics

• The Roadmap: what we know + some possibilities

• Business Model & Sizing The Opportunity

• Risks & Other Considerations

• Final Thoughts: Moat

Subscribe now

The Road to Wafer-Scale

The Cerebras origin story gets underway around 20 years ago.

SeaMicro

At the time, Gary Lauterbach was a highly experienced architect at Sun Microsystems. He had designed Sun’s latest generations of powerful server chips. Andrew Feldman was a former network equipment executive turned entrepreneur-in-residence at a VC firm. Both had, via different paths, arrived at the same conclusion: a new kind of efficient server hardware was needed to address the massively escalating challenge of datacenter power consumption. Feldman decided to start a company, but he needed a technical founder to join him. He had the right network and VC connections, while Lauterbach had the engineering expertise. A former colleague of Feldman (who

Read more

Friends of NFTBC

Yesterday was a busy day with four of our stocks of interest reporting. I intend to visit them all, but not necessarily in standalone posts - so keep an eye out in The NFTBC Digest. There’s also a new long-form post in the works for the coming days, so keep an eye out for that one too.

I began my BRKR deep dive last year with this 2010 quote from Nobel Prize winner Richard Ernst:

And what is the moral of the success story of the Bruker Corporation? Bruker has been active in a unique manner in a high-tech leading field over an incredibly long period of 50 years. Few other companies can match Bruker. Whatever Bruker has achieved has demanded novel technologies that in each case would have been impossible a few years earlier. For academia, such companies are ideal partners: they are highly ambitious, stimulating, and grateful for creative contributions.

This form of exchange can indeed be highly motivating for inventive entrepreneurs and for innovative researchers with endurance. Ultimately, it is courageous and creative efforts that are decisive. The example of Varian also shows that initial success can lead to complacency and the underestimation of motivated competitors. One further lesson that can be drawn concerns company leadership. One should guard oneself from companies that are managed by technically illiterate lawyers and MBAs. Technical expertise coupled to close contact with the market remains essential for avoiding flawed decisions. Bruker has in the past evaded such perils through wise personnel policies and long-term continuity. The company has produced beneficial innovations and added value for 50 years. Also investors might derive lessons: it is still the rock-solid technical expertise of a CEO together with their market awareness that offer the best guarantees against false decisions. With innovative companies, like the Bruker Corporation, the frequently forecast demise of Europe is not imminent. Even if the Bruker Corporation is of late governed by American law, the spirit of this remarkable company is still firmly European.

As I said at the time, this continues to more or less sum up the BRKR thesis, a decade and a half after it was written. Ernst was talking about Nuclear Magnetic Resonance, where BRKR now has 100% market share in high field systems - when you can do things that others cannot that’s usually a valuable proposition. A similar dynamic reaches out into numerous other fields with compelling longer-term tailwinds, whether that’s in the life science domain, semiconductor metrology, nuclear fusion, the list goes on - and it’s an eclectic list that generally goes quite under-appreciated. Over the course of the last 12 months, since I posted the deep dive, the company has continued to push in this direction by launching breakthrough instruments with the potential to open up new markets - these include timsOmni (functional proteomics), timsMetabo (4D metabolomics), PaintScape (3D genomics) and others. While it’s too early to know what the eventual financial impact of incremental innovations will be, the contribution to the order book is already showing up and the impact on intrinsic value likely to be material over time. In the context of last year’s dilution saga, things ain’t so bad then. Moreover, we can now be quite confident that the next major multi-year earnings growth story is finally underway, starting Q2 2026: back to the future.

In the remainder of this post we’ll take a look at some of takeaways from yesterday’s earnings materials. One last thing: for a limited period I have taken the paywall off the original deep dive for anyone who’d like to see it. I just ask that you kindly hit the like and share buttons.

[NFTBC does not give advice - please do your own research. I currently own Bruker shares].

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Friends of NFTBC

For as long as I have been looking at Regeneron, investors have been fixated on Eylea - specifically on its loss of exclusivity and the threat of competitors. So it’s notable to observe that yesterday REGN reported record Q1 revenue - up 14% on the prior record set in Q1 ‘23 (actually +23% if you rightly exclude one-off COVID revenue). During the same time, Eylea sales have declined from a hair-raising 47% of revenue to 13% today.

Yet, in a clean break with all convention, REGN did not elect to buy their way out of trouble - instead, all of the offsets (Dupixent, Eylea HD, Libtayo etc) came out of their own labs. But with the Eylea LOE eventually turning out to be a relative non-event, investors have now landed on a new fixation: the 2031 Dupixent LOE. In 2031 Dupi starts to lose some of its key patents and the bears are demanding their pound of flesh five years ahead of time. No matter that Dupi is still growing 20-30% with plenty of runway to go, nor that its IP might stretch a bit longer in the end. If you look at REGN through a conventional biopharma lens, there’s no escaping the conventional diagnosis: major LOEs can only be addressed through M&A or major business development. Trying to actually innovate your way out of such a hole is recklessness of the highest order and should not go unpunished:

This exchange from yesterday captures the dynamic very well:

Analyst

Len and George, my question is for you. So Regeneron spends aggressively on R&D, but the investment community lacks confidence that the company’s candidates will move the needle commercially, in particular versus established competitors. So could you please highlight the pipeline candidates in late-stage development that we’ll have cards turning over in the near term or relative near term, i.e., in the next, I don’t know, 18 months or so that you have the greatest confidence in that can generate multibillion-dollar peak sales that investors will be able to see more clearly in the next 18 months or so.

George Yancopoulos

Can I just say that past performance should be the strongest indicator of future performance. There’s only one company in recent history that’s had its own labs produce two $10 billion-plus blockbusters. And let me remind you that I think you, and probably a lot of other investors, never saw those coming or ignored what we were saying about them. So investor confidence, I think, should in large part be reflecting historical performance and the recognition that where blockbusters come from sometimes for the investor community can’t be directly anticipated. And the best way of producing very important big drugs is by having very exciting molecules across all stages of development that have enormous opportunity.

And if you just look at our oncology programs, whether it’s fianlimab-Libtayo, whether you look at Lynozyfic, whether you also look at odronextamab in follicular lymphoma. These are all potential blockbusters. The C5 franchise is a pipeline in a product, multiple blockbuster opportunities there. Our Factor XI customized approaches are looking more and more exciting, especially based on competitor data using, we think, inferior and less convenient approaches. And we just covered the obesity opportunity, which arguably could become the preferred obesity approach that not only addresses obesity, but more aggressively addresses cardiovascular morbidity.

So I don’t know, it’s hard to think of a more exciting pipeline in the entire industry.

Which is it then: a) the indefensible squandering of company money on R&D or b) one of the most exciting pipelines in the industry? These two poles could hardly be further apart! And perhaps we won’t get a definitive answer until hindsight offers us a helping hand.

With the remainder of this post, we’ll have a look at some of the notable points that came out of REGN’s earnings yesterday as well as the recent C5 Rountable event.

[NFTBC does not give advice - please do your own research. I currently own Regeneron shares].

Subscribe now

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

Friends of NFTBC

When I started looking at recently listed Ottobock (Ticker: OBCK) I was immediately intrigued, so I decided to write about it. Additionally, my impression is that global investors might not yet be especially aware of this German ‘hidden champion’. OBCK, with a market cap of around EUR 3.5bn, is the market leader in human bionics - most notably prosthetics, but also in tangential fields. The business itself has a long and colourful (if not especially public) history and its family owners have, I believe, built some quite compelling moats around it.

For various reasons I’m slightly behind schedule on getting this post out the door. Part of the reason for that is that I have been looking at OBCK fresh and the more I learnt, the more questions I had. There are numerous moving parts to understanding this business1 and various dynamics at play - I wanted to be sufficiently comfortable about them before writing to you about it all. Having done this now, one thing seems clear to me: OBCK is a very hard business to compete with and may even be a candidate for the most moaty business I’ve written about so far at NFTBC. In some instances, OBCK has exceptional pricing power. But I’m also pleased to note that in such instances its products bring tremendous value to their users - so long as reimbursement and access can be established. That’s all part of the story too. It’s also a largely recurring revenue business model with plenty of avenues towards durable growth - we’ll go into all of this.

My aim with this post is to try to bring OBCK newcomers up to speed in 40 minutes or so and to point you in the right directions should you decide to follow up with your own research. As a business with many thousands of SKUs, there is a lot that I won’t cover here, but I’m happy to continue the conversation elsewhere. Here’s what the basic layout will look like:

1. Introduction (free to read)

2. Background (partly free to read)

3. Basic thesis: Strategy, Moats & Optionality

4. Governance Structure & Management

5. The Business Today: Overview, Financials & Competitors

6. B2B: Products & Markets

7. B2C: Ottobock Patient Care

8. Risks / FAQ

9. Conclusions: 2026 & Beyond, Initial Thoughts on Valuation

While I do not yet currently own OBCK, I shall discuss later some of the conditions that could lead me to take action - potentially in the near term. As always, I will inform paying subscribers when and why I decide to acquire OBCK shares - should that time come. If you’re new to NFTBC, please check out this page for more information about the NFTBC approach and what to expect. I’ll no doubt make some mistakes in this post, which I shall endeavour to correct over time. With that, let’s get going.

[NFTBC does not give advice - please do your own research. I do not currently own Ottobock shares].

Introduction

For the last 100 years, OBCK has led the field in developing and commercialising next generation prosthetics. Most notably, in recent times, important developments include the microprocessor knee (MPK) - a breakthrough innovation that enables above-the-knee amputees to move with a more natural gait, safely tackle everyday tasks and even engage in much more adventurous pursuits. Today OBCK is the largest and most important prosthetics OEM by a wide margin - about 40% market share. Additionally, OBCK has been investing heavily into tangential fields in order to open up new and potentially even larger markets - more on that later.

While the company has only been listed since October 2025, we have sufficient information to know that OBCK has been a relatively consistent 7-8% top line grower for a long time. Revenue was under EUR 100m (equivalent) in 1990 - implying around 17x in 35 years or greater than 8% CAGR.

The numbers in the last several years also mask a more promising underlying story of 3-year ‘core’ organic revenue growth of 11%. With the company now close to having completed a series of non-core divestments, these revenue figures are now converging. Another compelling dynamic at play is the recent trend of margin expansion, which management expect to continue - we’ll get into some of the reasons behind that later. But first, some background…

Subscribe now

Share

Background: History, the Näders, EQT & IPO

A Little History

For the particularly keen, there is a three volume book on the history of Ottobock. I’m afraid I haven’t read it - but I don’t think we need to go into deep history here.

The founder, Otto Bock, started the business in Germany in the wake of World War I. At the time, prosthetics were one-off crafted items, and as far as processes go, this one was highly inefficient given the elevated need at that time. Bock industrialised the process turning it into the serialised mass production of components. In time, the business came to be managed by Bock’s son-in-law Max Näder. Under Näder the business continued to launch a series of important innovations including the breakthrough Jupa knee joint with braking mechanism, myoelectrics that help users grab objects, plastic prosthetics as well as the modular system that underpins modern prosthetics today. In 1990, at the age of 28, Hans Georg Näder took charge as CEO - the grandson of Bock and the son of Max Näder.

Hans Georg Näder (HGN)

Given the various Näders involved in this story, I shall refer to Hans Georg as HGN from hereon to avoid confusion.

HGN has taken a fair bit of heat in the German business media over the years, for example for apparently taking cash out the business, raising debt and for buying shipyards, breweries and such others things that eccentric billionaires are wont to do. There’s even a website dedicated2 to posting largely negative content about the man - if you decide to research OBCK you’ll probably find it quite quickly. I’m inclined to leave such tittle tattle to one side. For one thing, up until the point when EQT acquired a 20% stake in OBCK, HGN basically owned the business outright and was free to run his affairs in any way he saw fit. One thing worth noting (and that OBCK commented on some years ago) is the importance of drawing a distinction between Ottobock SE & Co. KGaA (i.e. OBCK) and the family holding entity Otto Bock Holding GmbH & Co. KG - it seems journalists have not always done this. In the end, the latter changed its name to Näder Holding GmbH & Co. KG perhaps in part to help address the confusion. Needless to say, the affairs of the holding company and the other assets it owns, are not directly relevant to OBCK shareholders except to the extent I shall identify shortly.

With that out of the way, let us now consider some of the great strategic decisions made by HGN - if you’ll pardon me for going off on a brief tangent. The biggest risk with successful family owned businesses is the possibility of dilution to management quality over time. So it’s notable that the grandson of the founder has been able to act with great strategic foresight. OBCK’s most important product of the last 25 years has been the C-Leg microprocessor knee. The first ever fully adaptive MPK, it’s been fitted more than 100,000 times and continues to be the market leader (after numerous iterations and branching products). Since the turn of the century, OBCK has compounded MPK units at 20% per year. While they don’t break it out, I think we can infer that the C-Leg is a substantial percentage of total revenue. The MPK was originally pioneered by maverick Canadian engineer Kelly James in the early 1990s. Per Ottobock lore, HGN waltzed into a prosthetics trade fair and penned a deal with James there and then, having recognised the significance of the technology.

The OBCK PR department seem to make a big deal out of this in various places, which made me kind of suspicious and drove me to try and verify it. And… it’s true: Kelly James independently retold the same in an interview with his alma mater University of Alberta 15 years ago. The context is telling and worth expanding on:

[Kelly James] travelled around the world to conferences trying to sell his leg to a manufacturer. At trade shows, he walked from booth to booth with his prototype slung over his shoulder. But most manufacturers were reluctant. What James was selling was, in retrospect, ahead of its time. “They all said, ‘I don’t know, it’s pretty complicated.’ Remember, in 1992 people didn’t know anything about computers.” James persevered, even when people in the prosthetics field told him the idea was crazy. But in 1992, German prosthetics manufacturer Otto Bock approached James at a trade show. In a closed-room meeting with the company’s owner and top engineers, they made James a deal for the leg, on the spot. “What really surprised me about the meeting was that they were so respectful of the leg. After I had talked to all the American companies and they had just slammed the door in my face, I spoke to the German company and they were so, ‘Wow, this is really amazing, and how did you do it?’”

But there were obstacles to overcome along the way. Otto Bock nearly decided not to manufacture and sell the leg after one doctor’s experience. The doctor, an amputee, was impressed with the leg after wearing it around for a couple of hours. But when he took it off to put his old prosthesis back on, he fell down the stairs. The C-Leg was too good in the sense that it was dangerous for someone to readjust to a less-sophisticated prosthesis. Otto Bock executives were upset, says James. “They were saying, ‘This is a no-brainer, we’re not going to build this knee … we can’t put amputees at risk.’”

James managed to convince them otherwise. “Let’s just change the number of phases the knee goes through during a step,” he said. The idea of easily tweaking a few lines of code for the knee’s microchip was an eye-opener for Otto Bock. Over time, the phases of the knee evolved, eventually eliminating any phase where the knee is locked and anything that made the knee too complicated for users to figure out. The knee’s simplicity is one of its main selling points. “You don’t read a manual, you don’t understand it, you don’t think about it—you just walk,” says James.

It’s kind of interesting to me because HGN was still relatively young and had only been in charge a couple of years. Would a professional salaried CEO have made the same decision? Kelly seems to suggest not. In the event, OBCK spent 5 years in development and launched the C-Leg in 1997. And it’s taken decades to build the case for reimbursement and grow access - a journey that continues. It’s tough to imagine these sorts of long-range investments being made by a non-family controlled business.

Read more

Welcome to The NFTBC Digest, an email received by full NFTBC subscribers only.

From week-to-week, there is all manner of news, links, marginalia and other curiosities concerning NFTBC stocks that I might not otherwise share with you. So this is what The NFTBC Digest is all about. My thinking here is that you can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable that tickles your fancy and that you haven’t already seen. There won’t be a set frequency - I don’t want to publish for the sake of publishing. Sometimes The Digest might be weekly when there’s plenty happening, but there could be fewer Digests during earnings season or if news is slower.

Read more

First things first, I’m working on a couple of projects/posts in parallel at the moment. The first will be an under-the-radar name in the healthcare space - I’m hoping to get this one out soon-ish, so keep an eye out for that.

Second, last week I recorded a podcast with Bill Brewster on the the Business Brew - I enjoyed the conversation. We talked about NFTBC and spent a bit of time on REGN, BMW and one or two other names. You can listen here, if you would like to:

• Apple

• Spotify

Be sure to subscribe to the Business Brew if you haven’t already. Bill’s a thoughtful host and asks great questions - I can only hope that I answered some of them intelligently.

I think my weakest answer was to Bill’s question about what the process looks like at NFTBC. I sort of waved the question away by saying there’s no formula or rulebook that I follow - while this is effectively true, it’s also a bit of a cop out. As an independent analyst-writer I think I’ve elaborated pretty well by this stage on the what and the why (e.g. here, here and here). But not as much on the how - the process. I suppose the process, insofar as there is one, is just something I do without ever having formally codified it or putting much thought into explaining it. What follows, is a few brief thoughts on how I might have answered the question better.

It’s true that I don’t work formulaically. I don’t like to have too many rules, preferring the flexibility of more of a principles based approach (instructed by the what and why).

Idea generation

Most (if not all) of the ideas for businesses to write about are sourced independently. In other words, in most cases, I probably won’t have seen or read a write-up by someone else before I dive in.

In terms of leads, they really can come from anywhere. Quite often, leads arise unexpectedly when I’m looking at or reading about something tangential or maybe something altogether unrelated. Other times, when I’m at a bit of a loose end, I like to sift through long lists of stocks just to see if anything jumps out - although I don’t run screens. I quite like to work quickly and visually at this stage - bringing up lots of different stocks, displaying financials in charts, scanning through information about management etc. I pass over practically everything for one reason or another.

Research phase

Many ideas look interesting to me intuitively and quickly. The next step is then about figuring out if the intuitions were any good and unearthing important considerations that are worth understanding. My process then essentially revolves around figuring out which questions to ask and then trying to answer them in a satisfactory way. And the deeper I go, the more questions arise. When I start out I don’t even know what most of the questions are going to be. Some of these questions are undoubtedly stupid ones. Some are perhaps more intelligent. I continue in this iterative process until I’m comfortable that I’ve identified the most important aspects of the business for my purposes alongside the relevant risks and uncertainties (i.e. questions that cannot be answered, for the time being).

Similar to the output, the process is far more qualitative than quantitative. But it’s not to say the numbers are not important, they are. Actually I insist on a somewhat laborious process of extracting figures myself manually from years’ worth of financial statements. I’ve never really trusted the fidelity of third party data. But more importantly, I tend to find that the numbers tell a more instructive story when I pull them out myself. The very extraction process becomes a learning exercise and I’m more likely to notice things and ask questions along the way - rather than skipping straight to a pre-populated Excel sheet.

It’s always my ambition to rule out dead ends as quickly as possible, before I invest too much time. But, alas, very occasionally I might get quite far before I assess the opportunity to be a no-go. Actually I’ve already had one of these this year, although there’s a possibility I could revisit that one in future.

Writing phase

The writing phase usually begins well before the research phase has finished - one feeds back into the other. Putting thoughts into writing ends up generating more research questions.

As to the writing process itself, I don’t have much of a template. Each post is generally bespoke. The approach and format I end up pursuing, emerge in a bottom-up fashion as the research progresses. Every business is unique and context-specific and flexibility seems to be the most accommodating way to deal with that.

AI

I’ve touched on AI before. It will always be the NFTBC promise not to use LLMs in writing, except in a very limited fashion and always identified as such. Neither do I routinely use AI for automating data-type tasks - other than for occasional genuine chores, where there won’t be any ancillary benefits from carrying out the heavy lift myself. I do however now use AI in the research part of the process, in an iterative/dialectic way. I find it very helpful in helping to surface new questions. Insofar as the AI purports to give me the answers, I approach these with scepticism - always seeking to challenge and verify, including from original sources.

Walking

I’ve always walked a lot, out in the fresh air. Insofar as I have ever have any original insights (up for debate) they’re most likely to occur when I’m out walking rather than sitting at a desk.

My hope is that the net result of all of this is something a bit different, interesting and potentially interesting.

Thanks for Reading.

Subscribe now

Share