Friends of NFTBC

With Cerebras’ stock now swooning like one of the great leading ladies of the Golden Age, it feels like the right time to temporarily drop the paywall on my original long-form post from May - do help yourselves if you would like:

We are, of course, in the midst of one of our occasional collective AI panics, and this is driving most of the activity - CBRS is by no means alone in its fall. The plan for this post is to share a few very CBRS-specific musings. I offer you no useful opinion on the remaining lifespan of the AI capex boom and such things. But what I do feel quite certain about is that, bubble bursting or otherwise, various important forces are at play and these are unlikely to be shunted off course in any significant way - perhaps even in the face of a capex recession. The genie is out of the lamp, and he’s not going back in.

First things first: to be able to see the CBRS bull case requires (in my view): A) an acceptance of the unique benefits of wafer-scale both today and perhaps even more so in the future, B) an acceptance that fast AI is going to be transformational, C) the realisation that CBRS’s primary business model is much closer to the recurring revenue model of a fully integrated AI cloud infrastructure provider like AWS or Google rather than a traditional chip/server business with customers placing discrete orders and D) an understanding that CBRS has ~$8bn of liquidity, giving it considerable flexibility to continue self-funding its expansion.

I won’t fully revisit A) here, having addressed it in various other posts but the central point comes down to the ‘thin pipes’ vs ‘fat pipes’ framework I talked about previously. The benefits to speed from minimising off-chip data movement down thin pipes are obvious and very significant. There are plenty of innovative approaches being pursued by chipmakers to widen the pipes between chips and chiplets, or bring memory and compute closer together. CBRS’s distinctive approach is to integrate a vast amount of compute and SRAM on a single wafer-scale processor, reducing the need to move data between separate chips in the first place. Moreover, wafer-scale could also benefit from advances in vertical integration and interconnect, with its large surface area offering considerable scope for many more additional connections. Smaller chips can exploit these technologies too, but CBRS uniquely gets to combine them with the advantages of its existing wafer-scale architecture. There are certainly some practical challenges to making wafer-sized chips economically and at massive scale, and consequently much of the CBRS case rests on whether or not you believe management can ‘execute’. For this reason, CBRS attracts plenty of scepticism. But, in all fairness, CBRS do have a solid track record in proving the sceptics wrong. I have little doubt that investors will continue to debate these issues for some time - so let’s just move on for now…

For the purposes of this post, I’m more interested in B), C) and D) and what they might tell us about how CBRS could perform if the AI capex cycle turns soon. One thought that I have been entertaining is the possibility that growth in the market for fast inference might not slow down by much, even if we were to enter some AI equivalent of the GFC. Today, fast inference is truly in its nascency - it’s barely started deployment yet, and not at industrial scale in any case. Only a minuscule percentage of users have even tried it yet. And only in Q4 2026 is CBRS actually starting to bring online material amounts of fast inference capacity, and this ramp will continue from hereon. And it’s similar for Nvidia’s new Groq LPX rack, which has only recently started production. The market for fast inference therefore is only really entering industrial scale production right about now. And it’s my contention that fast inference will transition very rapidly from ‘expensive gimmick’ » ‘nice to have’ » ‘essential’.

In my mind, I keep returning to the example of the iPhone launch in 2007. At the time, many people saw it as an expensive gimmick (I regret to admit to being one of them). And then the GFC arrived, bringing with it a sharp recession in consumer discretionary spending. In the fourth quarter of 2008, right at the moment when the global financial system was falling apart, Apple shipped more than 4m iPhones, up 90% year-on-year. And the growth rate accelerated to over 120% in the next quarter during the peak of the crisis. Incredibly, sales didn’t show much sign of a reversal until 2016, by which time Apple were shifting more than 200m units a year. The iPhone had now completed its transition: ‘expensive gimmick’ » ‘nice to have’ » ‘essential’. Smartphones really did change the world. One should always, of course, be a little careful with analogies. Apple helped its own cause with better products, much wider distribution and a lower upfront purchase price. And these figures cannot tell us how much faster the iPhone might have grown without the crisis. But they do demonstrate that a transformational technology, when market-ready and starting from a low base, can remain in vertical takeoff even under conditions of financial armageddon. I think CBRS, with its new CS-4 system designed for hyperscale, has a credible opportunity to follow a similar trajectory - provided the speed comes with compelling economics.

The point here is that very fast inference brings a fundamentally different experience and creates new possibilities to do things that couldn’t otherwise be done or even imagined. CBRS’s co-founder CEO likes to use the example of what broadband did for the internet, to suggest directionally what fast inference could do for AI. Whether it’s to do with realtime human-machine interactivity or even faster machine-to-machine or agent-to-agent communication - surely most of the best applications are yet to be explored.

OpenAI, for example, now uses CBRS internally for incident response and in shortening feedback loops between researchers, their agents and experiments. And, in fact, so highly does OpenAI seem to value these scarce capabilities they have, thus far, kept the greatest part of their CBRS capacity for themselves rather than moving to resell it all at high prices to customers - in other words, at OpenAI CBRS has potentially already become critical infrastructure in a matter of months. Here’s another very interesting use case emerging with a sense of urgency, from the recent announcement of a partnership with cybersecurity business Crowdstrike:

Cybersecurity is fundamentally changed by AI, as attacks are faster, more targeted, and harder to detect. Defending against AI-accelerated threats requires security that reasons and acts at machine speed. Adversaries move across domains in seconds, and the window to detect and respond has collapsed. Security teams must respond with larger, smarter models that run fast enough for real-time protection, stopping more threats sooner, including those that legacy systems would have missed

No doubt you’ll have already heard about the recent Hugging Face cybersecurity incident. During evaluations conducted with reduced safeguards, OpenAI agents circumvented controls intended to isolate them from the internet, coordinated through unauthorised channels, and compromised parts of both OpenAI’s research infrastructure and Hugging Face’s systems. It feels inevitable that incidents involving increasingly capable AI, including deliberate attacks directed by bad actors, will become more common. Open-weight models that can be modified to remove safeguards add another dimension to that concern. The genie is out, and ready to do his master’s bidding. The idea that these sorts of serious threats will simply disappear of their own accord or due to a capex pullback is magical thinking. In a recent podcast, one of OpenAI’s co-founders said that cyber spend “should actually vastly outstrip software spend” in future. His argument partly rests on software becoming much cheaper, with security and reliability accounting for more of the value. But even without accepting the full prediction, the opportunity seems sizeable and likely to feature prominently in budgeting priorities - including when spending elsewhere is being cut. The link to CBRS is strongest where security decisions require capable models to reason and respond immediately, although not every cybersecurity workload will need the fastest inference. The CrowdStrike partnership provides a concrete example of where it could matter - a lot. Once again: ‘expensive gimmick’ » ‘nice to have’ » ‘essential’.

But let’s hit pause on the point about cutbacks and reflect some more there. What if OpenAI runs low on cash and has to slash spending, what does that mean for the $20bn CBRS contract? This is obviously one of the possibilities investors seem to care about the most. Well, I would probably assume CBRS is one of the last things to get cut. I’ve already mentioned the critical infrastructure point, but also the possibility of selling fast inference as a higher margin scarcity product. But taking it to the next level, what if OpenAI fails entirely and defaults? Or what if the industry-wide ‘overbuild’ results in a gigantic inference capacity glut? I think it’s entirely plausible, and indeed likely, that a glut in slow tokens could coexist with a shortage in the fastest tokens. For 2027, CBRS is targeting up to 5,000 output tokens per second per user on large frontier models with CS-5, alongside substantial improvements in aggregate throughput. These are roadmap targets, rather than demonstrated production results, and competitors are moving quickly too. But achieving that combination of model capability, speed and throughput would represent a significant advance. The relevant test will be whether anyone else can match it on comparable models, contexts and economics - and there are no signs of this yet, in my view. The market for sub-200 tok/s inference serves different needs from the market for 1,000+ tok/s inference, even though some customers will naturally trade speed against price and other capabilities. A glut of conventional inference capacity would therefore not necessarily imply a glut of the fastest inference. Given how early we are in deploying these capabilities widely, I think demand could continue to grow rapidly even through an industry downturn. That doesn’t make fast inference immune to falling prices or eventual oversupply, but it does make me extremely wary of treating all inference capacity as interchangeable.

For this reason, I think CBRS would have a very credible route to finding alternative customers quite quickly if OpenAI suddenly failed. The infrastructure’s usefulness would not disappear with its largest customer. But if replacing a major contracted revenue stream, the models served, prices achieved and resulting cash flows could all be different - and not necessarily for the worse, mind you. Reassuringly, Cerebras Cloud pre-exists the OpenAI relationship and, in fact, they already maintain an extensive menu of open source models (both large and small) available today for enterprise customers with reserved capacity.

And then, returning to ponder CBRS’s business model again briefly. For obvious reasons, everyone thinks of CBRS as a chip business, and chip businesses being cyclical should be appropriately discounted as such. But CBRS is rapidly evolving into a cloud infrastructure and leasing business: cloud and other services already accounted for approximately 70% of reported revenue in Q2 and remain on a upwards trajectory. This requires more capital upfront, but multi-year capacity commitments can provide much greater revenue visibility than a succession of discrete hardware orders. OpenAI’s published agreement, for example, makes payment obligations non-cancellable, subject to the contract’s terms. Of course, that visibility is only as good as CBRS’s ability to deliver and its customers’ ability to pay. But there’s another potential advantage: as an integrated AI infrastructure provider designing its own accelerators, CBRS can avoid paying an external supplier’s margin on the core wafer-scale hardware. As the business scales and the new industrialised Nexus architecture ramps production, that could translate into attractive costs. It doesn’t automatically establish the lowest cost in the industry - CBRS still has to fund its own development, manufacturing and operations - but the opportunity is clear. Admittedly, CBRS also plans to deploy AMD’s Helios systems as part of its disaggregated inference solution. The eventual proportion of spending and power accounted for by that hardware remains to be seen, particularly as it will handle prompt processing and long contexts. But this is also a way to improve the economics of the combined system by assigning different jobs to the hardware best suited to them.

And then there is the flexibility of CBRS’s distributed deployment model, which can use smaller, pre-existing or co-located facilities rather than requiring all its capacity to sit in one enormous tightly connected cluster. GPU inference can also be distributed, so this isn’t an exclusive advantage, but CBRS’s deployment flexibility could still help it secure power and bring capacity online quickly. Ultimately, the meaningful comparison is the total cost of delivering useful inference at the required speed and model capability, rather than simply the cost per megawatt. While we’re yet to get a firm idea of what CBRS’s operating cash flows and capex outlays could look like in the years ahead, there are obvious and major advantages to avoiding the ‘Nvidia tax’ on the one hand and having to build superclusters on the other. I’ll leave you to carry out your own projections on CBRS’s future capital requirements, but I’m optimistic that these financial resources, supplemented by cash generated as contracted capacity comes online, could stretch a long way. That remains a judgment about future spending and cash generation, rather than a claim that the entire growth plan is already fully funded. CBRS was fortunate to IPO when it did, with the realised upsizing(s) and pricing. The point here is that CBRS could have considerably more financial resilience than is apparently assumed.

Naturally, in the short term I very much doubt any of this really matters or will influence the stock positively. None of this is intended to persuade or offer comfort that CBRS has near-term downside protection. If we are indeed heading into a capex bust I would fully expect CBRS to get carried out with everything else - and perhaps even more so given its lingering perception as a highly speculative bubble stock. All stocks are participating in a Beauty Contest after all and it matters not what you or I think about CBRS in the long run, but what other investors think that other investors think (to the nth order) and generally in the short run. But what I will say is that the lower CBRS goes, the more likely it is to attract acquisition interest from strategic suitors. I can think of at least several potential buyers for whom CBRS ownership would be highly strategic. And I can think of several industry players who most likely would not want certain competitors to own CBRS either. In other words, we have the possibility of the catalyst for a bidding war, and I would just note that Nvidia acquired bits and pieces of Groq last year for around $20bn. That doesn’t establish what anyone would pay for CBRS, but it does suggest the strategic value of these capabilities can be substantial.

Most importantly for myself as a long-term holder (and perhaps for some of you too), it’s helpful to have some degree of confidence that CBRS is capable of surviving whatever surprises might come down the road, and better yet - flourishing.

Thanks for reading.

Subscribe now

Share

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

This week Schneider Electric announced an agreement to acquire PTC in an all-cash deal. The transaction has been unanimously approved by PTC’s board of directors and represents a 42.3% premium to the prior closing share price. While the deal still requires various approvals (including that of PTC’s shareholders) and is expected to close by the third quarter of 2027, it looks very likely to go ahead. For this reason, and barring the unforeseen, this post will be the last of my musings on PTC.

Concurrently, I’m taking the paywall off all of my standalone PTC posts, including both parts of the deep dive - please help yourself (link here). There’s a good deal more PTC content within The NFTBC Digest, but given the impracticalities of separating it, I’m unable to open it up in the same way.

The purpose of this post is twofold:

1. To wrap up our shared PTC journey and reflect on the experience of the last 12 months - the bitter and the sweet.

2. To use the case study to highlight to you what I’m trying to do with NFTBC.

The Sweet

When I started writing about PTC, the company was in the midst of a major go-to-market transformation to reaccelerate organic ARR growth back into the double digits and to sustain it there. A year later, we can say that progress is encouraging: ARR growth is now starting to accelerate and management have outlined a credible near-term path back to double-digits.

There were three other specific elements of the thesis that I highlighted and explored as well: 1) untapped pricing power, 2) AI as a tailwind (in particular the foundational role played by PLM software in industrial enterprise AI adoption) and 3) Onshape - PTC’s uniquely disruptive cloud-native CAD system. As the year progressed, PTC demonstrated a shift towards longer customer contracts, as well as a new pricing model that is only beginning to flow through the business. This incorporates annual contractual pricing uplift of around 3–4%, compared with the 1–2% historically achieved through renewal negotiations - an extra couple of percentage points as contracts move onto the new model (pricing power ✅). We also saw PTC increasingly winning in competitive PLM displacements. Historically, displacing PLM has been notoriously difficult, but PTC has made a compelling case for a modern PLM implementation as table stakes for manufacturing businesses that want to incorporate AI into their product lifecycle workflows (i.e. presumably most of them, in the end), with Windchill as one of the best solutions. And, as promised, PTC continued to roll out thoughtfully developed AI features throughout the year (AI ✅). And last, Onshape - after years as a start-up in the shadows, Onshape finally broke out into the limelight, both in terms of coming to the fore in management discussions about strategic positioning and in a snowballing competitive displacement story, including the largest-ever Onshape enterprise deal (Onshape ✅).

Also on the Sweet side of things, and I’ll touch on this more below, I won’t deny that owning software stocks this year has been a headache. A headache that, in PTC’s case, potentially won’t fully subside until the company has well and truly proven the AI as a tailwind story. Precisely how long that might take remains unclear, but for now PTC sits very much in the ‘AI loser’ basket. In this context (again, I’ll touch on this later), the Schneider takeover has provided me with the opportunity to cash out with a respectable IRR of 30% while dispensing with all the ongoing drama wrought by AI.

The Bitter

With the journey cut short, there are lots of things that we’re not going to get to see play out, potentially missing out on a compelling longer-term compounding story. Again, not to rehash the complete thesis here (which includes other elements such as a traditional on-prem to SaaS conversion story), what I’ll miss most is observing the AI as a tailwind story develop. Notably, we saw a divergence in investor perceptions across the broader SaaS sector this year, and relevant for our purposes here, some of the data infrastructure players have already been judged ‘AI winners’ and accordingly re-rated. Well, if you believe that the industrial sector will ultimately adopt enterprise AI, then PTC’s Windchill is one of the most important data infrastructure plays. In PTC’s vision of the future, product lifecycle management is going to be agentic, with agents operating across the entire workflow within a framework of governance, permissions, traceability, security and controlled execution. At the centre of this would sit Windchill. As is the case with the other ‘AI winners’, the vision here also incorporates a growing consumption-based pricing component for advanced AI capabilities. But while PTC is moving with relative haste, we weren’t there yet - the debate about the relevant value capture was likely to continue for some time.

And, again, there’s Onshape. On the assumption that Schneider will be a good steward, I think Onshape might eventually be the biggest CAD system in the world in commercial terms - this is perhaps considered a fringe view, but nevertheless I favour the odds. Onshape is far better suited for the age of AI than legacy CAD systems. There are numerous reasons for this (which I won’t repeat - read the posts), but perhaps the most interesting one has really only recently emerged, and it relates to FeatureScript, Onshape’s native CAD feature programming language.

One prominent source of investor anxiety about PTC this year has been the increasing capabilities of frontier models in ‘one-shotting’ impressive-looking CAD models from text prompts. But as superficially impressive as some of these AI-generated designs are, there’s something else that frontier models are particularly good at and well suited to - I’m talking about coding, of course. And so Onshape’s pitch via FeatureScript is text-to-code-to-CAD. Your coding agent (e.g. Claude) writes the FeatureScript code, and FeatureScript natively implements the CAD geometry, with explicitly encoded engineering rules and design intent. The engineer still needs to specify and validate those rules but deterministic execution means the same code and inputs regenerate consistently. What interests me about Onshape here is the native integration and the creation of reusable engineering tools. The resulting custom feature becomes a native Onshape tool, with parameters engineers can adjust without returning to the AI model. Once developed and validated, it can be reused across designs and teams without repeated prompting or further AI token consumption.

But there’s more! One of the historical friction points to Onshape adoption was its relative lack of features compared with decades-old competitors. That’s one reason why FeatureScript exists - you can make your own custom features. But the problem was that it required coding skills, and not everyone could code. Onshape is, of course, introducing new features constantly, but if a user can’t wait, a coding agent like Claude can now help them build custom features themselves. This could lower one important barrier to switching: the difficulty of recreating specialised features and workflows.

There is, of course, the risk that frontier labs and others will more or less try to ‘brute force’ their models into being much better at CAD through specialised training, and it’s clear that some are trying. But one challenge is access to high-quality industrial CAD data together with the requirements, constraints and engineering rationale behind it. Much of the most valuable data remains within commercial businesses with little incentive to hand their IP over to labs - there may even be legal and regulatory reasons why they couldn’t even if they wanted to. This isn’t an insurmountable barrier: public datasets, synthetic data and feedback from executing generated CAD code provide other routes to progress. In any case, it’s still not entirely clear that improving CAD intelligence would necessarily disintermediate the FeatureScript approach. There’s much that remains unclear in the emerging world of AI as applied to design and engineering, but at the very least these were all lines of thought worth investigation.

NFTBC

Since starting NFTBC last year, PTC is the first stock to have completed its journey - albeit in a shorter interval than I might otherwise have expected. That said, when your focus is on businesses that are doing strategically fascinating yet under-appreciated things, it’s not uncommon for these businesses to be acquired from time to time. In the case of Schneider Electric, it turns out that they certainly recognised the strategic value in PTC and had been eyeing up the business for some time (as became apparent on their conference call) - the market action of 2026 provided the opportunity to strike. Additionally, for one reason or another, PTC seems to attract very little coverage outside the sell-side and tends to go very under-discussed in the usual places. This all made it an enticing choice for the NFTBC universe. So I was pleased by the reception when I started writing about it - it turns out that investors actually did want to read about this interesting business!

At NFTBC, my long-form posts are not intended to be pitches, so much as the start of a journey - an open-ended journey. Most of these businesses are trying something unconventional and possibly overlooked or misunderstood - or dare I say even ‘obviously wrong’. When a business promises to be higher growth and/or higher margin at some stage in the future, the typical response is scepticism, and rightly so much of the time. NFTBC endeavours to weigh things up ex ante and deploy a little imagination in doing so, rather than relying too heavily on extrapolation of the past or taking cues from others. There’s no avoiding the fact that, in many cases, it’s going to take time to see how these situations pan out - but that’s partly what makes it a satisfying and worthwhile pursuit. Actually, in most cases I would not expect the market to come round to my point of view quickly - there’s likely to be too little visibility and too much uncertainty for that. Again, with this context, I encourage readers not to take my long-form posts as near-term entry points.

Insofar as I provide any useful investment signals or insights, you should fully expect many of these to show up at some stage after the deep dive, most typically in The Digest - I tell full subscribers what I’m up to. For example, when I initially posted the PTC deep dive, I did not have a significant holding at the time - I had left plenty of room for averaging down and specifically highlighted this as my intention. In the event, I certainly did do some of that when the SaaSpocalypse got underway (see here and here, and also here). The long and short of it is that even though the agreed takeover price is only marginally above the share price at the time of my first PTC post, I still managed to wring out a respectable 30% IRR in 12 months of ownership. With hindsight, obviously the return could have been significantly higher if I had better timing and bought it all at the bottom - I hope some of you were able to do that!

Thanks for reading.

Subscribe now

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

A perennial topic with REGN is the question of whether they will ever consummate a big deal - will they, won’t they? For all his apparent disapproval of pharma M&A in general, CEO Len Schleifer has also been pretty clear that he’s not against it entirely - it’s just that the price has to be right and there has to be some kind of highly compelling strategic angle. With this post I raise the possibility that such a candidate could now be in sight.

But I also stress that the post is speculative, and I would assign only an outside chance of some version of its contents coming to pass (let’s say 10-20%). Nevertheless, I do believe there is a fairly strong logical foundation to it as well as a tantalising window of opportunity where the stars have all momentarily aligned. It’s probably now or never. And so, in the full knowledge that this outcome might be unlikely, it’s kind of intriguing to think it through and write about it anyway. If nothing else, hopefully we’ll learn something useful along the way.

Regrettably, there isn’t much of a free preview on this one (such as I typically like to provide) because I intend to come to the point quite quickly and that would give the game away too soon. That said, once you’re in, there’s plenty of fare to chew on: regarding background, strategic rationale, the uniqueness of the situation at hand and even a deal model you can download and play with if you want to. If you would like to read the post for free you have three options: 1) use your one free post voucher (if you haven’t used it yet) 2) use your week-long free trial or 3) successfully refer a friend:

Refer a friend

Or better still, I’d be delighted to welcome you into the elite tier of full NFTBC subscribers.

Subscribe now

Thanks for reading NFTBC.

Read more

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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, as well as an exercise in sharing some of my own updated thinking. You can simply scroll to whichever company you’re interested in and see very quickly if there’s anything notable 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

In a slight break with convention, The Digest is packed with earnings this week so I’m circulating a preview to the wider NFTBC mailing list - for the benefit of those who may be interested.

Here are a few of the highlights:

• BMW: following my prior best guess of a return to the 8-10% strategic automotive margin corridor by 2029, BMW are now officially communicating a target “by the beginning” of the 2030s - suggesting I was off by about a year. In either case, if they execute (and, breaking with consensus, I think they probably will), in several years EPS could be potentially multiples higher than the 2026 number.

• CBRS: expectations for the next couple of years are already very high, but I’ve been wondering lately if they’re high enough. The trail of breadcrumbs is growing longer. While the evidence is not yet conclusive, it is becoming harder to dismiss.

• OBCK: points for management credibility - 1) the business is now re-accelerating following the Q1 low ✅ and 2) they’ve signed a deal to divest the last ‘non-core’ business ✅ (in both cases as they said they would)

Subscribe now

Share

Read more

Friends of NFTBC

I hadn’t been planning another REGN post any time soon, but then something quite interesting happened this week. You can consider this post an addendum to June’s post on partnerships. It won’t be very long so I’m making it generally available.

This week Nature published a study describing a massive genomic study led by the Regeneron Genetics Center (RGC). The study unearthed rare truncating variants of the FNIP1 gene that are strongly associated with unusually broad improvements in metabolic health. In basic terms, carriers of these variants have lower BMI, more favourable fat distribution, less liver fat, higher lean body-mass percentage, lower atherogenic blood lipids, healthier blood sugars and a 60% lower odds of cardiometabolic disease.

To quote the headline news article (and RGC’s own Luca Lotta):

“For millennia, humans lived in a calorie-deprived environment, where every calorie counted,” says Lotta. “Pathways like this one helped the body spare calories for a rainy day - a sort of metabolic brake telling the body, ‘Don’t burn too much fat, you might run out.’” But we now live in a calorie-rich environment. Our genomes haven’t changed much, but our environment has, he says. “So, there’s a mismatch.”

In effect, disabling one copy of FNIP1 appears to fix this mismatch by acting on cellular metabolism. The prospect of translating the discovery into a medicine is therefore exciting. From the patent record it’s clear that REGN has already been contemplating drugging FNIP1 for at least several years. As to how close they might be to advancing a clinical candidate remains to be seen. When Len Schleifer was asked recently which of REGN’s MASH programs he was most excited about, his answer: “the ones I haven’t told you about. Stay tuned.” It makes you wonder…

RGC

There are still likely to be important translational questions outstanding on FNIP1 (I won’t try to address those here), but the therapeutic signal is unusually intriguing. In any case, weighing up FNIP1 is not the purpose of this post. It’s more about strategy and terminal value, such as we talked about in June. Here’s something else I wrote in June (in The Digest):

Aris Baras, head of the RGC, appeared on a podcast. It’s a great overview of what RGC is up to and some of their ambitions. Notably he talks about how REGN is currently putting around 10 RGC-discovered targets into pipeline programs each year. And he also mentions that they’re seeing non-linear benefits to scale. Think about it this way - they’re looking for rare genetic ‘superpower’ variants. You can only find these through sequencing populations at scale. But that’s not enough - you need to find enough people carrying the same variant (or, more commonly, different damaging variants in the same gene) to be highly confident that the gene-level associations are real. And that takes massive scale. They’re currently around 3m people sequenced and he seems to be suggesting that the pace of discovery is picking up at this scale, say, versus 1 or 2m. They believe they already have a clear path to 10m, but their new official ambition is 50m+ and as far as they know, no one else is trying to go that big. I think it’s plausible that REGN has invested a couple of billion dollars in RGC over the last 10 years, perhaps with a current run-rate of $300m+. Yet they have little to show for it yet in terms of new targets making it to revenue. Who else behaves this way in pharma? It’s exciting to see where RGC goes in the years ahead and what the eventual economic contribution could start to look like - I think it’s going to be big.

Size Matters

Arguably the most exciting publicly disclosed targets to come out of RGC so far are as follows:

Of the publicly disclosed novel targets so far, these are right at the high-end in terms of dataset size (i.e. people sequenced). FNIP1 was discovered in a substantially larger analysis than GPR75 or CIDEB and is among the clearest illustrations yet of the returns from mega-scale sequencing. Homing in on Aris Baras’ non-linear benefits of scale point, it’s pretty interesting that the FNIP1 signal emerged from an analysis of over 1m people, and is arguably the most compelling novel target so far. They found ‘ultra-rare protein-truncating variants in FNIP1’ that were only present in 1 in 7,000 people - this is what Baras was talking about by the need to go big.

Bearing in mind that these disclosures only take place at a significant delay to the underlying work (directionally patent filings give us an extra clue here), we can only imagine what RGC is currently cooking with a dataset that exceeds 3m and is on its way to 10m+ and perhaps eventually 50m. I think it’s pretty exciting.

I’ve seen numerous murmurings in the last 12 months to the effect that REGN is past its best and should now step aside on R&D in favour of smaller, more nimble, more innovative biotechs (that’s the general argument). But show me the small biotech (or even large pharma for that matter) that is willing and able to assemble all the required infrastructure and partnerships and put into effect and analyse a database of 10m+ people’s exomes/genomes (and increasingly, proteomes) linked to de-identified patient health records. It’s a massive undertaking and costly in both dollar terms and years. Over the last 12 years RGC has built a vertically integrated population-scale sequencing factory. Frankly, it is hard to imagine anyone recreating the same integrated system quickly or easily. So far as I can ascertain (and in line with Baras’ comments), no competitor has publicly committed to building a comparably integrated, 10 million-person sequencing resource for its own R&D purposes - the nearest disclosed pharma program I can find is materially smaller. In this context, RGC scientists have led or materially contributed to many (actually, I believe, most) of the largest exome-wide association analyses published yet.

Getting a Head Start

Now, if RGC’s discoveries were to remain effectively secret, the IP advantage would be inarguable. But in practice, RGC has certain data disclosure commitments - to partners and to its own researchers (REGN has always encouraged its scientists to publish, in the spirit of the original Genentech model). Moreover disclosing major discoveries in the name of science is, no doubt, the right thing to do in REGN’s own terms (‘doing well by doing good’). But that’s not to say that all important insights leak out - I’m sure they don’t. Nevertheless, major discoveries do not remain internal secrets - they enter the public domain through eventual publication in scientific journals or might be inferred from disclosures made in patent filings. So I think the bigger questions here are something like:

• How big of a head start can REGN get before actionable insights enter the public domain?

• How quickly can they act on that head start?

Both Len and George have talked recently about the increasing difficulty of getting a head start before someone follows. But it also looks like they’re adapting and acting a little differently to how they used to.

With HSD17B13 in MASH, REGN filed initial target IP and then around a year later told the world about the discovery at the same time they announced the Alnylam partnership to find and develop an RNAi drug - in practice they didn’t have much of a head start. And indeed, Arrowhead caught up very quickly. By GPR75 and especially CIDEB, REGN seems to have learned to feed RGC discoveries into a standing therapeutic-development machine privately, file both target-level and molecule-level IP, and only then publish the target. FNIP1 potentially takes this one step further, with almost three years between the earliest patent filing and the full scientific publication. Importantly, the patent itself became public in March 2025, so the target was not secret for that entire period. Nevertheless, REGN had approximately 18 months in which the target and therapeutic hypothesis were not public, followed by another 17 months during which competitors (if they went looking) could see the patent but not the complete evidence eventually published in Nature. I could be reading too much into it, but one plausible interpretation is that they’re allowing themselves valuable additional time for translational work, discovery and pre-clinical development before potential competitors get a first peek at the science behind the target. And if you think about it, the larger the dataset underlying the discovery, the less likely someone else is to have arrived at the same independently. FNIP1 took a million people vs 47,000 for HSD17B13. I suspect, at 3m+ people, REGN could have increasingly more headroom to act in the absence of competition.

And the same goes for target validation. Patents become public around 18 months after filing. And in these patents REGN leaves a trail of breadcrumbs around the target in question. At the very least patents can point competitors in the right direction and they can try to validate the target independently. But when variants become increasingly rare, independent validation becomes impractical in the absence of access to a massive dataset. The informational asymmetry grows as the variants get rarer. Taking it to the next extreme, suppose that RGC discovers a spectacular protective gene that’s 10X rarer than with FNIP1. The approximate carrier numbers become (illustratively):

Carrier numbers themselves increase linearly with cohort size, but the practical discovery benefit can be non-linear: an additional increment of scale can move a gene from statistically invisible to convincingly associated. At sufficient scale, RGC may have a significant window where it is the only group able to observe an ultra-rare association with persuasive statistical confidence (effectively a temporary monopoly on human-genetic insight). Until the complete evidence is published, competitors may be unable to reproduce the human genetic case - even if they can begin investigating the target through patents, public datasets and their own experiments.

Certainly once the translational work is sufficiently advanced we can assume that REGN can act expeditiously with their own VelociSuite antibodies, should the target in question be amenable to that approach. I think this is also the case with RNAi now via the Alnylam partnership - e.g. unlike HSD17B13 REGN appears to have a substantial head start on CIDEB, with their Alnylam-partnered molecule now in phase 2 (and no competitors in phase 1 yet, so far as I can see). In a similar vein, you can see REGN pushing the partnership strategy into further novel modalities to ensure they always have ready access to the appropriate technology (I find the new Parabilis partnership particularly intriguing in this regard).

Final Thoughts

Per Baras’ recent comments, RGC/REGN have put 50 RGC-discovered targets into actual pipeline programs (that’s +20 in the last two years, by the way). But in the clinical pipeline we can see only a small handful of these - the tip of the iceberg. There’s a lot that they just haven’t told us yet. Separately, my review of the public patent record identified at least 13 additional novel, genetics-derived targets around which REGN has filed patents, but for which it has not published detailed scientific evidence. That is a floor rather than a complete count - applications filed since February 2025 (give or take) will generally not yet be public. Patent filings do not, of course, prove that every target remains part of an active pipeline program however.

RGC practically never comes up in the general investment discourse on REGN - even for those who know about it it’s much too jam tomorrow to be considered relevant in a market environment that operates from quarter to quarter. But it is highly relevant in my view: a unique hard-to-replicate asset that 1) regularly turns out compelling novel drug targets and 2) whose capability and informational advantage only compound further over time through scale. In practical terms for investors, if RGC/REGN can in the coming years demonstrate 1) a consistent pipeline of compelling novel RGC-derived programs converting to the clinic and 2) a consistent ability to exploit their head start, then I think RGC will be difficult not to notice. I suspect that RGC’s increasing scale as well as the facilitating platform partnership approach make this more likely than not. So my contention is that if/when RGC does eventually enter the discourse as an obvious durable advantage, investors may have to revisit conventional thinking around terminal value.

But let’s not forget the sceptics: scale alone does not make medicines. Rare-variant associations may ultimately prove undruggable or intractable, target-level patents might be designed around, most pharma development programs fail and RGC has not yet contributed much revenue. Those caveats might explain why investors discount RGC - but not, in my view, why they should largely ignore it.

Thanks for reading.

Subscribe now

Share

It’s certainly a sign of today’s markets that Bruker can meet externally communicated expectations and maintain guidance for the year, yet still plunge more than 20% in one session (welcome to The Beauty Contest). I think there are two relevant explanations: 1) investors had assumed prior guidance to be too conservative and 2) BRKR’s recent run-up was closely tied to inclusion in various fashionable thematic trades (semiconductors and AI-in-biology). It’s probably a bit of both. If you’d put my feet to the fire, I’d have tentatively leaned towards a revenue beat for Q2 on the basis of recent results for LST peers but I wasn’t sure given BRKR’s idiosyncrasies (see here). But much more importantly I had been turning increasingly cautious on the stock, seeing it as highly prone to correction - full subscribers will have known this already (e.g. see here and here) and that I had been dialling down my own holding quite considerably. Herein we’ll take a look at the state of affairs and I’ll share a few reflections on BRKR’s future prospects - which, for the most part, I remain excited about.

Occasionally, an unrelenting share price decline in the face of no particular news is itself a source of signal to be taken seriously rather than noise to be ignored. I warned NFTBC readers about the possibility of Procept investors positioning for a negative guidance revision perhaps on the basis of privately obtained claims data here (and more recently here) - staring down the barrel of this inherent uncertainty is what prevented me from dialling up my own thus-far modest holding. And it looks like some version of this turned out to be the case with Procept in Q2. But the forward-looking picture is more complex than a simple cut to procedure expectations - there are a number of moving parts (some of which are quite encouraging) and the plan here is to weigh them up as best I can.

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

Subscribe now

Share

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

When I started writing about REGN last year, I talked about how Eylea had become “a distraction, a fixation even”. This is inevitably going to be the case in biotech when revenue is concentrated in any single molecule, but in the end, the fears proved to be unfounded - REGN has grown through the handoff from Eylea 2mg to Eylea HD. Eylea finally appears to be evolving from a hindrance into something potentially helpful once again.

One of the most entertaining quirks about Regeneron is the steadfast refusal to give revenue guidance of any kind.

I don’t like to make predictions about things that I don’t have any more information than you do. You can look at all the same information I have, and you can make your best guesses

- CEO Len Schleifer, speaking on the inSpired Podcast recently

He is, of course, being selectively candid here even if directionally fair. Ultimately there’s a tradeoff: REGN could perhaps provide considerably better forecasts than the analysts can. But it would come at the cost of having to play someone else’s game and entail a certain loss of control and freedom. To start operating that way might not suit the company philosophically either (also Goodhart’s Law etc). But the consequence of leaving analysts guessing is pretty wild volatility in revision trends. Here are consensus revenue trends over the last three years:

And let’s not forget that we’ve just about arrived at the two-year anniversary of REGN hitting its all-time high:

Yet despite a stock that remains down near 40% and a veritable rollercoaster ride in the forecasts (question: which one caused the other?), it looks as though REGN is on track to surpass future-year revenue forecasts made back at the highs - or the forecast for 2026 at the very least. When I wrote the deep dive last year, I suggested this advice:

if you err on the side of management’s ambitions rather than whatever sell-side analysts plug into their models, you’ll end up nearer to the real world. It sounds simplistic, but over the longer-term, the consensus numbers are always too low

It probably won’t have come as a surprise to NFTBC readers that revenue and EPS estimates for 2026 suddenly had to shoot upwards this week - I had been writing since last year that they seemed much too low (e.g. see here, here, here and here).

With that let’s take a look at some of the standout things from REGN’s latest quarter. You can find all the Q2 materials here.

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

Subscribe now

Share

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