Bullpen.News From Edwin Dorsey
Get Daily Recap Feedback

Saturday, October 31, 2026

Back to today

From X

@convequity

Convequity on X

Further thoughts on the $CBRS short-seller report I've already quote-tweeted this short-seller report, but I think it's quite funny. These days, a lot of tech short sellers seem to have a poor understanding of the technology itself. Perhaps they're better at timing market liquidity and identifying moments when sentiment is most vulnerable. Going back to the report's central claim, I think it's arguing that Cerebras hasn't adequately addressed the challenges of using a giant substrate to support an entire wafer, or of maintaining reliable connections across such a large chip. It's similar to the challenges facing $NVDA's Rubin Ultra. The idea is to connect four compute chips, with a combined size roughly eight times the reticle limit. At that scale, thermal expansion becomes a major issue, as does maintaining consistent communication between the chips. For Cerebras, these are certainly significant challenges. That's why it took so many years to turn the original idea into something that actually works. I would say the third-generation Wafer-Scale Engine, WSE-3, was the first generation to become commercially attractive. Earlier generations had lots of problems to work through, much like other chips, for example Google's TPU from v1 to v3. Even WSE-3 hasn't solved everything. One of the biggest remaining issues is parametric yield. The simplest way to explain it is that, when a chip spans an entire wafer, you can't guarantee that every part will operate reliably at the same high frequency. (Frequency is how many cycles per second the chip runs, so a higher frequency means more compute from the same silicon.) You therefore have to lower the operating frequency of the whole system. A normal chipmaker would cut the wafer into many separate dies and set each one's speed on its own. The argument is that those dies could then run at frequencies 50% higher, or perhaps even twice as high. Cerebras keeps the whole wafer as one chip instead, so it gives up some of the compute the same silicon could otherwise deliver. In return, it gets an enormous amount of compute and on-chip memory connected as one wafer-scale chip. Data moves between the cores over wiring on the wafer itself, rather than through slower, more power-hungry links between separate chips. That is what lets Cerebras serve AI models so quickly, especially for inference. That's the trade-off, and parametric yield has been a constraint for Cerebras. The workaround has been to lower the specifications and operating frequency, leaving enough headroom for the system to work reliably. But that also means leaving some of the wafer's potential unused, despite manufacturing it on an expensive, supply-constrained TSMC N5 process, and potentially on more advanced nodes in the future. That's why I find the latest CS-4 rack-level system, using the WSE-3T, interesting. My understanding is that WSE-3T uses the same wafer design as WSE-3 but delivers more than 50% higher performance. They've apparently found new ways to address the parametric-yield issue, allowing them to unlock frequency headroom they previously couldn't use reliably. It looks as though they're actively addressing the problem and making good progress. The earlier implementation may have been suboptimal, but even that version was already demonstrating commercial competitiveness. This is a common problem with short sellers today: they build a thesis around old information while the technology is evolving rapidly. They point to a technical limitation that may still exist, but fail to distinguish between a problem that makes a product unworkable and one that merely reduces its performance. Between roughly 2015 and 2021 or 2022, these issues may genuinely have been showstoppers. By late 2026, I don't think that's necessarily true anymore. They may still hold performance back somewhat, but they don't prevent the product from working. Meanwhile, Cerebras has the advantage of being the only company to have made a wafer-scale chip work. In my view, that benefit outweighs the remaining limitations. The short seller doesn't seem to recognize that distinction. I think this also explains why Oracle, AWS, OpenAI, and other companies have begun testing Cerebras chips this year. The technology seems to be reaching a point where these problems are largely, though not entirely, resolved. Previously, perhaps they were only halfway to a solution, and the product wasn't sufficiently practical. Now that they're much closer, large-scale commercial deployment becomes more attractive to customers. Of course, it remains to be seen how this plays out. They still haven't crossed the deployment threshold I have in mind, and once they do, we may see new problems emerge. For now, though, my sense is that these are well-known issues among people familiar with the technology. The short seller is largely repeating things that those people already understand. The more interesting question is the direction of progress: how quickly are the problems being resolved, and what does that mean for the product's competitiveness? The report doesn't really address that. The rest of the report seems to cover the usual short-seller material, raising fear, uncertainty, and doubt about adoption and other peripheral issues. But the central technical thesis strikes me as quite weak. It reminds me of previous short reports on companies such as $SMCI.
@convequity

Convequity on X

Thoughts on Elon's recent neocloud comments "Neocloud" has always been a weak label. These businesses aren't really clouds, and they aren't particularly new. The more accurate term is GPU VPS. Most of them rent bare-metal GPUs. They generally don't offer redundancy across three availability zones, and they don't sell hundreds of platform services on top of the hardware the way a true cloud does. That puts them closer to VPS providers than to cloud platforms. CPU history shows the difference. CPU cloud concentrated among three to five players because platform services lock customers in and drive a demand flywheel. Customers use more of the stack over time, rather than only renting the hardware. That is how AWS-style clouds kept high margins and strong moats. CPU VPS went the other way: many providers selling essentially the same product, in a brutally competitive, commoditized market. If GPU supply stops being constrained by chips or energy, GPU VPS starts to look like CPU VPS, with gross margins of something like 50%. Better marketing and operations can lift margins and grow revenue, but they won't push gross margins to 75-80%. That is the long-term risk Musk pointed to when he predicted most neoclouds would go out of business. Most of these names are VPS businesses earning cloud-like profits and valuations from a temporary imbalance between supply and demand. $CRWV is part of that group. CoreWeave and $NBIS are also the two furthest from a pure GPU VPS, because both already sell platform services on top of the GPUs. CoreWeave is still the most exposed on financing. It pushed hard to maximize margins on each contract and took on interest-rate risk against fixed contracted revenue. Its counterparties, OpenAI, $MSFT and $META, may have played those deals more intelligently. So CRWV is the most vulnerable name on its balance sheet, yet one of the least commodity-like on product. That gap is where the return potential sits for investors willing to stick with CRWV.

Investor Letters

All letters

TSOH Weekly Roundup (10/30/2026)

Markel spent $315 million repurchasing shares in Q3 as Thrive agreed to buy ADT for $7.6 billion and Hasbro cited rising adult toy demand.

MKL $23B
TSOH Investment Research 1 click

Another Haircut with Avy the Barber

SoFi’s stock case hinges on interest rates, tangible book value and private-credit loan sales after its $1.5 billion equity offering.

SOFI $22B
WHAT ARE WE DOING? “Contrarians At The Gate” 3 clicks

Twilio CEO: ‘Buy the dip!’

Twilio CEO Khozema Shipchandler urged investors to “buy the dip” after earnings sent the software maker’s shares down 17% after hours.

TWLO $39B
The Technology Letter 2 clicks

CoStar: 3Q26 Business Recap

CoStar’s 3Q26 revenue growth slowed to 14% as bookings fell to $37 million, while cost cuts lifted EBIT margin to 8.8%.

CSGP $12B
Speedwell Memos 3 clicks

My Last GFL Update - Hopefully

GFL’s take-private structure hinges on BC Partners rolling 60% of its stake to preserve debt portability under the 40% permitted-holder…

GFL $16B
310 Value’s Newsletter 5 clicks

Are Neoclouds Dead?

Neoclouds face tighter debt financing as AI infrastructure capex outruns cash flow, leaving cash-burning GPU providers more exposed than…

NBIS $46B
Daniel Romero 9 clicks

Reddit: Deep Dive Analysis

Reddit generated $805 million in Q2 2026 revenue, 95% from ads, but slowing Google search referrals expose AI-driven traffic and…

RDDT $26B
Hated Moats 6 clicks
SBAC · long

Sba Communications CORP

SBA Communications trades at 14.9x forward AFFOPS, near a record low, despite being pitched as the market’s cleanest tower operator.

SBAC $20B
Value Investors Club

Publications we follow

Every publication on this site, hand-picked.

129 publications

Newsletters 109

YouTube 4

Blogs 15

Seeking Alpha 67 of 266

Nominate a publication

X accounts we follow

Every account on this site, hand-picked.

91 accounts
Nominate an account

Investor letters we follow

190 firms

Sectors

Ordered by how often our editor reaches for them.