Bullpen.News From Edwin Dorsey
Get Daily Recap Feedback

Thursday, September 17, 2026

Back to today

Power Semiconductors: The Follow-Up

STMicroelectronics, Navitas and Wolfspeed offer contrasting power-semiconductor cases tied to NVIDIA’s 800V AI data-center architecture, spanning an incumbent leader, a speculative…

STM $47B
SixSigmaCapital: Navigating the Financial Markets 1 click

From X

@RagingVentures

Raging Capital Ventures on X

But if returns on capital are declining for everyone, the choice is whether to eat or to be eaten? $T is willing to subsidize fiber as a loss leader to protect their wireless margins. $CHTR / $CMCSA are likewise willing to give away wireless to protect their broadband margins… which comes back to the question: which is stickier, your phone or broadband line? I think $CMCSA works because of the spin unlock and sheer strength of their balance sheet. $CHTR is a very long dated call option with more risk, but much more upside if broadband losses attenuate and/or they allocate capital intelligently (although $CHTR’s balance sheet is more levered, there is tremendous optionality in its duration, cost and structure).
@blondesnmoney

Cluseau Investments on X

Disclaimer: I have sold this position and have no view on the valuation or future prospects My craziest OTC adventure this year was discovering a single-asset refinery that sells Jet Fuel to the US Military. It sounds kinda crazy, but it's the exact type of ridiculous setup funny enough to look into. In July, refining crack spreads were blowing out, and nearly every refiner $PBF, $PARR, $VLO, $PSX were steamrolling. I wanted some exposure, but I felt sheepish chasing names that were already tearing it up. I started looking at smaller refiners, but nearly everyone was in the 3x - 5x forward earnings range, not expensive, but for a cyclical, also not really that cheap. It turns out that, aside from the megacap refiners and $PARR, there's only one other US listed microcap refiner, Blue Dolphin Energy $BDCO. The stock is heavily insider owned, and at the time, was trading at just over 4x Q1 earnings ($0.99). Crack spreads are publicly viewable, and other refiners were demolishing estimates - $PARR came in at $11 in Q2, triple Q1, so it didn't take too much math to realize BDCO's Q2 was gonna be strong given jet fuel prices, and that the stock was potentially trading under 1x forward earnings. Of course, it's an OTC name with limited liquidity so you can't realize size it up, but how often do you get a chance to buy someone who supplies jet fuel to the Department of Defense under 1x annualized earnings? It was almost comedic, the story was funny enough to swallow the OTC stigma and just pick up a little bit up. About a week later, Q2 results came in at $1.19 (putting 1H earnings at ($2.18), and the market suddenly realized a stock trading at $4 on $2.18 of two quarter's EPS was just too cheap. The stock hit $18 today - I've been trimming since entry and am finally out. It was a tiny position, but I continue to find it ridiculous that these OTC setups exist. With a small amount of capital and a love for reading filings, you can probably find a slew of OTC names that are actually decently profitable that are simply too small for larger investors to take a stab at. I have no position in BDCO and have no opinion on it, $4 and $18 are totally different prices, but the fact this setup existed and somehow flew under the radar until earnings is pretty entertaining.
@convequity 1 click

Convequity on X

Huawei just updated the Ascend 960 & 970 roadmap. $NVDA's competition pressure is increasing from both $AVGO and Huawei continuously as we've anticipated. → 960DT rollout pulled forward from 4Q27 to 1Q27. (Seems like they are finally getting the HBC-like, tailored-memory approach closer to high-volume production.) → 960PR by 3Q27 with 8 PFLOPS FP4—double the 960DT’s FP4 compute. (Closing in on NVDA’s Blackwell Ultra, which tops out at 15 PFLOPS dense FP4. Rubin moves that target to 35 PFLOPS dense FP4.) → 970 by 2028 with 14 PFLOPS FP4, up from the previous 8 PFLOPS target. Overall, this looks consistent with our Tau Scaling thesis: a higher compute ceiling, more tailored chip designs, and a faster path to production. Not every specification has doubled, but a near-doubling of the 970’s planned FP4 compute alongside an accelerated 960 rollout is a meaningful change. The bigger story, in our view, is Huawei’s CIDM model. Optimizing the entire vertical—not just the logic die—creates more room to change the compute design and shorten the development cycle. Huawei’s disclosed co-optimization already spans devices, circuits, chips and systems. This is the kind of vertical integration Elon Musk is targeting with Tesla and SpaceX’s Terafab. But Huawei is already putting its ecosystem version to work. (Would be quite funny if the Chinese ecosystem gets accelerator-based EUV into production lithography first, too.) Against NVDA, our thesis is that the single-chip gap narrows from roughly three years toward two—and potentially less as the roadmap accelerates. From a PPAC perspective, Tau Scaling could bring the logic die within a year—or less—while memory remains further behind. Bandwidth, not simply capacity, looks like the harder bottleneck. But Huawei may be able to lean into memory capacity and co-design the LLM and serving strategy around that constraint, rather than simply copy something optimized for NVDA’s hardware balance. System-level performance still needs to be proven at the cluster sizes Huawei is targeting. But we would not be surprised if it works better than expected. Even before the full UnifiedBus vision is realized, we think Huawei could gain system-level performance leadership on some large-model workloads sooner than many expect.

Investor Letters

All letters

A2Gold: Research Note

A2Gold's McIntosh drilling found 0.23 g/t gold over 65.5m, 370m southwest of the existing resource, and extended mineralization 86m…

KGC $34B
Atrium Research - Equity Research 2 clicks

GLP-1 winners and losers

GLP-1 drugs are gaining traction across Asia as Chinese developers advance candidates and Indian generics slash prices, reshaping prospects…

NVO $175B
Asian Century Stocks

Adobe 3Q26 Update

Adobe’s fiscal 3Q26 revenue grew 13%, while ARR growth slowed to 9.3% excluding Semrush.

ADBE $99B
Speedwell Memos 2 clicks

Ferrari Exploratory Report

Ferrari sells about 13,000 vehicles annually, generating €7 billion of revenue and 28% operating margins; its shares trade at 32 times…

RACE $73B
Speedwell Memos 1 click

Meta: 4Q25 Business Update

Meta’s 4Q25 revenue rose 24% as AI lifted ad performance, but 2026 costs are guided up about 41% and capex to $115–135 billion.

META $1.7T
Speedwell Memos 1 click

Affirm (AFRM): Buy Now, Cry Later

Affirm, a $25 billion buy now, pay later lender, faces a short thesis comparing its model with failed subprime monoline finance companies.

AFRM $23B
Kerrisdale Capital

Portfolio Update - January 2026

Portfolio update exits The Trade Desk, QXO and Park Hotels, then adds Tencent at a 35% discount, SoftBank for AI exposure and HelloFresh at…

QXO $13B
Modern Value Investing

Publications we follow

Every publication on this site, hand-picked.

108 publications

Newsletters 91

YouTube 4

Blogs 12

Seeking Alpha 29 of 159

Nominate a publication

X accounts we follow

Every account on this site, hand-picked.

69 accounts
Nominate an account

Investor letters we follow

78 firms

Sectors

Ordered by how often our editor reaches for them.