Bullpen.News From Edwin Dorsey
Get Daily Recap Feedback

Sunday, October 11, 2026

Back to today

From X

@convequity 4 clicks

Convequity on X

OpenAI and Anthropic will rent most of the AI capacity they plan, but mainly from the big clouds rather than the listed neoclouds Convequity's AI Bubble Barometer (ABBX) tracks the data-centre capacity each company plans and how much of it it expects to rent from partners rather than own. 1. OpenAI plans 21.2 GW, of which 18.0 GW is to be supplied by partners. Anthropic plans 13.8 GW, all of it from partners. Together the two labs plan to rent 31.8 GW. 2. That is 2.8 times the 11.2 GW that the five hyperscalers, $MSFT, $AMZN, $GOOGL, $META and $ORCL, plan to rent from others. 3. The labs' named suppliers are mostly AWS, Oracle, Google, Azure and SB Energy. CoreWeave, a listed neocloud, and Nscale, a private one, appear on the lists, but the bulk of the 31.8 GW is not with listed neoclouds. 4. The seven listed neoclouds, $CRWV, $NBIS, $IREN, $APLD, $CORZ, $CIFR and $BTDR, plan to build 33.62 GW of their own. Much of that is gross site power rather than contracted IT capacity. 5. The labs are also behind most of the recent growth in AI capacity plans. Since 14 Jul, the total capacity that the 15 ABBX companies say they plan to have has risen by 22.19 GW, from 131.23 GW to 153.42 GW on 7 Oct. OpenAI and Anthropic raising their own plans accounts for 13.7 GW of that rise (62%). We think the labs' rental demand mainly supports $AMZN, $ORCL, $GOOGL and $MSFT. The neoclouds will need to fill their pipelines largely with other customers, including the hyperscalers' own 11.2 GW of planned rentals, so we would not read the labs' plans as demand for the neoclouds.
@StableBread

Fajasy on X

The Department of Energy conditionally committed $17.5B in loans in June 2026 for long-lead equipment for up to 10 AP1000 reactors (Westinghouse’s large U.S. design), and only “a limited number of heavy engineering plants” can make the components, per World Nuclear News. BWX Technologies $BWXT bought Precision Components Group (PCG) in July 2026 to get a U.S. foothold. PCG makes heat exchangers and other heavy components at plants in York, Pennsylvania, and Florence, New Jersey, but it can’t make the largest parts, like an AP1000 reactor pressure vessel, CEO Rex Geveden said on the Q2 call. So $BWXT is weighing New Jersey, Indiana, and one other location for a new U.S. large-component plant, and Geveden expects “a final investment decision in the coming months.” He also expected at least one new-build equipment order by the end of 2026, from the three additional small modular reactors (SMRs) planned at Darlington in Ontario, the AP1000 projects, or BWRX-300 SMR projects in the U.S. John MacQuarrie, president of Commercial Operations, widened that window at $BWXT’s Investor Day on September 29, 2026, saying awards should start “either later this year or early in 2027.” $BWXT’s Investor Day targets call for high-teens annual revenue growth in the commercial segment through 2030, with adjusted EBITDA margins rising to the high teens from 13% in 2026. Reading “high teens” as 16-19%, commercial revenue would grow from ~$1.1B in 2026 to ~$2.0-2.2B by 2030, and adjusted EBITDA from ~$140M to ~$320-420M. That would make commercial about a third of $BWXT’s adjusted EBITDA, up from about a fifth.
@convequity 4 clicks

Convequity on X

Neoclouds trade at 2.6x invested capital, less than half the hyperscalers' 6.1x The seven neoclouds in Convequity's AI Bubble Barometer (ABBX), $CRWV, $NBIS, $IREN, $APLD, $CORZ, $CIFR and $BTDR, lost 7.2% of their combined enterprise value this week, from $190.2B at the 2 Oct close to $176.5B at the 9 Oct close. Over the same week the five hyperscalers, $MSFT, $AMZN, $GOOGL, $META and $ORCL, gained 2.4%, from $13,316B to $13,637B. 1. Enterprise value divided by invested capital (EV/IC) at the 9 Oct close was 2.6x for the neoclouds and 6.1x for the hyperscalers. At the 2 Oct close it was 2.8x and 6.0x. 2. ABBX estimates the weighted average cost of capital (WACC) at 14.6% for the neoclouds and 9.7% for the hyperscalers. 3. At the 9 Oct close, neocloud share prices were 20% to 45% below each stock's best Friday close since 17 Jul: CIFR −44.5%, CORZ −33.3%, BTDR −25.8%, IREN −24.6%, APLD −24.1%, CRWV −22.0% and NBIS −20.4%. Among the hyperscalers, ORCL was 10.9% below its best, AMZN and META 4.4% and GOOGL 1.3%, while MSFT closed above its best. We think most of the discount reflects the neoclouds' higher cost of capital rather than a mispricing. We would not call 2.6x cheap until they report their July to September backlog and customer prepayment figures.
@convequity 2 clicks

Convequity on X

A laser that only one company can supply will not scale to the volumes AI needs. That is the main reason $LITE is a less attractive investment to us than it was a few quarters ago. Our post on Lumentum got more attention than we expected, so we want to set out the argument in plain terms, including what we think about microLED. We did not say microLED will certainly happen. 1. The part in question is Lumentum's ultra-high-power external laser source. In co-packaged optics, the optical parts sit inside the chip package. The laser sits outside, in its own module, and its light is carried into the package by fibre. That outside module is the external laser source. Ultra-high-power means one laser produces enough light to feed many optical channels at once. 2. If only one vendor can make that laser, it has little chance of reaching meaningful volume. One supplier cannot carry a part the whole industry needs. So the industry has to find other paths around that bottleneck. 3. The first likely path is micro-VCSELs. A VCSEL is a vertical-cavity surface-emitting laser: a tiny laser that sends light out of the top of the chip rather than out of its edge. They are made many at a time on a wafer, and each one is low power. A micro-VCSEL is a very small version. Many small lasers do the job of one very powerful one. 4. The second path is a less powerful laser packed inside the NPO package. CW stands for continuous wave: the laser emits a steady beam, and a separate part, the modulator, switches that light on and off to carry data. NPO is near-packaged optics: the optical engine sits on the board right next to the chip, rather than inside the chip package or in a pluggable module at the front of the switch. In this path the CW laser is still fairly high power, but not ultra-high power, and it sits inside the optics package instead of in a separate external module. 5. The third path is microLED. A microLED is a very small light-emitting diode, the same kind of device as in a display, not a laser. Each one is switched directly to carry data, and many of them run in parallel over short distances. 6. On microLED, the consensus on X is that it will not work for AI interconnects. We are not saying it will certainly happen. It is one of several alternatives. But there is good research work on it, and it has a chance. Most of the criticism we see is aimed at one architecture. The research we rate uses microLEDs in a different architecture, and the criticism does not address it. 7. Lumentum will still grow. But compared with the previous few quarters it is a less attractive investment, because its ultra-high-power laser is the part the industry has to find a way around. We trimmed most of our Lumentum position earlier this year, around when the share price began to flatten.
@StableBread 1 click

Fajasy on X

Neither of the two commercial prospects BWX Technologies $BWXT has named for its BANR microreactor is a data center. 1) Prodigy Clean Energy, a Montreal-based developer of factory-built nuclear plants: Two BANRs for a transportable pilot plant in Belledune, New Brunswick, built mainly at $BWXT’s Ontario plants and running in the early 2030s. New Brunswick’s government signed a letter of intent (LOI) to buy the power. $BWXT and Prodigy are still negotiating a definitive agreement, and the release provides no dollar value. 2) Tata Chemicals Soda Ash: A Wyoming soda ash producer that signed an LOI to explore deploying up to eight BANRs. $BWXT also markets BANR to data centers. However, I found no data center contract, customer, LOI, or revenue in any $BWXT press release from January 2025 through October 2026, in the FY2025 10-K, or in the 2026 10-Qs. $BWXT’s own mentions stop at a product page saying BANR can provide “clean electricity for municipalities, data centers and campuses,” and slides 11 and 25 from its Investor Day on September 29, 2026, which list “AI/data center demand” as a driver of nuclear demand in general. I didn’t hear the phrase “data center” mentioned once during Investor Day, and slide 49 lists data centers in the third phase of $BWXT’s microreactor plan. CEO Rex Geveden said at Investor Day that “the incremental demand from AI right now is 60 GW” and “winning at AI means winning at power, and winning at power probably means having a nuclear solution.” If he’s right, $BWXT benefits by supplying components, fuel, and services to more reactors, not by selling microreactors to data centers. Advanced Reactor Design and Engineering, the product line that includes Project Pele (its prototype microreactor for the Pentagon) and $BWXT’s other microreactor programs, brought in $203.8M in 2024 and $147.1M in 2025, then $52.6M in H1 2026, down from $69.8M a year earlier. At 3.0% of $BWXT’s $1,761.8M of H1 2026 revenue, microreactors are still a small, early business.
@StableBread

Fajasy on X

BWX Technologies $BWXT is in military base power three ways, through (1) Project Pele, a prototype reactor it’s building for the Pentagon, (2) the TRISO fuel its microreactors run on, and (3) Janus, the Army program that chose $BWXT’s commercial reactor for Fort Campbell, Kentucky. Pele is a 1.5-megawatt (MW) transportable, gas-cooled microreactor built for the Pentagon’s Strategic Capabilities Office under a 2022 award. $BWXT delivered its full TRISO fuel core to Idaho National Laboratory (INL) in December 2025, received the Department of Energy’s (DOE) preliminary safety approval in September 2026, and targets delivering the reactor to INL in 2027. TRISO (tri-structural isotropic) fuel, which the Department of Energy calls “the most robust nuclear fuel on earth,” packs uranium into particles coated in carbon and ceramic layers. $BWXT produces TRISO at Lynchburg “in the hundreds of kilograms quantities,” Joe Miller, president of Government Operations, said at its Investor Day on September 29, 2026. It also supplied the fuel for Antares, a microreactor startup whose reactor was the first new reactor to reach criticality (a self-sustaining chain reaction) under DOE’s reactor pilot program in June 2026. POWER magazine, a power-industry trade publication, reports $BWXT is “the only Janus vendor that self-supplies its TRISO fuel.” $BWXT is also negotiating a commercial TRISO plant in Wyoming with Kairos Power, an advanced reactor developer, at up to $500M of capex. CFO Mike Fitzgerald said at Investor Day that $BWXT needs to see high-assay low-enriched uranium (HALEU, the 5-20% enriched uranium most advanced reactors need) become available before committing to the plant, and “probably will not make a decision on that for a little while.” The Army chose BANR (the BWXT Advanced Nuclear Reactor), a 20 MWe (megawatts of electricity) gas-cooled design scaled up from Pele, for Fort Campbell. $BWXT plans to start construction at the site in late 2028 and have the reactor running in the early 2030s. Three details matter here: 1) No contract value: $BWXT’s release describes “a phased contracting approach.” For comparison, the Janus agreement for Radiant, a microreactor startup, is worth up to $750M. 2) $BWXT will own the reactor: The Army chose vendors “to own, construct and operate nuclear microreactors” ($BWXT’s release), under fixed-price, milestone-based agreements in which the Army pays at milestones rather than buying power. 3) Not first in line: With operation in the early 2030s, $BWXT’s reactor won’t meet the September 30, 2028 deadline a May 2025 executive order set for an Army-regulated reactor running on a U.S. military base. Radiant and Antares, two of the other Janus vendors, both target 2028 for their first reactors. After Radiant’s agreement, $1.45B of the program’s $2.2B is left, and an even split among the other four vendors would give $BWXT ~$360M over fiscal 2027-2031, or $72M/year, under 2% of 2026 revenue guidance.

Investor Letters

All letters

TSOH Weekly Roundup (10/09/2026)

Paramount Skydance’s debt-heavy Warner Bros. Discovery acquisition, streaming bundle pricing and alcohol-industry diversification anchor…

WBD $78B
TSOH Investment Research

Boyz in the Hood

Robinhood’s AI agents, 24/7 equity trading and 10:1 crypto perpetual futures encourage higher retail trading velocity, leverage and…

HOOD $91B
WHAT ARE WE DOING? “Contrarians At The Gate” 1 click

Top Rebound Stocks for Oct’26

Intuitive Surgical, down 32% from its January 2025 peak, faces slower US procedure growth and new robotic-surgery rivals despite $10.1…

ISRG $149B
Rebound Capital 1 click

Publications we follow

Every publication on this site, hand-picked.

129 publications

Newsletters 107

YouTube 4

Blogs 17

Seeking Alpha 49 of 225

Nominate a publication

X accounts we follow

Every account on this site, hand-picked.

88 accounts
Nominate an account

Investor letters we follow

189 firms

Sectors

Ordered by how often our editor reaches for them.