Thursday, November 5, 2026

Back to the front page

From X

@convequity 1 click

Convequity on X

Circular AI financing keeps getting bigger. A record $63 billion vehicle, partly supported by Broadcom ($AVGO), will buy Broadcom chips for Anthropic, and it only pays off if enterprises adopt AI beyond coding. The vehicle borrows, buys Broadcom XPU accelerators (custom AI chips) and networking gear, and leases them to Anthropic. Deliveries are expected from 2027. Broadcom guarantees a residual value of $40B+. This is the latest example of the banks, hyperscalers, model labs and chipmakers are frontrunning AI demand they expect to be insatiable. Without faster enterprise adoption, that demand may not arrive, and the financing stays circular. AI has to spread from coding into the rest of the enterprise. Coding can be checked, because code has to compile and pass tests. Other domains need the same verifiability, which means codifying context that lives in employees' heads, or at best across email, Slack, WhatsApp and work-management platforms. Vendors that can codify that context should extract immense value. Palantir ($PLTR) is the closest so far. Palantir's ontology, a structured account of an enterprise's objects, relationships and rules that software can use to check a result, is critical because it is instrumental to verifiability. We think it can make an average Nemotron model state of the art for that enterprise's own work. If the result is good enough, enterprises will use it and achieve sovereign AI. Once that context is codified, enterprises will not want to blindly run closed frontier models on it and give their know-how away. Broader adoption depends on data sovereignty. The shift has been visible since Palantir's Alex Karp warned earlier this year that handing closed frontier labs their data, models and institutional know-how means losing control of the value. The path is intelligence sovereignty plus more verifiability, then greater enterprise adoption, which would justify this frontrunning. It would also benefit Nvidia ($NVDA), which needs enterprise adoption because hyperscalers and model labs are leaning toward custom chips.
@StableBread

Fajasy on X

35% of TOYO Co., Ltd’s $TOYO H1 2026 revenue came from one customer it doesn’t name, and 22% came from companies under its own controlling shareholder. $TOYO manufactures solar cells, the processed silicon wafers that turn sunlight into electricity, and assembles some of them into panels (modules). Its cells use TOPCon (tunnel oxide passivated contact), a common high-efficiency silicon cell design. $TOYO runs three factories: 1) Ethiopia (4 GW of cells): Started production in April 2025, six months after $TOYO announced the plant, added a second 2 GW line in August, and reached its 4 GW nameplate in October 2025. 2) Vietnam (2 GW of cells): Running since H2 2023, but its cells now go to non-U.S. markets like India and Taiwan, since the U.S. Commerce Department set anti-dumping and countervailing duties (taxes on imports sold below cost or subsidized) on Vietnamese cells in April 2025. 3) Texas (2 GW of modules): A leased 567,140-square-foot plant near Houston that started its first 1 GW line in October 2025 and had completed its second by early October 2026. It assembles panels from $TOYO’s Ethiopian cells. VSUN (Vietnam Sunergy JSC) is the Vietnamese panel maker $TOYO was carved out of, and both companies sit under the same majority shareholder, Abalance, a Japanese solar group. VSUN is still $TOYO’s lender, owed $51.4M including interest, and a past wafer supplier through its Chinese subsidiary. $TOYO’s $261.0M of H1 2026 revenue came from four lines: → Third-party cells (56.5%): $147.6M of cells sold to outside module makers. → Related-party sales (22.0%): $57.5M of cells and modules sold to VSUN and its affiliates. → OEM services (14.2%): $37.2M of fees for cells made under one customer’s name, which $TOYO books net, counting only its fee as revenue. → Third-party modules (7.2%): $18.7M of Texas-made panels. Upstream, 70% of the polysilicon (the refined silicon that wafers are cut from) in $TOYO’s Ethiopian cells comes from a U.S. producer and the rest from OCI’s plant in Malaysia, Rhone Resch, $TOYO’s Chief Strategy Officer, said on the Q2 call. Resch also said its 2026 wafers come from non-China production, “including a designated facility in Indonesia.” $TOYO markets its supply chain as “non-FEOC,” free of any foreign entity of concern. U.S. buyers care, since the One Big Beautiful Bill Act (OBBBA) denies the main clean-power tax credit to projects with too much equipment from Prohibited Foreign Entities, a label aimed at Chinese suppliers.

Trading Post November 5, 2026

MercadoLibre’s logistics expansion cut margins by 60 basis points, while Mercado Pago grew credit 62% with 7% NPLs.

Intel: The CPU Comeback, Discounted

Intel's Q3 2026 results were hit by memory costs and an Israel shutdown, but enterprise AI inference could lift CPU demand as server…

Problems at Twist Bioscience (TWST)

Twist Bioscience’s $200 rebound followed an Anthropic-linked short squeeze, yet its DNA and protein services face broad competition while…

VSH Earnings Review

Vishay reported record Q3 revenue of $979 million, up 24%, and guided Q4 to about $990 million as rising backlog and margins contrast with…

Thursday Tailwinds #13: Nov 5, 2026

Florida marinas are 92% full as boat counts rise and permits constrain new supply, creating pricing power for existing operators such as…

eToro: Expanding Beyond Crypto

EToro trades at 2.89x EV/FCF with $304 million in trailing free cash flow, expanding from crypto into equities, commodities, currencies and…

SILC · long

Silicom LTD

Silicom (SILC) is pitched as inexpensive on its existing hardware business, with upside framed as a cheap call option.

Research Note

Nine Mile Metals is the subject of a research note.

Publications we follow

Every publication on this site, hand-picked.

129 publications

Newsletters 109

X accounts we follow

Every account on this site, hand-picked.

91 accounts

Investor letters we follow

359 firms

Sectors

Ordered by how often our editor reaches for them.