Sunday, November 1, 2026

Back to today

From X

@convequity 3 clicks

Convequity on X

Renting AI capacity protects Microsoft ($MSFT) if the AI buildout busts. It leaves Microsoft exposed if compute gets scarce. On Q2 2026 data, Microsoft's implicit and off-balance-sheet infrastructure commitments were 17 times its headline borrowings according to our AI Bubble Barometer. 1. Microsoft signs a multi-year deal to rent capacity. The partner, a specialist AI cloud such as $IREN or Nebius ($NBIS), typically raises the capital, builds the site, buys the chips and owns the assets. 2. Construction risk, the risk that the GPUs lose value, and most of the financing sit with that partner and its lenders. 3. Microsoft swaps capital spending today for rental payments later. Compare Alphabet ($GOOGL) at 10.2 times. Almost all of its off-book commitments are purchase orders, $811bn of them, for chips and equipment Google will own. We read Microsoft's choice as less than full confidence in the buildout, and Google's as the opposite. In our view the market underweights the risk that compute gets scarce. The partners own any capacity beyond their Microsoft contracts and can sell it to whoever pays more. Microsoft would then have to wait or pay up, as it owns fewer of the data centres and chips it uses. That means higher compute costs and tighter margins. On our reading, Microsoft wins if AI demand falls short, because its partners carry most of the asset risk. If compute gets scarce, Microsoft loses, because renting gives it less control of supply, and Google wins because it will own the chips it ordered.
@calvinblissett

Calvin Blissett on X

I feel AI can answer this better than me, kind of just giving more context on my other response. 54 sources used. Good read for $YB investors: Acquisition tactics repeatedly described as traps Chinese media, Black Cat (Heimao) complaint records and consumer-association notices describe a variant of the industry “magic-cube” (魔方) funnel that regulators have tried to curb since 2020–2021: - Pop-up and short-video/mini-program ads offering “first month ¥0.6 / ¥1,” “free claim,” or “¥1 for high coverage.” - Low-friction flows (SMS code, minimal identity checks) that default or bury auto-renewal, password-free WeChat/Alipay debit and upgrade from a one-month “experience” policy to a full-price monthly plan (often ¥90–800+). - Terms, health declarations and price jumps placed in small print or behind easy “next” clicks. Elderly users in lower-tier cities are disproportionately cited; children frequently discover multi-month or multi-year deductions only when reviewing bank or Alipay statements. - Documented examples include continuous deductions totaling ~RMB 1,500–2,200 over months and one multi-year case claimed at ~RMB 22,400; several policies sometimes opened on the same person. Refunds are often obtained after platform or media complaints, sometimes with partial fees withheld until complaints are withdrawn. A 2022 Shaanxi banking-insurance regulator fine (RMB 10,000 on the brokerage entity plus warning/fine on an executive) cited failure to conduct internet brokerage business as required. In 2024 the Gansu Consumers Association publicly criticized a related agency (Shouxin) for enrolling and debiting a consumer without authorization. SEC 20-F risk factors discuss dependence on third-party traffic channels, evolving CBIRC/NFRA online-insurance rules (including limits on intermediary marketing of personal lines), reputation harm from complaints, and the need to retain consumers, but do not specifically disclose the auto-debit or “first-month” funnel controversies. Not unique to YB The same pattern is widespread among Chinese internet insurance intermediaries. Waterdrop (水滴保) has drawn larger complaint volumes on the same platform (reports of 2,300–3,200+ entries) and has received larger fines (including a RMB 1 million penalty tied to “first-month zero” sales). Zhonghui Mutual and other carriers appear in parallel cases. Regulators have repeatedly flagged “first-month ¥1,” default upgrades and unperceived auto-debits; a 2026 multi-agency network-marketing rule (effective 30 September 2026) adds mandatory reading timers, bans one-sided promotion of the first-period premium, requires one-click popup closure, and prohibits default bundled sales and unauthorized automatic charges. Management statements and evidence of change - 2022 company responses (Sina Finance and others) defended “first-month ¥1” products as carrier-designed and regulator-filed, said many listed complaints predated tighter rules or were mis-attributed, and claimed near-100% resolution of its own cases with relatively low complaint counts versus peers. - August 2026 responses to Pengpai/The Paper on a specific ¥0.6-to-full-premium case asserted that backend logs showed completed login, disclosure reading, health notice and bank authorization, called the enrollment compliant, and issued a full refund “to resolve the dispute and improve service,” not as an admission of fault. The company stated that 2026 flows added mandatory reading countdowns and stronger confirmation steps; independent media tests of the WeChat mini-program confirmed longer forced-read sequences and the removal of some ¥0.6 experience creatives. Quantitative signals are mixed. Black Cat counts cited across 2022–early 2026 cluster around 850 rising to ~1,200 and then largely plateauing while new policies scaled into the tens of millions (one outlier report cites a much higher cumulative figure under broader search terms). Absolute complaint volume is therefore small relative to 30 million+ annual policies, yet the cases that surface are concentrated on elderly users, multi-month silent debits and difficult cancellation—exactly the pattern that generates media coverage. Many disputes end in refunds once escalated. High marketing intensity continues, consistent with reliance on low-friction paid traffic. How large is the problem? It is real and recurring, documented in complaint platforms, consumer-association notices, small regulatory actions and consistent anecdotal patterns, but it is amplified by media focus on elderly victims and by the emotional visibility of “unknown deductions.” It is an industry feature of high-CAC internet insurance distribution in China rather than a YB-only practice; peers show comparable or higher complaint intensity. Upcoming 2026 marketing rules directly target the tactics described. Filings treat the issue mainly as general regulatory and reputation risk rather than a disclosed operational failure.

Verra veerer

Verra Mobility’s Avis termination threatens 2027 earnings, while Government Solutions profit rose 25% and implied fourth-quarter guidance…

VITL · long

Vital Farms INC

Vital Farms is a long idea after bird flu-driven egg supply losses coincided with a structural shift.

The Data Center Bridge

Small- and mid-cap data center stocks trade as one despite sharply different funding bridges: 12 of 15 cover fixed costs on signed…

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

190 firms

Sectors

Ordered by how often our editor reaches for them.