$UPS CEO Carol Tome told Reuters that data center logistics represents a $15 billion annual revenue opportunity.
The AI buildout gives UPS another route to earn more from fewer, higher-value shipments, from flying in chips to delivering equipment to project sites.
Replacement parts could extend demand beyond construction. UPS hasn't disclosed data center sales, so the $15 billion sizes the opportunity, not revenue already won.
🚨 GM (General Motors) is leaving the NYSE for Nasdaq after 110 years.
The switch is expected on November 17.
The interesting part? GM says the move makes it eligible for potential Nasdaq-100 inclusion.
Coca-Cola and Procter & Gamble warned of more price hikes next year as energy, raw material and freight costs rise.
Consumers are already pulling back from name-brand groceries, leaving staples companies to balance cost pass-through against promotions. P&G now expects a $1.4 billion profit hit from input costs, while $KO fell 1.34% and $PEP dropped 2.02% today.
$JPM's 55ip is working with Anthropic to bring tax-smart technology into Claude for financial advisors, per Reuters.
The distribution angle is important.
Specialized wealth tools could reach advisors inside their AI workflow. Invesco also announced work with Anthropic today on custom portfolios and practice management, putting Claude in the mix for more than tax tools.
The US has offered $CEG $5 billion in financing to support nuclear power projects, Bloomberg reports
The offer could help Constellation fund projects with large upfront costs. For shareholders, the key is whether it provides cheaper funding than the company could otherwise obtain—and which projects it would finance.
SoftBank Group is considering a $50 billion investment project in Malaysia, likely targeting AI and data centers, per Yomiuri Shimbun.
Its Oct. 8 framework agreement with Petros and Grab gives the Sarawak AI infrastructure push a local foothold. The funding uncertainty matters: how much SoftBank would actually invest and how it would raise the money remain unclear.
$AMD will invest $1 billion in $META's equity offering as the companies expand their AI infrastructure partnership, per Reuters.
AMD's Helios racks will power Meta's frontier models and services. The chip supplier is helping fund its customer's AI buildout, with its $1 billion check equal to about 2.7% of Meta's $37.5 billion financing.
The strongest point is that Anthropic already has access to major enterprises, and there is so much room to sell much more inside each one.
Gavin is highlighting this combination (broad adoption, but spending that remains small relative to customers’ size)
There’s an even more useful detail in the S-1: its top 10 Fortune 500 customers spend over eight times the average share of expenses on Anthropic. That suggests deeper adoption is possible within existing accounts.
But also note that total expenses are a very generous denominator. It includes things like merchandise, raw materials and shipping, not all spending that AI can realistically replace.
Wall Street firms have committed financing for a $63 billion deal to buy Broadcom AI chips and networking equipment and lease them to Anthropic. Deliveries are expected to begin in 2027.
A separate company will purchase and own the equipment, with Anthropic paying to use it. Bank of America, Citi and Morgan Stanley are arranging the senior debt, which gets repaid first. Blackstone is providing the committed junior debt, which ranks behind it.
Broadcom is also providing support tied to the equipment’s future resale value, covering an amount equivalent to the senior financing. That means Broadcom is helping protect the financing as well as supplying the hardware.
The significance: this gives Anthropic a way to secure computing equipment without paying the entire purchase price upfront, and helps Broadcom finance customer purchases. The announcement does not disclose the lease payments or the detailed conditions of Broadcom’s support.
Nokia’s Q3 earnings are out.
There is stronger evidence that AI networking demand is lifting sales and profits, but cash generation is lagging. All figures below are in euros.
• Revenue €5.15B, up 10% reported and 9% at constant currency. Comparable operating profit rose 35% to €605M, with margin improving to 11.7% from 9.5%. Comparable diluted EPS was €0.09 versus €0.06.
• AI/cloud customer sales grew 123%, with €1.5B in new orders. IP Networks grew 43% and Optical Networks 18%, both at constant currency. Earlier customer wins are now producing revenue.
• The biggest forward signal: Management expects combined Optical and IP sales growth of 25–30% in 2027. That gives the AI networking opportunity more substance beyond this quarter.
• 2026 operating-profit guidance remains €2.1–2.6B, with management now expecting to finish above the midpoint.
• Cash outlook weakened.
Expected free-cash-flow conversion fell to 35–55% of comparable operating profit, reflecting an additional €400M of working capital to secure supply, plus accelerated restructuring.
My analysis:
The strongest development is that Nokia’s AI exposure is showing up in shipments and operating profit. Optical and IP networking are the clearest near-term earnings drivers.
But the growth has costs: Network Infrastructure’s gross margin fell to 41.2% from 42.7%, and reported operating profit fell 22%, reflecting restructuring and integration charges. Mobile Infrastructure also benefited partly from deliveries brought forward from Q4.
I’m constructive on the business momentum. The next proof is whether Nokia can deliver its 2027 growth outlook while improving margins and turning the inventory investment back into cash.
JUST IN: The US and Canada are discussing relief from tariffs of up to 50% on Canadian steel and 25% on autos, per Reuters.
Automakers could benefit on two fronts: cheaper steel inputs and lower duties on finished vehicles. But a source cautioned that any deal could be narrow. The scope matters: relief on autos alone would leave the steel tariff of up to 50% in place.
$MSFT is offering buyouts to around 60,000 employees, per Business Insider.
Payroll savings depend on how many accept. The offer puts labor costs in focus as Microsoft and its tech peers signal no slowdown in AI spending with their combined spending set to exceed $730 billion this year.
Anthropic disclosed roughly 900% net revenue retention. Eligible customers spent about nine times what the same cohort spent a year earlier.
So why worry about demand?
Because Anthropic has to buy capacity before customers decide how much they’ll use.
Its S-1 disclosed roughly 900% net revenue retention at September 2026. That measures spending growth from an existing customer cohort. It doesn’t mean those customers have committed to spend that much next year.
Roughly 80% of first-half revenue was usage-based. Anthropic, meanwhile, has signed long-term compute agreements with minimum spending obligations.
Here’s how that can go wrong: imagine a company reserves enough computing capacity for customer spending to triple. Spending only doubles. Revenue still looks fantastic, but some of the capacity it committed to may sit unused, while minimum bills remain payable.
Growth can stay strong and still fall short of the spending plan.
If demand fills the capacity at attractive prices, securing it early pays off. If it doesn’t, lower utilization can squeeze margins and increase financing needs.
For IPO investors, I’d watch how quickly new capacity turns into paying usage, and how much cash that usage generates after compute costs.















