Microsoft and Google both reported, and Ram works through the numbers from the driver's seat, starting with Anthropic: the market kept waiting for a shoe to drop, and he can't find one. His read is that semis are cheap and back, and that the hyperscalers' spending flows straight to the CapEx receivers.
Microsoft and Google both reported, and Ram works through the numbers from the driver's seat, starting with Anthropic: the market kept waiting for a shoe to drop, and he can't find one. His read is that semis are cheap and back, and that the hyperscalers' spending flows straight to the CapEx receivers.
On Microsoft, he calls the OpenAI deal incredible: a revenue share, the IP and an early stake, while OpenAI carries the CapEx. He sees Microsoft's hold on financial services and other regulated sectors as durable, even as GitHub and LinkedIn come under pressure, and he hears from venture investors that the big banks see AI coming.
On Google, he still sees a long-term winner that is spending on data centers now to earn the return later. He weighs the two companies' backlog against their CapEx, asks how much of it rests on the frontier labs and on venture-funded demand, and closes on why he thinks agents put us back in 1995, with adoption still early.
Chapters
(00:00) Anthropic, and the shoe that never dropped
(01:02) Semis are cheap and the midterm risk is priced in
(01:35) Microsoft's sales next to Anthropic's growth
(02:23) How semiconductors stopped being a commodity
(03:21) Microsoft's revenue capture from OpenAI
(04:19) CapEx receivers over the hyperscalers
(04:38) Grok on Microsoft's quarter
(05:49) What Microsoft earns from OpenAI
(07:08) Microsoft's stake in OpenAI
(08:27) Microsoft as one of the shrewdest investors in AI
(09:29) Private and public investors read revenue differently
(09:55) How the OpenAI deal came about
(10:59) Chipmakers financing their customers
(11:15) OpenAI broke the contract and Microsoft took the edge
(12:30) Why OpenAI's S-1 won't look like Anthropic's
(13:04) Microsoft's lock on financial services
(13:49) GitHub and LinkedIn under pressure
(14:17) The big banks see AI coming
(15:45) Grok on Microsoft's earnings call
(17:51) Free cash flow and credit to Satya Nadella
(18:37) Did ego cost Google the early deals?
(19:01) Google is still a long-term winner
(19:48) Grok on Alphabet's quarter
(22:07) Backlog against CapEx
(23:03) How much of the backlog is the frontier labs
(23:59) Most of the growth comes from real businesses
(24:30) Still early in AI adoption
(25:08) The CapEx dilemma for CEOs and boards
(25:53) Venture-funded demand is less durable
(26:19) A venture round that has heads exploding
(28:06) All of venture is AI now
(28:45) If Google gets cheap, it's a time-horizon call
(29:15) Meta has no cloud business to monetize
(30:04) Google and Microsoft can afford to be slow
(30:54) Don't underestimate Google
(31:12) Why agents put us back in 1995
(31:35) A processed-foods CEO meets AI
(32:36) Change management is extremely hard
(33:11) Now is the time to build
Lumida Wealth Management LLC is a registered investment adviser. SEC registration does not constitute endorsement by the U.S. Securities and Exchange Commission and does not indicate a particular level of skill or ability.
References to companies, industry themes, market statistics, prior investment activity, or case studies are included for context and should not be construed as a recommendation or endorsement.
Prospective investors should conduct their own diligence and consult their own legal, tax, and financial advisers as appropriate.