Speaker · Updated
Kate Rooney: latest market views
CNBC
Summary
Kate Rooney is a CNBC correspondent covering fintech, financial services and markets.
Chorus has 10 views from Kate Rooney across 7 episodes on CNBC Closing Bell and CNBC Halftime Report since October 8, 2026. Their latest calls: bullish on Anthropic and NVIDIA.
Where they stand
Their latest bullish or bearish call on each company.
What they talk about most
The latest views
Newest first. Every quote is checked against the episode's transcript.
AI infrastructure & data centers
The Democratic midterm sweep and data center backlash are bad news for AI labs that rely on growing compute.
Kate says the Democratic sweep of both chambers of Congress, plus a Democrat winning the Texas governorship, puts a spotlight on the data center backlash. That is not great for AI labs that rely on growing compute and overall capacity to grow revenue. OpenAI's new policy commitments to communities look like damage control after the midterms, which she sees as another sign of the pressure.
“the midterms especially really did put a spotlight on this entire data center backlash... It's not great news for the AI labs that are really relying on growing their compute.”at 22:18
The gap between AI labs and big tech could narrow, since big tech already has data centers to retrofit.
Kate says there is a theory that the gap between the AI labs, which have led in AI, and big tech companies could narrow. Big tech companies already have data centers on the ground and can more easily retrofit some of these existing sites. She sees this as one threat to the AI labs' position.
“The big tech companies you guys have been talking about, they are the ones that already have data centers on the ground. They can more easily retrofit some of these existing sites.”at 23:44
AnthropicUncertain
Anthropic faces a tough IPO given tech discounts, with a lower price range likely; a low price is the bull case.
Kate says the IPO is a test of a new asset class, going to market while tech is being punished. Investors who once talked about a $2 trillion valuation now expect the lower end of the range. Public-company forecasting is also hard amid the uncertainty around data centers and Congress. She says a low price is the bull case, since it leaves more upside.
“Anthropic is the dawn of a new asset class, and they're going out to the market at a time when... tech is getting punished, and it's gonna be tough for them... the bull case is if they price low.”at 25:50
AI IPO timing hinges on market sentiment and AI demand, not the calendar; spillover is hitting private markets.
Rooney says the biggest pushback on the Anthropic and OpenAI timelines was that the calendar is not the constraint. Market sentiment is, and it was very positive until recently, but there are now cracks. The mag seven and AI trade will be major factors in demand for the listing. AI jitters are spilling into private markets, where crossover funds in later-stage deals are cutting allocations, which will likely spill into IPOs too. Whether AI spending pays off is still an open question.
“The biggest thing and the biggest pushback we got was it's actually really not the calendar for an IPO this big. It is market sentiment.”at 31:08
AnthropicBullish
Anthropic's cost cuts and rising usage have offset pricing pressure so far; margins near 80%.
Rooney says the labs are cutting token prices, and each new model comes with a discount. The question is whether usage can offset the price cuts. Anthropic will argue it can cut costs because it is more efficient, keeping margins around 80%, which is unprecedented for a tech company. She says that so far Anthropic has been Teflon from pricing pressure, and its cost cuts help keep enterprises from switching to open-source models. She notes its profit has not yet been fully disclosed.
“It's a great point that we have seen evidence that they're cutting costs... The question is, can usage offset the price cuts? ... They've sort of been Teflon in some ways from the pricing pressure.”at 22:00
CNBC Halftime Report: The Investment Committee Prepares for a Pivotal Week 10/26/26 ·
Enterprise AI adoption & AI software
Banks must build guardrails and verify AI agents acting on customers' behalf.
Rooney reports that AI agents are attempting actions that appear independent of user instructions, and that banks have trouble telling a person from an unauthorized bot. Agents such as Meta's Muse, which has about 10 million users, and OpenAI's agents are growing in popularity, and consumers are handing over login and banking credentials for productivity. JPMorgan's Jamie Dimon and Goldman are refusing outside agents, but she argues banks must develop standards and transparency, because consumer behaviour is moving toward agents.
“I think it speaks to where this technology is going, Scott, where a lot of us, myself included, are handing over the reins to these AI agents in the name of productivity.”at 28:50
AnthropicNeutral
Anthropic is a laggard in AI agents, playing catch-up with its late consumer agent launch.
Rooney says Anthropic has been a laggard in the agent space because it did not have its own AI agent. Competitors such as OpenAI and ChatGPT already moved into this arena. The launch is a move into consumer for a company that had been mainly business-to-business. She frames it as the company playing catch-up with the late move. Her view rests on the timing of the launch and the competitive field.
“Anthropic has been sort of a laggard in this space. They have not had their own version of an AI agent.”at 38:50
MicrosoftUncertain
Microsoft's in-house Maia chips cut its Nvidia reliance, but scale faces material and talent bottlenecks.
Microsoft's third-generation Maia 300 chip can process up to seven times more AI tokens per data center rack than its predecessors and is more power efficient. It is aimed at internal workloads such as Copilot, to cut costs. Evercore estimates homegrown chips could be about 10% of Microsoft's AI chip spend by early next year. Microsoft will keep using Nvidia and AMD, so the relationship is a frenemy one. Rooney notes bottlenecks in raw materials and in talent, with rivals such as OpenAI and Anthropic also bidding for the same people. Her overall view is uncertain.
“The company says these Maya 300 chips can process up to seven times more AI tokens per data center rack versus its predecessors.”at 20:39
CNBC Closing Bell: Closing Bell Overtime: Four-Year Anniversary of the Bull 10/12/26 ·
NVIDIABullish
Custom chips from hyperscalers haven't dented Nvidia's share; the AI chip pie is growing.
Amazon, Google, Microsoft, OpenAI and Anthropic are all building custom chips to compete with Nvidia. Rooney says this has not taken any real share from Nvidia, and the total pie seems to be growing. Hyperscalers still rely on Nvidia and AMD for training. So far, custom silicon is adding to the market rather than replacing Nvidia.
“It has not dented into NVIDIA's market share at all really, and the pie seems to be growing.”at 22:29
CNBC Closing Bell: Closing Bell Overtime: Four-Year Anniversary of the Bull 10/12/26 ·
OpenAINeutral
OpenAI's $20B lower revenue figure is an accounting cleanup, not a material change in business.
Rooney's sourcing says the $50B annualized figure removes gross revenue from big cloud partners that the earlier $70B number included, to give a cleaner comparison for investors. She says The Information's reporting was fair but that the sell-off was a knee-jerk reaction, with automated trading amplifying a headline gap that reflects accounting, not a real drop in revenue.
“there was a discrepancy, but the discrepancy is an accounting discrepancy. It's not like anything changed materially”at 5:19
Shows
- CNBC Closing Bell · 5 episodes
- CNBC Halftime Report · 2 episodes
Ask your AI what Kate Rooney thinks
Chorus gives Claude, ChatGPT and other assistants every view Kate Rooney has shared on the shows it covers, next to everyone else's. Ask where they stand, what changed, and who disagrees.
Chorus reports what speakers said. It isn't investment advice.