CNBC Closing Bell · · 51 min
Anthropic IPO, AI Trade Reset and Market Sell-Off on Closing Bell
With Adam Parker, Alex Kantrowitz, Dario Amodei, Lo Toney, Rick Heitzmann
ChorusBring the market's conversation to your AI. Learn more
Summary
CNBC's Closing Bell episode centred on Anthropic's IPO and the wider AI trade reset. Lo Toney of Plexo Capital, an early Anthropic investor, argued its enterprise focus and compute efficiency are now being recognised, and that AI labs like Anthropic and OpenAI will move up the stack into applications and software M&A, comparing it to Oracle buying PeopleSoft. Alex Kantrowitz was bearish on Anthropic's return potential at a roughly $1 trillion-plus valuation, while arguing that AI tools are lowering the software barrier to AMD chips, weakening Nvidia's CUDA lock-in. Adam Parker was bullish on AMD and semiconductors, saw AI-driven volatility as a sentiment problem rather than an earnings problem, and argued that AI now drives most of the S&P 500's return variability. Oliver Renick flagged the IPO's supply pressure with the VIX near 30 and heavy Nvidia and Micron put buying as hedging rather than bottom-signalling.
Key takeaways
- Lo Toney (bullish on Anthropic): Anthropic's compute efficiency and enterprise focus made it an attractive early investment
- Adam Parker (bullish on Advanced Micro Devices): Buy AMD on the dip; its CPU-heavy agentic shift should drive share gains
- Alex Kantrowitz (bullish on Advanced Micro Devices): AI tools are removing the software barrier to AMD chips, so AMD gains share
- Oliver Renick (bearish on Anthropic): Anthropic's IPO added more supply than demand could absorb in a high-volatility market, so it was zero-sum.
- Rick Heitzmann on Enterprise AI adoption & AI software: AI money will slowly rotate from infrastructure into application-layer companies over the next six months.
What was said, by company and theme
18 views from 5 speakers. Every quote is checked against the episode's transcript (how).
Anthropic
Across the shows: 145 speakers in other episodes, 256 bullish and 117 bearish calls, past 30 days
Where everyone stands on Anthropic
Lo Toney · Plexo CapitalBullish
Anthropic's compute efficiency and enterprise focus made it an attractive early investment
Plexo Capital began looking at Anthropic around Q4 2023 and invested in its Series D at an $18 billion valuation. Lo Toney cites management's focus on safety, the choice to target the enterprise market, and the efficiency of its compute. He argues the focus on enterprise let Anthropic avoid expensive businesses like video, and that this efficiency is now being recognized.
“the efficiency that they were able to have with their compute. Their focus around the way that the models were created, the focus on enterprise allowed them to not have to do things that were very expensive like video”at 3:21
Alex Kantrowitz · Big TechnologyBearish
Anthropic's IPO is unlikely to deliver a Facebook-style 16x return from here
Alex Kantrowitz compares Anthropic to the Facebook IPO, when shares could be bought at about $20 to $40 and were worth around $600 years later, a 16x return. Anthropic is already a roughly $1 trillion-plus company at the IPO. For a similar return it would need to reach about $32 trillion, so he thinks the chances of that kind of return are low.
“Chances are low, you're gonna get that kind of return. It would have to go to 32,000,000,000,000.”at 7:56
Lo Toney · Plexo CapitalBullish
Anthropic can move up the stack into applications and earn premium frontier pricing
Lo Toney argues that if every company uses software, every company becomes an AI company because AI is eating software. Anthropic has the chance to expand the surface where its models compete and to earn premium pricing at the frontier, through specialization and application-building. He points to the competition with Claude Code and Cursor as the tension Anthropic will face, and sees the opportunity in using its models to go after applications.
“I think that's where Anthropic has a big opportunity to not only expand the surface where their models can compete to provide some premium pricing at the frontier, some specialization”at 9:50
Oliver Renick · CNBCBearish
Anthropic's IPO added more supply than demand could absorb in a high-volatility market, so it was zero-sum.
Renick compares Anthropic's debut with SpaceX's. SpaceX listed when the VIX was around 17, which helped ignite animal spirits. Anthropic listed with the VIX near 30, the highest since April, so the deal added supply the market could not handle and looked zero-sum. Anthropic itself swung almost 20% intraday. He expects its options, trading from Monday, to carry higher volatility than anything in the Nasdaq 100 or S&P 500.
“Instead of igniting animal spirits, today's IPO looks like it was a zero sum game that added more supply than this market's demand could handle.”at 42:06
Advanced Micro Devices
Across the shows: 21 speakers in other episodes, 28 bullish and 1 bearish calls, past 30 days
Where everyone stands on Advanced Micro Devices
Adam Parker · Trivariate ResearchBullish
Buy AMD on the dip; its CPU-heavy agentic shift should drive share gains
Adam Parker would buy AMD at these levels for a twelve to twenty-four month view, maybe less. He says the AI complex was built around eight GPUs for each CPU from Nvidia. New technologies will move the other way, with four CPUs for each GPU, and AMD will win that shift. He expects AMD's growth to be well above the semiconductor industry average. He thinks the Hunterbrook short report timing looks fishy.
“I like AMD. I'd buy the shares down here, for a twelve, eighteen, twenty four month view, maybe less. And the reason is the whole complex was kinda eight GPUs NVIDIA to one CPU. The new technologies are gonna be more the other way, like four CPUs to one GPU, and AMD's gonna win.”at 15:18
Alex Kantrowitz · Big TechnologyBullish
AI tools are removing the software barrier to AMD chips, so AMD gains share
Alex Kantrowitz argues AI makes it easier for companies like Anthropic and OpenAI to use AMD chips. Anthropic's roadshow slide showed that it typically took six to twelve months to get running on an AMD chip, but with Methos it took three days. He says the software barrier to alternatives to Nvidia is falling, so AI is how companies get around Nvidia's lock-in. He thinks the AMD bull case does not depend on the Jensen Huang news.
“I don't think you need the Jensen news to be encouraged about building or buying into AMD... AI is making it easier for the the companies like the Anthropics and OpenAI's of the world to use AMD chips.”at 15:48
NVIDIA
Across the shows: 177 speakers in other episodes, 184 bullish and 172 bearish calls, past 30 days
Where everyone stands on NVIDIA
Lo Toney · Plexo CapitalNeutral
Nvidia will lose some share to alternative chips but remain fine as a business
Lo Toney says having options beyond Nvidia is good procurement discipline for AI companies, and AMD will gain some share. Nvidia may lose a little but will still be fine as a business. He notes that Nvidia built a moat through the CUDA platform and its more than 4 million developers, but says demand for compute is so high that more supply from multiple players is a good thing.
“it's good procurement discipline for these companies to have options other than NVIDIA... NVIDIA is gonna maybe lose a little bit. It's still gonna be fine as a business.”at 17:43
AI chips & semiconductors
Where everyone stands on AI chips & semiconductors
Adam Parker · Trivariate Research
AI stocks and semis will go much higher; the sell-off is options-driven, not fundamental
Adam Parker says the AI trade has an elastic band on it because of zero-day options and levered ETFs. He points to Micron, which rose from $3.50 to around $7.15 since March. Its gain looks huge, but the volatility has spooked people. He thinks AI stocks and semis will be much higher before the end of this investment cycle, even though sentiment and momentum are currently bad.
“I think this thing has an elastic band on it because of the zero day options and the levered ETFs... it's just the vol has got people freaked out. I still think you're gonna see AI stocks, AI semis way higher”at 5:20
Adam Parker · Trivariate Research
Semis are not over-earning much; earnings are unlikely to collapse as prices imply
Adam Parker argues the market thinks semis are over-earning massively, but he disagrees. Semis grew about 2% above nominal GDP with lots of volatility for forty years. He expects them to grow 7-8% above nominal GDP at much higher gross margins. They may be over-earning a little, but he thinks earnings are unlikely to collapse the way some prices imply. He sees the sentiment as an odd game.
“semis grew two percent above nominal GDP with a lot of volatility for forty years. They're gonna grow seven to 8% above at much higher gross margin. They could be over earning a little, but I think it's unlikely that earnings collapse”at 22:04
US stock market outlook
Where everyone stands on US stock market outlook
Adam Parker · Trivariate Research
AI now drives the market: 18% of S&P 500 firms have AI revenue, 45% of market cap
Adam Parker says 18% of S&P 500 companies have meaningful AI revenue. They make up 45% of market cap and explain 92% of the variability of returns. So he argues there is nothing outside AI in the public companies that matters for returns right now, which is why the AI factor is so important to the broad market.
“18% of S and P five companies have meaningful AI revenue. It's 45% of the market cap. It's 92% of the variability of returns.”at 11:06
AI infrastructure & data centers
Where everyone stands on AI infrastructure & data centers
Adam Parker · Trivariate Research
Data centers can't be switched off in two to three years; electricity-bill backlash will fade
Adam Parker says data centers take ten to fifteen years to build, so there is almost no way to turn them off in the next two to three years. He concedes AI demand could push some power bills up in certain nodes, but other nodes make economic sense for data centers. He thinks the political issue will die down slowly, and he does not expect it to be a lightning-rod issue in 2028.
“data centers take ten, fifteen years. There's almost no way you can turn things off in the next two to three years.”at 12:07
US consumer
Where everyone stands on US consumer
Adam Parker · Trivariate Research
The consumer is stronger than expected; the K-shaped economy is a balance-sheet issue
Adam Parker says the consumer has been much stronger than people thought. The K-shaped economy is more about balance sheets than income statements, because people can get jobs if they want them and wages are pretty good. He cites low 90-day credit card delinquencies as evidence that the economy is holding up.
“the consumer's been way stronger than people thought... this k shaped economy is much more of a balance sheet thing than an income statement thing. Meaning, people can get jobs if they want them.”at 13:03
US–China AI & chip race
Where everyone stands on US–China AI & chip race
Alex Kantrowitz · Big Technology
Stricter chip export controls to China and nearby countries seem inevitable, hitting AMD and Nvidia
Alex Kantrowitz says the world is about to see stricter export controls on selling chips to China and nearby countries, whether or not the AMD short report is accurate. He expects this could hurt AMD and Nvidia's businesses depending on how things play out. He argues Anthropic and OpenAI benefit because they lose competition for compute, so more compute will be steered to them, which makes them better businesses.
“we're about to enter a world where there's gonna be stricter export controls to... About selling chips to China and countries nearby. It seems inevitable to me.”at 16:49
Cybersecurity
Where everyone stands on Cybersecurity
Adam Parker · Trivariate Research
Large-cap cyber stocks trade at 50–90x earnings, pricing in more than the franchise is worth
Adam Parker says the large-cap cyber names like CrowdStrike and Palo Alto trade at very high multiples. Even if they are unassailable and nobody gets rid of them, investors are paying 50 to 90 times earnings two to three years out. He sees the cyber franchises as unassailable but the price as demanding, and the truth is in between. So he would rather own semis than cyber.
“the larger cap ones traded such massive multiples that even if they are unassailable and nobody ever gets rid of CrowdStrike or Palo Alto, you're paying somewhere between 50 to 90 times earnings”at 20:59
AI pricing & unit economics
Where everyone stands on AI pricing & unit economics
Adam Parker · Trivariate Research
Customers won't lock into long single-vendor AI deals, so pricing competition will intensify
Adam Parker argues that no business wants to commit to a single AI vendor for twenty years with no exit. Customers will choose with a budget and test different providers. He expects more pricing competition among AI providers as customers mix and match.
“Every business doesn't wanna impregnate an oracle for twenty years that can never get out. They're gonna let their customers choose with a budget. And I think you're gonna see more pricing competition.”at 22:04
Software & SaaS
Where everyone stands on Software & SaaS
Lo Toney · Plexo Capital
Anthropic could buy software companies with customer bases and lock-in, as in the PeopleSoft era
Anthropic now has more currency for acquisitions. Lo Toney would look at software companies with strong customer bases and some lock-in, where Anthropic can use its models. He argues Anthropic and OpenAI will move up the stack to get closer to customers, with better margins, lock-in and data for improving the product. He compares this to Oracle buying PeopleSoft, for the AI era.
“I would look to some of those software companies where Anthropic can use the model and then go to these software companies that have a great customer base and a little bit of lock in.”at 23:48
Market sentiment, positioning & volatility
Where everyone stands on Market sentiment, positioning & volatility
Oliver Renick · CNBC
No big-money put selling signalling a market bottom; market makers are widening volatility ranges instead.
Renick says he looks for large put sales near the money in the S&P 500 as a sign of a possible bottom, and does not see them today. Instead, market makers are widening their ranges for the VIX and VOL. Volume in SPY and QQQ ran 50% above average. The biggest single trades were large purchases of in-the-money puts on Nvidia ($240 million) and Micron ($180 million), which is hedging or bearish positioning rather than bottom-signalling buying.
“I'm looking for big money put sales near the money strikes in the S and P to signal a possible bottom, but I don't see them in today's action.”at 42:47
Enterprise AI adoption & AI software
Where everyone stands on Enterprise AI adoption & AI software
Rick Heitzmann · FirstMark Capital
AI money will slowly rotate from infrastructure into application-layer companies over the next six months.
Heitzman argues application-layer companies are driving most of the AI value, including on the model side. He expects a slow rotation out of AI infrastructure into applications over the next six months. He cites demand for AI-native companies, including vertical applications in legal and healthcare, which he says have superior financial results to companies weighed down by last-generation technology. He also expects the Anthropic IPO to be the start of a wave of new IPOs.
“The application layer companies are driving most of the value even on the model side or applications themselves, and I think you're gonna see a slow rotation out of infrastructure into applications over the next six months.”at 44:37
Across the shows
What other episodes Chorus covered said about the same companies in the 30 days up to this one.
- NVIDIA177 speakers · 184 bullish, 172 bearish calls
- Anthropic145 speakers · 256 bullish, 117 bearish calls
- Advanced Micro Devices21 speakers · 28 bullish, 1 bearish calls
Topics
- AI concentration in the index
- AI pricing competition
- AI semis sell-off and volatility
- AMD share gains from CPU-heavy AI workloads
- AMD share gains from software portability
- Anthropic enterprise strategy and compute efficiency
- Anthropic moving into applications
- IPO return potential
- IPO supply and volatility
- Nvidia competition from alternative chips
- US consumer balance sheets
- chip export controls
- compute spending commitments
- cyber valuations
- data center political backlash
- infrastructure to application rotation
- options positioning and market bottoms
- semiconductor earnings sustainability
- software M&A by AI labs
More from CNBC Closing Bell
- Anthropic’s Record IPO, AI Selloff and Market Rout · November 20, 2026
- Nvidia Slumps, Anthropic IPO Looms and Markets Test Key Levels · November 19, 2026
- Anthropic IPO, NVIDIA CEO Search and AI Trade · November 18, 2026
- NVIDIA Earnings Beat, SEC Probe and AI Trade Reset · November 17, 2026
- NVIDIA Earnings Stakes, AI Infrastructure and Tech Outages · November 17, 2026
Ask your AI about this episode
Chorus gives Claude, ChatGPT and other assistants every view from this episode and every other show Chorus follows. Ask who agrees, who changed their mind, and what else was said about the same companies.
Chorus reports what speakers said. It isn't investment advice. How Chorus works · About Chorus