-+ 0.00%
-+ 0.00%
-+ 0.00%

Wall Street tech veteran Niles is optimistic that Anthropic (ANT.US) will win the AI smart price war! But it warned that the stock price might break

Zhitongcaijing·11/20/2026 11:33:19
Listen to the news

Zhitong Finance App learned that as of the closing of US stocks on November 19, EST, the global AI application leader, Claude developer Anthropic (ANT.US), issued at $83 per share. It opened at $111.50 on the NASDAQ on Thursday, hitting an intraday high of $115.10, and finally settled at $97.17. The allotted investors earned 17.07% on the first day, while investors who bought at the opening price and held until the close lost 12.85%. After this record US stock IPO, senior Wall Street fund manager Dan Niles is optimistic that Anthropic will win the model price war with its corporate customer base, but he has reservations about whether the market can maintain the $83 IPO price.

On the same day, the Dow fell 1,101 points, or 2.19%, while the S&P 500 index fell 2.02%. The low closing for six consecutive trading days showed a weak trend, and the NASDAQ fell 2.25%. The Philadelphia Semiconductor Index fell 2.41% to 9,419.72 points. This week's cumulative decline as of Thursday was 7.16%, retreating 35.63% from the historical closing high on June 22. This global semiconductor benchmark index is already mired in a technical bear market. American memory chip giant Micron Technology, one of the big winners of the AI computing power frenzy, closed down for six consecutive trading days, a sharp retracement of 43.11% from the highest point in history at the close of June 25.

Facing this wave of AI trading sell-off, Michael Berry, who has the title of “big short,” abandoned his subscription to Anthropic, and while increasing bearish on US technology stocks, bucked the trend and increased his holdings of undervalued Chinese technology assets such as JD, Tencent, and BYD. Dan Niles, on the other hand, said that Anthropic has served companies willing to pay from the beginning, which is more beneficial than seeking AI subscription services from ordinary C-end consumers who are used to free services, and emphasized that Anthropic is more likely than OpenAI to eventually win this round of global AI intelligence competitions; although it is optimistic about its corporate customer base, it also questions the short-term stock price at a time when the model price is reduced and computing power promises are high — that is, whether the stock price can maintain the $83 issue price during the 30-day green shoe period.

Dan Niles (Dan Niles) is the founder and portfolio manager of Niles Investment Management and has been researching and investing in technology stocks on Wall Street for over 35 years. He has a master's degree in electrical engineering from Stanford University. In his early years, he worked as an engineer at Digital Equipment Company (DEC). He joined the investment bank Robertson Stephens in 1990, switched from mergers and acquisitions to stock research, joined Lehman Brothers in 2000, responsible for semiconductor and computer hardware research, and founded the Satori Foundation in 2004.

From 1999 to 2003, he was selected for the “Institutional Investor” National Research Team's semiconductor group for five consecutive years. His views have important reference value in judging the technology industry cycle. What is particularly noteworthy is how he implemented major technological changes into corporate pricing rights, profit margins, and stock valuations.

Dan Niles is optimistic that Anthropic will survive the AI smart price war, but he is cautious about short-term stock prices

Anthropic's initial public offering performance exceeded expectations, but Dan Niles warned investors to keep an eye out for valuation risks, a slowdown in AI revenue growth, and increasingly fierce price competition.

Judging from the book performance, Anthropic's debut was remarkable. But Dan Niles of Niles Investment Management said that more severe tests are yet to come. The portfolio manager believes that as the underlying model of intelligent AI gradually becomes homogenized, Anthropic is more likely to survive the competition than OpenAI. At the same time, price competition is squeezing profit margins, and stock prices will also be tested by a 30-day green shoe period.

Anthropic rose 17% on the first day of its listing on the NASDAQ on Thursday, and showed a decline of more than 6% before Friday's market.

The issue price and number of shares issued have both been raised, and the amount of capital raised has even exceeded the original plan. Niles said in a recent post on social media platform X that Anthropic's current IPO has achieved its due goals. The company raised the number of shares issued and the issue price at the same time, raising more capital than planned, and rose 17% at the close of the first day of listing. For such a large-scale release, Niles usually sees a 10% to 15% increase on the first day as a success, and this performance has exceeded this threshold.

However, Niles pointed out that when the stock opened, it was 34% higher than the issue price. The increase reached 39% within a few minutes, then continued to fall until the close. In his view, a healthy listing debut should be paid above the opening price, which means that investors who failed to get a share are more willing to go after it than the original holders are willing to take advantage of the listing to cash out. Anthropic failed to close above the opening price on the same day. Here is the explanation of the intention to buy and sell in the report according to Niles's original post.

“Whether this IPO is actually a success or not, the real test is the green shoe period over the next 30 days. The AI-related stock market is in a difficult situation. Can it stay above the $83 issue price? I'm not that sure.”

If you could only choose a company that is a superleader in AI models that have survived for a long time, Niles would choose Anthropic

There is reason for Niles's optimism about Anthropic's long-term prospects. He said Anthropic has been seeking corporate customers willing to pay since day one, while OpenAI initially targeted consumers who are accustomed to using free services. “Anthropic is the first listed company that investors can directly invest in with a big language model as its core business. It has scarce value and is also likely to be quickly included in relevant indices.”

Referring to the negative factors, Niles said that the month-on-month growth rate of OpenAI and Anthropic's combined revenue slowed from about 75% in the second quarter to about 50% in the third quarter. According to his judgment, OpenAI's price cut boosted its revenue growth rate from about 17% to 50% month-on-month, while reducing Anthropic's growth rate from 144% to 47%. Anthropic's adjusted operating margin fell from 19% to 15%, while OpenAI's loss increased from about US$10.1 billion to US$10.9 billion.

“Price competition is intensifying, the combined revenue growth rate of the two leading big language model companies is slowing down, operating margins are also declining, and they are already undertaking huge capital expenditure commitments. However, the market opportunities are also huge, and intelligent AI is still in its infancy.”

“If I could only choose a company that can survive after homogenizing the underlying model, I would choose Anthropic. But as far as the AI sector as a whole is concerned, it's not one of my most promising investment targets recently.”

What do retail investors think of Anthropic? According to retail weather vane data compiled by the Stocktwits platform, retail investors' sentiment about the stock is still “extremely bullish.” News about the stock surged 748% on the Stocktwits platform over the past 24 hours.

From a fair value of $92 to a target price of $155, what does Anthropic rely on to support its trillion-dollar valuation after the global model is drastically reduced in price?

Morningstar gave a fair value of $92, while Melius is optimistic that the stock price will reach $155 within 24 months. The valuation difference between the two falls on a specific question: when the model price reduction brings in more corporate adoption, can Anthropic rely on increased expenses and improved reasoning efficiency to cover computing power costs and reverse the trend of falling adjusted operating margins from 19% to 15%?

The popularity of AI agents that focus on efficient proxy workflows can simultaneously drive up inference usage and reduce model unit prices. Enterprise organizations, on the other hand, can hand over standardized processes such as document extraction to low-cost models on a large scale, leaving complicated programming and expensive steps of rework after errors to cutting-edge models. What the buyer actually compares is the total cost of successfully completing a task, which includes not only the quotation for each token (token), but also the number of inference rounds and retries required for the model to complete the task.

Anthropic revealed that through price adjustments and improved execution efficiency, Fable 5.1 reduced the cost of typical customer tasks by about 25% compared to the previous generation, and reduced complex programming and highly automated agent tasks by up to 45%.

For shareholders, profit depends on whether the computing power required to complete similar tasks can be reduced simultaneously, and whether lower prices can be exchanged for sufficient amounts to be paid. 79% of the company's revenue for the first nine months was confirmed by usage, and procurement and leasing commitments reached US$909 billion, almost all of which involved computing power and infrastructure.

The prospectus has given the financial expression of this tension. Anthropic's third-quarter revenue increased about 47% month-on-month to US$16.956 billion. Adjusted operating profit increased from $2,235 billion to US$2,463 million, but the adjusted operating margin fell from 19% to 15%. According to US accounting standards, the company switched from operating profit of 974 million US dollars in the previous quarter to a loss of 567 million US dollars, and included 1.96 billion US dollars in non-cash charitable donation expenses for the quarter.

Price competition is a variable that Niles is concerned about, and the company's profit also depends on computing power procurement costs and cost caliber. The chip market also continued to ferment, and the Democratic Party's victory led to a slowdown in data center construction expectations — the well-known market research agency TrendForce predicted on November 19 that although demand for high-performance storage in North America is still strong, Chinese cloud computing vendors and some North American cloud companies will cut enterprise solid-state drive orders. Coupled with the expansion of flash memory bit supply, the NAND contract price may drop sequentially from the first quarter of 2027.

Anthropic's third-quarter revenue increased 47% month-on-month, but the adjusted operating margin fell from 19% to 15%. Therefore, Wall Street analysts' valuation differences focused on how much profit corporate customers could bring to Anthropic as model price competition intensifies. Morningstar's estimated fair value per share was $92, lower than the first-day closing price of $97.17. More optimistic, Melius predicts that the company's revenue in 2029 is expected to exceed 400 billion US dollars and give a 24-month target price of 155 US dollars.

The difference between the latest valuation given by Wall Street and the target price is large enough to change buyers' profit and loss judgments on the first day. Morningstar estimates a fair value of $92 per share, which is higher than the issue price and lower than the closing price on the first day. Citizens JMP gave a target price of 125 dollars, Needham gave 150 dollars, and Melius gave a clear 24-month target price of 155 dollars, which is about 29%, 54%, and 60% higher than $97.17, respectively.

Both Melius and Citizens expect the company's revenue to exceed $400 billion in 2029. There is also a clear date for short-term demand: MSCI has announced the inclusion of Anthropic in the large-cap sector of the Global Standard Index, effective December 1. At the same time, underwriters have the right to purchase up to 147 million additional shares at the issue price within 30 days, and can also make up for overallotments in the market. $83 is Niles's actual purchase price, and the green shoe mechanism does not guarantee that the stock price will maintain it.