Musk Talks Down Neocloud Rivals? Dan Niles Says Microsoft Will Be the Ultimate Winner in Enterprise AI

1. Anthropic Formally Pursues Nasdaq IPO

Artificial intelligence company Anthropic filed a registration statement with the US Securities and Exchange Commission on Oct. 26 for an initial public offering. The company plans to list on Nasdaq under the ticker ANT.

Anthropic confidentially submitted its IPO filing in June. It has now disclosed financial information and other details to public investors. Goldman Sachs and Morgan Stanley are the lead underwriters, with JPMorgan, Citigroup and Bank of America also participating in the underwriting syndicate.

Anthropic raised funding in May at a $96.5 billion valuation. That was about 2.5 times the $38 billion valuation it received in February.

That valuation topped OpenAI’s $85.2 billion.

Anthropic’s rapid growth has been driven in part by Claude Code, its AI coding tool. Demand for enterprise AI services has expanded as companies use AI more aggressively in software development and workflow automation.

The listing comes as worries about AI investment are mounting. Shares of AI infrastructure companies such as CoreWeave and Nebius have fallen more than 30% in October as funding burdens and credit risks increased.

Anthropic’s IPO stands to be a key test of investor confidence in AI companies and the capital market’s willingness to keep financing them.

The offer price, deal size and listing schedule have not been set.

2. Musk Says 200 Neocloud Companies May Collapse

Shares of neocloud companies fell sharply after Elon Musk warned that more than 200 of them could collapse.

In premarket trading on Oct. 26, CoreWeave, Nebius and Iren were each down more than 4%.

Writing on X a day earlier, Musk said the bond market had effectively stopped financing companies that have not generated enough revenue outside AI data centers.

He added that more than 200 neocloud companies would go through large-scale failures and industry consolidation, with few survivors.

Neocloud companies buy large quantities of Nvidia graphics processing units and lease computing resources to AI companies. The business requires heavy spending on GPUs and data-center construction, leaving these firms highly dependent on bond and loan markets.

Funding costs for the sector have been soaring as investors grow more wary of AI infrastructure spending.

One notable point is that SpaceX, which Musk leads, also competes with neocloud providers.

SpaceX operates its own data centers and sells computing resources to outside customers. Musk said the company has raised $40 billion in debt to buy GPUs.

He argues SpaceX differs from typical neocloud companies because it also generates revenue from Starlink satellite communications and rocket launches.

If competitors scale back because of funding strains, SpaceX could gain an advantage in winning customers, buying GPUs and building data-center capacity. Reduced competition could also strengthen its bargaining power on computing-rental pricing.

Social Capital founder Chamath Palihapitiya also said the neocloud industry could face sweeping restructuring in 2027.

He said delays in data-center construction and a GPU glut could prove more severe than expected, making 2027 a turbulent year for the industry.

SpaceX is not immune. Bloomberg reported that terms including the final interest rate on the $40 billion financing for GPU purchases are still being discussed.

SpaceX plans capital expenditures of $35 billion to $40 billion in the fourth quarter alone. If the broader AI computing market falls into oversupply, SpaceX could also face lower returns on investment.

3. Morgan Stanley Says Chip Stocks Could Fall Another 10% to 15%

Morgan Stanley said the correction in semiconductor stocks is in its late stages, but the group could still fall another 10% to 15% from current levels.

Strategists led by Mike Wilson expect the semiconductor correction to end around mid-November.

Based on technical analysis and comparisons with past moves in silver prices, they said the Philadelphia Semiconductor Index, or SOX, could find support near 9,500.

The SOX closed at 10,779 on Oct. 23.

The main reason chip stocks remain under pressure is that funding for AI investment is becoming harder to secure.

Morgan Stanley said companies that benefited from semiconductors and heavy capital spending are now at the center of funding concerns.

AI companies need vast sums to build data centers and buy GPUs. At the same time, governments and other industries are demanding more capital, intensifying competition for funding.

The strain is showing up even more clearly in credit markets.

The spread on five-year credit-default swaps tied to CoreWeave rose above 1,100 basis points at the end of last week.

Based on market pricing and assumptions including recovery values, that implies roughly a 50% probability of default within five years.

A wider CDS spread means investors are assigning greater credit risk to CoreWeave.

Morgan Stanley also said Federal Reserve policy is adding pressure to financial markets.

The bank’s analysis suggests Fed Chair Kevin Warsh is willing to tolerate tighter financial conditions, including shocks to equity and credit markets.

As demand for capital rises while liquidity remains limited, companies with weak balance sheets and richly valued growth stocks are facing greater pressure.

Despite those risks, Morgan Stanley maintained its year-end target of 8,000 for the S&P 500.

4. Dan Niles Says the AI Bubble Hasn’t Peaked, With Agentic AI as the Next Growth Driver

Wall Street technology investor Dan Niles said the recent selloff in AI-related stocks does not mean the AI investment boom has reached its final peak.

Niles said AI is the biggest technological revolution in history and that related investment amounts to about 3.5% of gross domestic product.

He acknowledged overinvestment in the AI industry, but pointed to the internet buildout of the 1990s, which also went through repeated sharp corrections as it expanded.

He divided AI’s development into three stages: training, inference and agentic AI.

Agentic AI, in particular, could create new computing demand because it uses 10 to 100 times more tokens than traditional inference.

In the near term, however, funding conditions for AI companies and slowing growth are weighing on share prices.

Niles said CoreWeave’s unsecured corporate bonds are trading at yields of about 15%.

Anthropic’s quarterly revenue growth also slowed to 45% in the third quarter from 145% in the second quarter.

The spread of open-weight AI models has pushed average selling prices for AI services down about 50% since May.

Even if AI usage rises, rapidly falling service prices could weigh on revenue growth and profitability.

Limits on data-center construction are also emerging as a new risk.

In the US, opposition to data-center construction is spreading because of power consumption and environmental concerns. Amazon recently withdrew plans for a large data center in Loudoun County, Virginia.

Even so, Niles said Microsoft is most likely to emerge as the ultimate winner in enterprise AI.

Microsoft has the advantage of being able to sell Copilot to its existing corporate customers. Copilot seats increased from 15 million in December last year to more than 30 million in June this year.

He also expects Azure growth to accelerate to 45% in the September quarter from 43% in the June quarter.

Niles said Alphabet is well positioned to win in consumer AI thanks to the Android ecosystem.

Still, he said market expectations for Google Cloud growth have become too high and worsening cash flow is a risk.

He expects Amazon to use physical AI to improve efficiency in logistics and e-commerce. AWS revenue growth also accelerated to 37% in the second quarter from 28% in the first quarter.

But tighter restrictions on data-center construction could also constrain AWS growth.

Niles said this week’s Big Tech earnings will be driven less by revenue and profit than by capital spending, cash flow and whether companies need to raise additional funding.

He said AI’s long-term growth potential remains high, but investors should raise cash holdings and stay cautious in the short term.

5. Yardeni Says Constraints on AI Investment Could Be Positive

Yardeni Research said financial and physical constraints on AI investment could jolt stocks in the short term, but could play a constructive long-term role by curbing overinvestment.

In a report on Oct. 26, Yardeni said the US midterm elections on Nov. 3 could bring fresh uncertainty to the stock market.

Prediction market Polymarket puts the odds of Democrats taking the House at 92% and controlling the Senate at 68%. The chances of winning both chambers stood at 67%.

If Democrats take Congress, President Donald Trump’s policy agenda could face constraints. Oversight and regulation of AI data centers could also be tightened.

In the US, local opposition to data centers has been growing because they can raise electricity bills and consume large amounts of water.

If state and local governments tighten regulation, data-center construction could be delayed and costs could rise.

Still, Yardeni said those constraints could ultimately benefit the AI industry.

As financing becomes harder to secure, companies will scrutinize investment returns and efficiency more carefully.

Power shortages and restrictions on data-center construction could also help prevent indiscriminate capital spending.

By contrast, large technology companies with strong balance sheets are projected to keep investing.

Capital expenditures by Amazon, Alphabet, Meta, Microsoft and Oracle are forecast to reach about $780 billion in 2026.

From 2027 through 2030, their annual capital spending is projected to exceed $1 trillion.

Meanwhile, small-cap and value stocks have been outperforming in the US market as large-cap growth shares undergo a correction.

Of 48 major US stock indexes in the week ended Oct. 23, 30 posted gains. That compares with 14 indexes that rose a week earlier, suggesting market breadth improved.

The S&P 500 large-cap pure growth index has fallen 17.5% from its peak.

By contrast, the Russell 2000 value index gained 2.1% last week, marking its strongest return for a second straight week. The Russell 2000 index and the S&P 600 small-cap value index each rose 2.0%.

The shift suggests investors are rotating into relatively undervalued small-cap and value stocks as concerns about AI investment weigh on valuations for large-cap growth names.

Yardeni said the US economy and corporate earnings should continue growing, making any pullback driven by political uncertainty and AI investment worries a buying opportunity.

6. Oil at $140 Could Trigger Demand Destruction

International oil prices fell about 2% on Oct. 26. The risk of further attacks on Saudi energy facilities remains, but concern is growing that if high prices persist, oil consumption itself could start to fall in a phenomenon known as demand destruction.

US West Texas Intermediate futures fell 2.19% to $101 a barrel. Brent futures dropped 1.89% to $114.51 a barrel.

Earlier, Yemen’s Houthi rebels claimed they had attacked major airports and oil facilities in Riyadh, the Saudi capital. Saudi authorities said they intercepted additional missiles and drones.

Michael Alfaro, chief investment officer at Gallo Partners, told CNBC that current oil prices include a geopolitical risk premium of about $20 a barrel.

He said oil could surge if Saudi Arabia’s major pipelines or refining facilities come under further attack.

But if Brent rises above $120, $130 and $140 a barrel, demand destruction would become more pronounced.

Demand destruction occurs when prices rise so much that consumers and companies cut back use of a product, reducing actual demand.

Oil is essential for driving, air travel, freight transport and industrial production, so demand tends not to fall sharply in the short term even when prices rise.

But if high prices persist, consumers may drive and travel less, while companies may raise prices or cut output to offset higher transportation and production costs.

If households spend more on gasoline and heating, they have less money for restaurants, travel, clothing and appliances.

Companies may also cut investment and hiring as costs rise.

If higher oil prices push inflation back up, the Fed could become more cautious on rate cuts even as growth slows.

That could create a situation in which expensive oil fuels inflation, curbs consumption and investment, and raises recession risk.

There is historical precedent.

In July 2008, WTI rose to about $147 a barrel. But as the global financial crisis deepened, oil demand collapsed and prices fell to around $40 by the end of that year.

The decisive factor behind that crash was the credit squeeze and recession caused by the financial crisis.

Another variable is the US shale industry’s reluctance to ramp up output aggressively.

Even with oil above $100 a barrel, US producers are hesitating to sharply increase capital spending.

That means another attack on Saudi energy facilities could send prices sharply higher again.

On the other hand, if oil prices fall from here, Wall Street may not read that as an unqualified positive.

If prices decline because demand is weakening rather than supply is increasing, that could be taken as a recession signal.

7. Is AI Charging Every Customer a Different Price?

Concern is spreading in the US that retailers could use artificial intelligence to charge different prices to different customers.

The Wall Street Journal reported that Nick Nado, 38, of Belleville, Illinois, recently tried to buy a hard drive through Walmart’s app and found the same item listed at different prices on different phones.

His phone showed a price of $369, while his wife’s phone showed a price that was $110 lower.

Nado said he suspected Walmart was using cookies or stored personal data to offer him a higher price.

Walmart said the discrepancy was caused by a system error, not price discrimination based on personal data, and that it had fixed the issue.

Still, the incident reflects what some US consumers increasingly describe as AI pricing anxiety.

Consumers worry that retailers, restaurants and ride-hailing services could use AI to analyze purchase histories and spending patterns, then charge different customers different prices.

The spread of electronic shelf labels is adding to those concerns.

Electronic shelf labels replace paper price tags with digital displays, allowing product prices to be changed quickly.

Some consumers suspect companies could combine AI with electronic shelf labels to suddenly raise prices depending on the customer or the situation.

Existing studies in the US and Europe, however, have found no evidence that electronic shelf labels lead to price increases during demand surges.

In some European grocery stores, they are instead being used to lower prices on products nearing expiration and reduce food waste.

Walmart has also moved directly to calm those concerns.

Chief Executive Officer John Furner said in a public letter last month that Walmart would not raise prices based on a shopper’s income, purchase history or urgent need to buy.

“We price products, not people,” he wrote.

McDonald’s also said it does not use AI to set menu prices or change prices based on real-time demand.

One reason the debate is spreading is prolonged inflation.

Based on the US consumer price index, overall prices have risen more than 29% over the past six years.

With living costs already high, consumers are more sensitive to the possibility that they could pay more than someone else for the same item.

State and federal governments in the US are also stepping up regulation.

New York state has passed a bill requiring companies to disclose when they use algorithms to set product prices.

Maryland, Connecticut and New Jersey have also passed laws restricting certain forms of personalized pricing or data-driven pricing.

The Federal Trade Commission has also said it could take action against deceptive or unfair pricing that uses personal data.

Allegations of AI-driven personalized price discrimination have not yet been fully proven. But the debate over how personal data is used and how transparent pricing decisions should be is widening.

A new fault line is emerging in US retail: whether AI becomes a tool for offering consumers more discounts or a tool for analyzing purchasing power and charging them more.

Shin-young Park, New York correspondent, Hankyung.com, nyusos@hankyung.com