Market confidence in the AI investment narrative is fraying as a series of negative signals emerge ahead of Wednesday’s open. Internally at Meta, CEO Mark Zuckerberg reportedly told staff that AI agent development is progressing more slowly than leadership anticipated. At the same time, contrarian investor Michael Burry has placed short positions across several AI-infrastructure stocks, and pre-market data showed notable pullbacks for major names: NVIDIA (NASDAQ:NVDA) fell 1.42% to $194.14, Meta (NASDAQ:META) declined 1.57% to $605.93, and Palantir (NASDAQ:PLTR) led the group with a 2.88% drop.
NVIDIA occupies a central role in the AI supply chain as the primary GPU supplier supporting the broader buildout. The stock’s pre-market movement left it roughly 17% below its 52-week high of $236.54. Given NVIDIA’s position, any substantial and sustained reduction in enterprise commitments to AI would likely depress chip demand before other links in the ecosystem feel the impact.
The internal update at Meta, reported to have occurred during a town hall on July 2, included a blunt admission from Zuckerberg that AI agents "haven’t accelerated the way leadership expected." That remark stands out because it contradicts the rapid-deployment rationale behind Meta’s recent restructuring and investments.
Investor sentiment is also being influenced by survey data cited by UBS via Intellectia.ai, which found that about 60% of businesses are cutting back on AI expenditures. The survey raises questions about whether the massive infrastructure investments underway are producing clear returns for the average enterprise or primarily benefitting large cloud providers and chipmakers.
Adding to the mood of caution is Michael Burry’s entry into short positions across multiple AI-infrastructure firms, as reported by 247wallst.com. Specific tickers tied to his positions have not been disclosed publicly. Burry’s activity has revived comparisons with the late dot-com phase; Barchart analysis published on July 7 draws an explicit parallel, noting the fate of Cisco, the infrastructure bellwether of that earlier cycle, which lost more than 80% of its value following the bust.
Oracle (NYSE:ORCL) was highlighted in the same vein as a contemporary echo: the stock has declined nearly 40% over the past year and was positioned for additional pre-market losses of 2.02% to $138.74 ahead of Wednesday’s session. Microsoft (NASDAQ:MSFT), meanwhile, has fallen more than 22% over the last year and was trading down another 1.37% in pre-market activity as of the same snapshot.
Competitive dynamics are exerting added pressure on Western hyperscalers and chip suppliers. Reuters reported on July 2 that a new low-cost Chinese AI model is performing at parity with Anthropic and OpenAI on standard benchmarks. That development recalls the DeepSeek disruption of early 2025 and fuels doubt about whether the scale of Western AI capital expenditure will translate into durable advantage if frontier-level models can be replicated at lower cost.
Further complicating the market narrative, Reuters reported on July 3 that Alibaba prohibited employees from using Anthropic’s Claude coding assistant over alleged "backdoor risks." That move crystallizes a recurring worry within the sector: the possibility that large-scale AI infrastructure could be used primarily for data collection or surveillance rather than broad commercial productivity gains.
Concerns about data-center obsolescence were captured in a comment by Mark Douglas, CEO of MNTN, who told Fortune on July 2 that U.S. data center capacity "is not going to age well," estimating roughly two years before obsolescence pressures mount. That perspective, coupled with sector-specific selloffs, appears to be prompting investors to reassess which parts of the AI value chain will sustain long-term commercial demand.
Palantir, a company with deep roots in government surveillance and a business mix heavily weighted toward defense and intelligence contracts, experienced the largest pre-market percentage decline in the cohort at 2.88%. That reaction suggests market participants may be discounting how much of the sector’s near-term value is tied to public-sector work as opposed to commercial adoption.
Looking ahead, the near-term outlook centers on a few key catalysts. Upcoming earnings calls for Microsoft, Meta, and Alphabet are likely to draw intense scrutiny over whether hefty AI capital expenditures are translating into measurable revenue growth or primarily inflating operating costs. Separately, Barchart’s July 7 coverage flagged the Federal Reserve’s next rate decision as a near-term market-moving event; for high-multiple AI-related stocks, any indication of a prolonged restrictive policy could steepen the discount applied to earnings that remain several years in the future.
Michael Burry’s short positions function as a real-time market test of the argument that AI-infrastructure valuations resemble late-stage dot-com froth. If AI-infrastructure shares continue to underperform while open-source and lower-cost models from China narrow the capability gap, pressure will mount on Western hyperscalers to justify the scale and pacing of their spending well before year-end.
Summary: A combination of internal admissions at Meta, short-selling by Michael Burry, survey data showing reduced corporate AI spending, pre-market declines in major AI-related stocks, and fresh competitive pressures from low-cost Chinese models is increasing investor skepticism about the durability of the current AI investment cycle.
- Key points:
- Meta’s CEO privately acknowledged AI agent development is progressing slower than expected, undermining part of the company’s investment rationale.
- Michael Burry has reportedly shorted multiple AI-infrastructure firms, reviving comparisons to late-stage dot-com valuations; broad pre-market weakness hit NVIDIA, Meta, Palantir and others.
- Competitive and security developments - a low-cost Chinese model matching frontier benchmarks and Alibaba’s ban on Anthropic’s Claude - raise questions about the long-term payoff for large-scale Western AI capital expenditure.
- Risks and uncertainties:
- Enterprise retrenchment in AI spending - approximately 60% of businesses in the UBS/Intellectia.ai survey report cutbacks - could reduce demand for GPUs and cloud services, affecting semiconductors and cloud infrastructure.
- Valuation risk in AI-infrastructure names, underscored by Burry’s shorts and the dot-com comparison in Barchart’s analysis; hardware and software infrastructure stocks may face steep re-pricing under prolonged investor doubt.
- Geopolitical and security-driven restrictions on AI tools, exemplified by Alibaba’s ban on a coding assistant, could limit commercial adoption and re-shape demand toward public-sector and surveillance applications, impacting firms with different customer mixes.