VECTOR WIREAI INTELLIGENCE
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models Deals Regulatory Sources

Apple Explores AI Chip Startup Acquisitions as Internal Server Silicon Falls Short

Apple is in talks with bankers and semiconductor startups about acquisitions to strengthen AI server chips, as its Baltra chip slips past its 2026 target.

Vector Wire — AI-assisted editorial illustration

Apple has been actively exploring acquisitions of semiconductor startups to bolster its AI server capabilities, according to a report from The Information cited across multiple outlets2,3,4,6. The company has spent recent months talking to investment bankers about possible deals and approaching chip startups to gauge their willingness to sell.

The move represents a departure from Apple's longstanding preference for building silicon in-house. The catalyst: Apple's current AI data centers run on its own M2 Ultra chips, which handle lighter AI tasks but fall short on the heaviest inference workloads. For more demanding tasks, Apple still relies on Nvidia GPUs running in Google Cloud, including for parts of the Gemini-powered Siri backend.

Apple's next-generation server chip, codenamed Baltra, has slipped past its planned 2026 debut. The company's next-generation M7 Ultra server chip is not expected until 2029, according to MacRumors as cited by Gagadget. In the interim, Apple plans to upgrade its data centers with M5 Ultra chips.

Apple has recently held talks with PrismML, a company specializing in model compression. The report also notes Apple acquired Q.ai earlier this year, suggesting a pattern of using external technology to address weaknesses in its proprietary AI chip development.

ANALYSIS The acquisition strategy underscores a tension between Apple's vertically integrated chip design philosophy and the pace required to compete in AI infrastructure, where Nvidia's dominance in data center GPUs has created a capability gap that Apple's mobile-optimized silicon cannot bridge on its current timeline.

Apple has not commented on the report, and Reuters noted it could not independently verify the details.

Separately, Apple cleared a key regulatory milestone in China. On July 15, the Cyberspace Administration of China released a new list of approved generative AI services for mobile devices, with seven AI products receiving filing approval1. Among them, Apple Intelligence — submitted by Apple Technology Development (Shanghai) Co., Ltd. — officially passed the filing on July 8 and is intended for use on Apple phones. The approval removes what the report described as the most critical regulatory hurdle for AI compliance in China, bringing the Chinese version of Apple Intelligence closer to official launch.

Apple's stock rose nearly 3% in early trading on July 16, reaching a new all-time high and trading at $324.11. The South Korean outlet Asiae reported that the stock gains reflected growing visibility of Apple's AI-sector progress, noting the company has been considered a relative latecomer compared to Google and OpenAI5.

ANALYSIS The two developments — chip acquisition exploration and China regulatory clearance — address distinct bottlenecks in Apple's AI strategy: server-side compute capacity and geographic market access, respectively. Together they signal an acceleration of effort across multiple fronts after a period in which Apple's AI positioning drew skepticism relative to peers.

CORRECTIONS: none for this article · this piece updates automatically as the story develops · corrections policy & trail →