Apple wants to become less dependent on Nvidia for demanding AI tasks – and is considering a step the company has avoided for years: larger acquisitions. A report describes how Apple is specifically seeking out chip companies to advance its own AI server chips.
Apple already processes some of its AI tasks in its own data centers using its own chips, but for computationally intensive tasks, it still relies on third-party hardware. For heavy workloads - including the Gemini model behind the revamped Assistant - calculations run on Nvidia chips within the Google Cloud. According to a report by The Information, Apple wants to reduce this dependency, and sources familiar with the matter indicate that the company is now actively seeking suitable chip manufacturers. Apple has reportedly been in talks with investment bankers about potential deals in recent months.
Why the pressure is increasing right now
The timing is no coincidence. Apple's own AI server chip, internally codenamed Baltra, was originally slated for launch this year, but according to the report, it has been delayed. As long as this chip is missing, the gap in computationally intensive tasks will persist – and with it, the dependence on Nvidia, the dominant provider of AI server hardware.
This gap became apparent during the development of the new assistant. Apple's engineers reportedly attempted to run Google's Gemini models on its own server infrastructure, but ran into a limitation: the existing chips were designed for Mac tasks and couldn't handle a model of this scale. Therefore, Apple had to offload parts of the workload to Nvidia chips in Google's cloud. Apple's chip team has traditionally focused on battery-powered mobile devices, not on the high-performance server chips needed to compete with Nvidia.
A break with the previous acquisition strategy
Acquisitions would represent a significant shift in direction for Apple. Historically, the company has limited itself to smaller acquisitions in the hundreds of millions and avoided large deals. This approach is currently changing: In January, Apple paid nearly two billion dollars for the acquisition of the Israeli company Q.ai – the second-largest acquisition in the company's history after the three billion for Beats in 2014. Q.ai specializes in interpreting speech based on the subtlest facial movements.
This aligns with a signal from the latest quarterly report: Chief Financial Officer Kevan Parekh announced that Apple will no longer aim for a balanced cash-to-debt ratio. The company did not provide a reason, but the additional flexibility could free up capital for larger acquisitions. The fact that Apple's own chip expertise itself originated from an acquisition – the takeover of PA Semi in 2008 – lends a certain logical weight to this potential strategic shift.
Acquisitions as a complement to in-house development
Acquiring external chip expertise would complement, not replace, ongoing internal development. Apple is currently developing a server chip based on the M5 Ultra. A later M7 Ultra is expected to significantly boost AI performance - to a level that could rival Nvidia's Blackwell chip - and with support for up to 1.5 terabytes of RAM. However, a server chip based on the M7 Ultra is not expected to be available before 2029.
Acquisitions are just one of several ways Apple plans to reduce its reliance on Nvidia. The collaboration with Broadcom on an AI server chip has been ongoing since 2024, and both companies recently extended their partnership until 2031. Purchasing additional specialists would accelerate this effort, rather than replace it.
Why the change in leadership could tighten the course
The potential change in strategy comes at a time of significant personnel changes. On September 1st, hardware chief John Ternus will take over as CEO from Tim Cook. At the same time, chip chief Johny Srouji has been given expanded responsibilities, now encompassing all hardware development in addition to semiconductors. A leadership team with both hardware and chip expertise could provide further momentum for a more aggressive deal-making strategy.
Whether major acquisitions will actually occur remains to be seen – the report describes exploratory talks, not finalized contracts. The only certainty is the direction: Apple wants control over its AI infrastructure, just as it already controls the chips in iPhones and Macs. The path to achieving this currently leads through more than just its own development department. (Image: Shutterstock / Soft grass)
- OpenAI's first device is intended to be a speaker with a camera
- iPad mini 8: Leak mentions 60Hz display
- iPhone 20: Factory apparently ready for glass design
- M6 Pro and M6 Max: AI is the reason for Apple's decision to forgo AI
- M7 Ultra: Up to 1.5 TB of RAM in Mac Studio
- New Apple Pencil with replaceable battery – due to pressure from the EU
- iPhone 17: Report predicts significantly lowered demand forecast
- iPhone Ultra: New leak reveals battery capacity
- Apple is testing significantly larger AI models directly on the iPhone
- iPhone 18 Pro Max: Component costs could rise by almost $300
- More affordable Vision Pro: Display project apparently on the verge of collapse
- iPhone Ultra: Mass production underway, no delays expected
- The iPhone 18 Pro could be noticeably thicker
