Blab Tech
Apple Eyes Chip-Startup Acquisitions After M2 Ultra Limits AI Server Performance
Apple Inc. is reportedly looking to buy semiconductor startups to strengthen the company’s AI server processors, a move that marks a rare departure from its traditional avoidance of large acquisitions. The Information reported Wednesday that Apple has approached several chip firms directly and engaged bankers to explore potential deals.
Apple’s current AI servers run on its in‑house M2 Ultra chips, which have struggled to keep pace with demanding workloads. The company had planned to launch a new server chip, codenamed Baltra, later this year, but the release has been postponed. During a recent Siri update, Apple attempted to run Google’s Gemini language models on its own servers but had to fall back on Nvidia processors in Google’s cloud because the M2 Ultra could not handle the load.
Apple’s last significant purchase was the $2 billion acquisition of Israeli audio‑AI firm Q.ai in January. With cash and cash‑equivalents totaling $45.57 billion as of March 28, the company has the financial flexibility to pursue new hardware capabilities.
The company’s hardware strategy has long centered on in‑house design. Senior Vice President Johny Srouji, who oversees chip development, has led the creation of battery‑driven processors for iPhones, iPads, Macs and Vision Pro. Apple’s experience has been with mobile and consumer devices, not high‑performance server silicon that competes with Nvidia’s data‑center GPUs.
Baltra is being developed in partnership with Broadcom, according to reports. Originally slated for a 2026 launch, the delay has underscored the need for faster server‑grade silicon. While the M2 Ultra is powerful for desktop use, it lacks the throughput required for large‑scale AI inference.
Leadership changes may also influence the strategy. New hardware head John Ternus will take over as CEO in September, succeeding Tim Cook, and Srouji’s responsibilities have expanded to include broader hardware engineering duties.
Industry observers note that acquisitions would bring specialized expertise and accelerate the development of server‑grade chips. The Information report suggests Apple is in talks with multiple semiconductor firms and banks to explore acquisition options.
Apple’s approach contrasts with its historical preference for internal development. The Q.ai deal, focused on silent‑speech technology for on‑device AI, was the last sizeable purchase. The current push for server chips indicates a strategic shift toward building a more robust AI infrastructure.
Apple’s data‑center operations are expanding as the company grows services such as Apple Music, Apple TV+ and iCloud. Stronger AI processors could improve performance for machine‑learning workloads, content recommendation and voice‑assistant services.
No acquisition has been announced, and Apple has not confirmed any deals. The company’s exploration of chip‑startup purchases remains at the research and discussion stage.
In summary, Apple is actively exploring the acquisition of semiconductor startups to address limitations in its current AI server processors. The delay of the Baltra chip, reliance on Nvidia for heavy workloads, and substantial cash reserves suggest a strategic push to strengthen its AI infrastructure. Leadership transition and expanded hardware responsibilities may accelerate decision‑making. The next steps will likely involve further discussions with potential targets and financial institutions, but no concrete agreements have been reported.
Apple’s current AI servers run on its in‑house M2 Ultra chips, which have struggled to keep pace with demanding workloads. The company had planned to launch a new server chip, codenamed Baltra, later this year, but the release has been postponed. During a recent Siri update, Apple attempted to run Google’s Gemini language models on its own servers but had to fall back on Nvidia processors in Google’s cloud because the M2 Ultra could not handle the load.
Apple’s last significant purchase was the $2 billion acquisition of Israeli audio‑AI firm Q.ai in January. With cash and cash‑equivalents totaling $45.57 billion as of March 28, the company has the financial flexibility to pursue new hardware capabilities.
The company’s hardware strategy has long centered on in‑house design. Senior Vice President Johny Srouji, who oversees chip development, has led the creation of battery‑driven processors for iPhones, iPads, Macs and Vision Pro. Apple’s experience has been with mobile and consumer devices, not high‑performance server silicon that competes with Nvidia’s data‑center GPUs.
Baltra is being developed in partnership with Broadcom, according to reports. Originally slated for a 2026 launch, the delay has underscored the need for faster server‑grade silicon. While the M2 Ultra is powerful for desktop use, it lacks the throughput required for large‑scale AI inference.
Leadership changes may also influence the strategy. New hardware head John Ternus will take over as CEO in September, succeeding Tim Cook, and Srouji’s responsibilities have expanded to include broader hardware engineering duties.
Industry observers note that acquisitions would bring specialized expertise and accelerate the development of server‑grade chips. The Information report suggests Apple is in talks with multiple semiconductor firms and banks to explore acquisition options.
Apple’s approach contrasts with its historical preference for internal development. The Q.ai deal, focused on silent‑speech technology for on‑device AI, was the last sizeable purchase. The current push for server chips indicates a strategic shift toward building a more robust AI infrastructure.
Apple’s data‑center operations are expanding as the company grows services such as Apple Music, Apple TV+ and iCloud. Stronger AI processors could improve performance for machine‑learning workloads, content recommendation and voice‑assistant services.
No acquisition has been announced, and Apple has not confirmed any deals. The company’s exploration of chip‑startup purchases remains at the research and discussion stage.
In summary, Apple is actively exploring the acquisition of semiconductor startups to address limitations in its current AI server processors. The delay of the Baltra chip, reliance on Nvidia for heavy workloads, and substantial cash reserves suggest a strategic push to strengthen its AI infrastructure. Leadership transition and expanded hardware responsibilities may accelerate decision‑making. The next steps will likely involve further discussions with potential targets and financial institutions, but no concrete agreements have been reported.
