Why has Apple always refused to use NVIDIA?

In the era of AI explosion, NVIDIA has almost monopolized the AI chip market with its powerful GPU chips, becoming a partner that many tech giants are eager to pursue.

However, Apple has maintained a subtle distance from NVIDIA, and it can even be said that it is deliberately avoiding them. In fact, in the 2000s, there was a brief "honeymoon period" between the two, but over time, their conflicts have intensified.

This inevitably raises curiosity: why has Apple consistently refused to use NVIDIA? What kind of "grievances" and strategic considerations are hidden behind this?

Apple has always sought to create a complete ecosystem, and a large procurement of NVIDIA's GPUs would undoubtedly weaken Apple's dominance in the AI field. To break free from its dependence on NVIDIA, Apple has adopted various strategies.

However, as the competition in AI deepens, Apple faces pressure to train larger and better models, which will require more high-end GPUs. In the short term, the competitive and cooperative relationship between the two may still exist.

Historical Grievances: From "Honeymoon Period" to "Ice Age"

The cooperation between Apple and NVIDIA was not hostile from the beginning. As early as 2001, Apple used NVIDIA's chips in its Mac computers to enhance graphics processing capabilities. At that time, the relationship was good, and it could even be described as a "honeymoon period."

However, this honeymoon period did not last long.

The first significant event that caused a rift in their relationship occurred in the mid-2000s. At that time, Steve Jobs publicly accused NVIDIA of stealing technology from Pixar Animation Studios (of which Jobs was a major shareholder), which undoubtedly cast a shadow over their relationship.

In 2008, the tension between the two further escalated. A batch of defective GPU chips produced by NVIDIA was used in several laptops, including Apple's MacBook Pro, leading to widespread quality issues, referred to as the "bumpgate" incident.

NVIDIA initially refused to take full responsibility and compensation, which angered Apple and directly led to the breakdown of their cooperative relationship. Apple had to extend the warranty period for the affected MacBooks and suffered significant economic and reputational losses.

According to insiders cited by The Information, NVIDIA executives have long viewed Apple as a "demanding" and "low-margin" customer, unwilling to invest too many resources in them. After the success of the iPod, Apple also became more assertive, believing that cooperation with NVIDIA was difficult. Additionally, NVIDIA's attempt to charge licensing fees for the graphics chips used in Apple's mobile devices further exacerbated the conflict.

The Game of Business and Technical Strategies

In addition to historical grievances, Apple's refusal to use NVIDIA is also closely related to its consistent business strategy.

Apple has always emphasized comprehensive control over its product hardware and software, striving to create a complete ecosystem. To achieve this goal, Apple continuously strengthens its independent research and development capabilities and reduces reliance on external suppliers In the chip field, Apple is at the forefront of the industry. From the A-series chips for iPhone to the M-series chips for Mac, Apple continuously launches high-performance self-developed chips, gradually freeing itself from reliance on traditional chip giants like Intel. Against this backdrop, Apple is naturally unwilling to be constrained by NVIDIA in the AI chip field.

Apple hopes to have complete control over key technologies to ensure product performance optimization and differentiated competitive advantages. Massive procurement of NVIDIA's GPUs would undoubtedly weaken Apple's dominance in the AI field, limiting its product innovation and technological direction.

Moreover, although NVIDIA's GPUs are powerful, they also have issues with high power consumption and heat generation, which pose a significant challenge for Apple's pursuit of lightweight and portable products. Apple has always been committed to making its products lighter, thinner, and more efficient, while NVIDIA's GPUs somewhat contradict this design philosophy.

Apple has repeatedly requested NVIDIA to customize low-power, low-heat GPU chips for its MacBook, but has not succeeded. This has prompted Apple to turn to AMD and collaborate with them to develop customized graphics chips. Although AMD's chips are slightly inferior in performance compared to NVIDIA's, their performance in power consumption and heat dissipation better meets Apple's needs.

New Challenges in the AI Wave

In recent years, the explosive development of artificial intelligence technology has brought new challenges to Apple. To maintain competitiveness in the AI field, Apple needs to train larger and more complex AI models, which undoubtedly requires more powerful computing capabilities and more GPU resources.

To break free from reliance on NVIDIA, Apple has adopted a multi-pronged strategy.

First, Apple primarily rents NVIDIA's GPUs through cloud service providers like Amazon and Microsoft, rather than making large purchases. This approach can avoid significant capital investment and long-term dependence.

Second, Apple has used AMD's graphics chips and collaborated with Google to use its TPU (Tensor Processing Unit) for AI model training.

Additionally, Apple is working with Broadcom to develop its own AI server chip, codenamed "Baltra," which is expected to enter mass production by 2026. This chip will not only be used for inference but may also be used for training AI models.

Although Apple has been striving to break free from reliance on NVIDIA, the competitive and cooperative relationship between the two may still persist in the short term. Mastering core technologies is essential to remain undefeated in the fierce market competition