For smartphone and PC companies, artificial intelligence is a story of risk with uncertain reward. Producers need to invest in AI capability at the peril of being left behind in a potentially game-changing industry shift. Yet the benefits have been limited so far, and the future payoffs are hard to value.

S&P Global Ratings does not anticipate that AI will accelerate the replacement cycle over the next two years, even as devices are rapidly becoming more AI-capable. The trend could eventually boost premiumization of PCs and smartphones; but at this stage, the boost is more on average selling prices (ASPs) than margins.

The narrative could quickly change on breakthroughs in AI applications. For example, shipments could jump if AI search and other AI software are enhanced and further integrated to create "must-have" services across AI smartphones and PCs. This would be similar to how fully touchscreen smartphones altered people's way of digital life.

Another potential driver of change is a diverging path for China and rest-of-world on AI-development, due to technology restrictions. Regulation remains a wild card.

AI Is Not Shaking Up Market Share Or Driving Replacement Cycles

AI is not yet a meaningful differentiator among smartphones and PCs, given the commoditized nature of hardware products in these industries. Most PC and smartphone brands are investing in AI capability, and while some are ahead of the others, the differences are not material (see tables 2 and 3 in our Appendix). Hence, we do not expect their market shares to appreciably shift.

Please see our Appendix for breakdown on the various AI-related offerings for leading smartphone and PC competitors. This includes the various paths they are pursuing, including Android versus Apple's self-developed operating system (iOS).

Earlier Replacement Cycles Look Unlikely

AI capable devices will make up a growing share of shipments--at more than 30% of smartphones by the end of this year, from just 5% in 2023, according to consultancy IDC (see chart 1).

Chart 1

Yet we anticipate this trend, which entails costs for the original equipment manufacturers (OEMs), will not meaningfully pull forward their device replacement over the next few years. Rather, the rising AI smartphone adoption is mainly supply-driven and reflects consumers' purchases of the latest flagship models at their normal replacement cycle. S&P Global Ratings forecasts global smartphone shipment growth will slow to 2% in 2025 from 6% in 2024, after an accumulated 14% decline over 2022-2023.

As for PCs, the market is still waiting for more attractive AI PC offerings that should hit in the coming months. Meanwhile, some replacement activity is occurring on the commercial PC front due to the expected phase -out of Windows 10 support. If more

If more affordable AI PC models with better use cases fail to hit the market over the next two quarters, then momentum could stall after enterprises refresh their PCs for Windows 11

Why isn't AI driving a quicker replacement cycle?

In our view, a large portion of consumers may not rush to buy a new AI device and will likely wait until their usual replacement period. For example, many of the currently available AI functions, such as photo editing, translation, search with image and Gen AI chat-box, can be done by existing third-party apps or software, possibly with a few more steps.

Multiple recent surveys by third parties also reveal that the majority of consumers are not very excited by AI smartphones, especially outside Asia.

Chart 2

Chart 3

Currently announced AI capabilities seem to be an improvement on existing functionality, rather than driving new user behavior. This is like many other hardware upgrades in smartphones' history, such as camera lenses and fifth generation telecommunication (5G) technology, which drove improved customer satisfaction but failed to result in a meaningful and sustained acceleration of device replacement.

Chart 4a

Chart 4b

For PCs, one barrier is the unknown for businesses

AI-generated content could be another reason for the slower-than-expected adoption of AI PCs in the workplace, especially given businesses nowadays account for 60%-70% of PC end demand. It is difficult to test the logic and accuracy of AI generated content, and leakage of proprietary information could potentially result in serious legal consequences.

AI applications may need to be tailor-made and carefully tested and monitored for enterprise customers, which entails higher costs. However, we expect as best practices develop and become industry standards, the proper use of AI-generated content will become less of an obstacle for AI PC adoption in the workplace. On the contrary, retail consumers have been faster to embrace new AI features and adoption of AI PCs.

How Are Smartphone Vendors Joining The AI Race?

Apple's stakes are highest.  Given its unique operating system, Apple's route to AI integration has the most potential to beat competitors--or fall behind the pack. Compared with other players, Apple makes more in-house AI related hardware and software investments. If Apple and OpenAI fall behind Google in their LLM capabilities, Apple's AI products may lose some tractions with consumers.

Nevertheless, Apple has the strongest financial resources and proven track record on R&D capability. Its strong brand recognition and differentiated operating system help support high customer loyalty. Its inability to launch Apple Intelligence in China is also likely to impact its already-declining market share there. Potential U.S. tariff on Chinese imports and retaliatory measures by China would hurt Apple the most among the major smartphone OEMs.

Samsung is a few paces ahead of other Android competitors.  The brand showcased its software integration capability with AI functions in its flagship smartphone lineup at the beginning of 2024, leveraging Google's Gemini LLM. While Google makes some AI features (such as Circle To Search) available to other brands in the Android camp at a certain point of time, Samsung's early move on AI has helped boost sales of its premium phones during the first half of 2024.

Most Chinese brands will face tense competition.  They are investing to develop their on-device LLM and AI applications, mainly for their domestic products. For their international products, we believe they will rely on Google's Gemini or other third-party LLM, thereby limiting Chinese vendors' differentiation. Competition is heightening due to their time clash in new AI domestic product launches that started last year, as they align their flagship launch cycle with Qualcomm's release of its latest Snapdragon chipsets.

Xiaomi aims for premium payoffs.  Its approach to AI integration is similar to other Chinese OEMs. The AI functions on its domestic smartphones are incorporated in its Android-based "HyperOS" system. Xiaomi has been pushing a premiumization strategy for its smartphones in recent years, success of which depend in part on its development of AI features that currently resemble those offered by its Chinese competitors. There are execution risks, especially in the premium smartphone markets overseas. Xiaomi is also increasing its investment in EVs. It remains to be seen if this new business could dilute its corporate resources for AI smartphones.

Huawei's software stands out among the Chinese brands.  We believe Huawei has better software development capabilities, owing to its proprietary smartphone and PC operating system. The challenge for the company is on its supply chain. Whether it can source sufficient advanced chip sustaining its smartphone-based AI-features is a key determining factor to its success. This is also an exogenous factor for other brands' smartphone market share in China.

AI Upgrades Don't Come Cheap…

At this stage, the real winners in the AI bonanza are the makers of processors, software or related services that need to be purchased by other tech companies.

On the hardware side, device makers are paying higher prices to secure the advanced chips and other AI-enabling components.

On the software side, third-party software developers such as Google should have the advantage in this growth industry with their massive data collection and processing capabilities.

Currently, the most popular AI-related software is made by Google, Microsoft and OpenAI. In China, internet companies such as Douyin, Baidu, Alibaba and Tencent as well as AI start-up companies Moonshot (the Beijing-based start-up that launched Kimi LLM) and DeepSeek (a quant-fund backed start-up based in Hangzhou) are now leading the way on AI development. Their initiatives have been accelerating, as tech restrictions may limit the use and selling of global AI-related software to the country.

We think the story can rapidly change as more AI competitors enter the market and the applications rapidly broaden. At this stage, however, paying for processing is expensive and it's not clear if the costs can be passed along.

Pricing schedules for smartphones so far show that manufacturers have not increased like-for-like prices on AI-capable smartphone models relative to previous year's models despite higher component costs (see table 3 below).

Table 1

For PCs, we expect to see a similar trend but with a delay in AI penetration and higher ASPs. The lukewarm overall PC demand this year, higher prices on AI PCs and wait-and-see attitude by enterprise customers are key reasons for the delayed AI-driven upgrade. As more affordable AI models become available in the coming quarters, PC makers could see an ASP and revenue boost. We currently expect AI PC to account for about half of total global PC shipment by 2027, similar to IDC's estimate.

…But Could Pay Off If AI Software Accelerates Interest

In our view, the hardware is ready. Radical change hinges on software development

We believe smartphones have the best potential to monetize the AI trend if the software applications get more interesting. This is how innovation typically works.

An example of a "game changer" was Apple's release of its first iPhone. As the number of smartphone applications increased exponentially and the hardware improvements were made, an increasing proportion of online activities have been moved from PCs to smartphones, eventually replacing cameras, maps, newspapers, radio, etc. in many people's lives.

An AI-related "killer app" would, in our view, meaningfully accelerate device replacement, and increase ASPs and margins. In our view, the leading OEMs would likely be best positioned to benefit.

Regulation Could Be A Wild Card

Governments across the world are rapidly releasing new AI regulations. A strict regulation could limit versatility of AI features and reduce differentiation. The inconsistent launch time for Apple Intelligence across countries and regions globally reflect the regulatory challenges for AI software. For example, Apple delayed its Apple AI rollout in the EU market due to certain regulatory concerns. Nor has the company cleared the hurdles for launching Apple AI in China, and that contributed to its soft smartphone sale in 2024.

In China, regulators are focused on information security and content moderation. Gen AI service providers in China are required to register their AI models with the government and have their AI-generated content closely monitored. Currently, LLMs developed by foreign companies, such as ChatGPT, Gemini and MS Co-pilot are not available in China. This will somewhat favor Chinese vendors that are in better position to address the regulator's security concerns and consumer preference, while developing their AI applications.

Chart 5

Foreign smartphone and PC OEMs therefore rely on Chinese-developed LLMs, such as from Baidu, for AI features in their products sold in China. Restrictions on AI software developed overseas and a growing push to adopt domestically manufactured tech products could result in a further share loss of foreign brands in China. This is especially true for Apple's smartphone products sold in China.

A Tough Risk-Reward Scenario For Credit Ratings

We expect our rated smartphone and PC OEMs to remain relevant through the rapidly shifting AI landscape. Apple Inc., Samsung Electronics Co. Ltd., Xiaomi Corp., Lenovo Group Ltd., Dell Inc. and HP Inc. will continue to invest in AI innovation and release AI-capable products into the market (see Appendix). This transition is not without risks.

The fast-evolving AI technology and software could reduce OEMs' visibility on demand and may bring challenges to inventory management and cash flows. If the AI-driven demand falls significantly short of the initial expectations, inventories could rise quickly across the sector, and OEMs would need to go through another round of destocking. As such, a decline in ASP, revenue and margins and negative working capital will strain companies' cash flows.

This technology shift will also require higher R&D investments and component cost inflation, which may pressure the EBITDA margins in the smartphone and PC hardware segment if these companies can't pass on the costs. Over the longer term, smartphone OEMs may have better potential than PCs to monetize new AI software and services, for example by charging fees on more AI privileges to users.

All technology shifts bring disruption and unknowns. The smartphone and PC OEMs we rate are leaders in their respective sectors, and have stayed on top of the game in past technology cycles. They have sufficient financial buffers to absorb AI-related expenses, including potential margin hits. Our view remains that when, not if, AI smartphones and PCs take off, the current market leaders will be best placed to reap the gains.

Appendix: The Commoditized Competitive Landscape

Overall, the leading smartphone makers currently have pretty similar AI-related offerings.

Table 2

Apple may continue to differentiate itself with self-developed application processors and iOS, which allow it more flexibility to integrate more AI features in its devices. Android smartphone vendors, in markets outside mainland China, will generally rely on Google's system and Gemini large language models (LLM), then make their own local modifications.

Smartphone makers' competitive edges depend on both their hardware design and software development capacity as well as the LLM competition between Gemini and OpenAI, similar to the pre-AI era.

Likewise, we expect the competitive landscape for Windows-based PC makers to largely remain status quo, in part due to capabilities that currently appear to be undifferentiated (see table 4).

The AI PC models included below are compatible with Microsoft Copilot and hence share the Copilot capabilities. The table does not capture the differences in processing power, better battery life and better form factor that can affect the value to users. Moreover, not all PCs are equally compatible with Copilot, and this application is not available in some countries and regions.

Table 3

Table 4

Editor: Cathy Holcombe

Digital design: Tim Helyer

Related Research

• Industry Credit Outlook 2025: Technology, Jan. 14, 2025
• Full Analysis: Xiaomi Corp., Jan. 13, 2025
• Tear Sheet: Lenovo Group Ltd., Nov. 22, 2024
• Gen AI Is Writing A New Credit Story For Tech Giants, Nov. 13, 2023