Love Share

Apple is prepared to spend billions to catch the leaders in generative AI technology, according to Bloomberg Apple watcher Mark Gurman.

In his Power On newsletter Sunday, Gurman wrote that three Apple execs — John Giannandrea, Craig Federighi, and Eddy Cue — are leading the company’s generative AI push and are on a course to spend US$1 billion a year on the initiative.

He noted that Giannandrea is overseeing the development of a new AI system, including a smarter version of Apple’s digital assistant, Siri.

Federighi’s group, Gurman continued, is working on adding AI to Apple’s mobile operating system, iOS, filling it with features running on the company’s large language model (LLM).

Meanwhile, Cue’s group will be trying to add AI to as many Apple apps as possible, Gurman added, including Apple Music, Pages, and Keynote.

He also revealed that Apple executives were caught off guard by the AI fever gripping the industry and have been scrambling since late last year in an effort to play catch-up with first movers in the technology.

Playing the Long Game

Apple is less concerned than others might be about first-mover’s advantage, contended Anshel Sag, a senior analyst with Moor Insights & Strategy, a technology analyst and advisory firm based in Austin, Texas.

“There is still a lot to shake out in the AI space both on hardware and software, and I don’t think Apple believes it needs to rush to catch up,” he told TechNewsWorld.

“That said,” he continued, “it also cannot fall too far behind, and if you look at Apple’s continued silicon investment in AI compute on the iPhone, it’s quite clear that Apple is not letting the Microsofts and the Googles of the world get too far ahead.”

Apple’s focus on privacy and security also may have made its advances in AI take longer than Google and Microsoft, he added.

“Generative AI is not something you’d expect them to be way out in front on given their focus on hardware and other things, maintained David Smith, a vice president and analyst at Gartner, a research and advisory company based in Stamford, Conn.

“They’ve been doing quite a bit using AI. They just don’t have anything like ChatGPT to put out there. So, I’d expect them to try to have more,” he told TechNewsWorld.

He noted that $1 billion a year in R&D may be a large spend for some firms but not for Apple. “Apple spends $30 billion a year in R&D, so a billion isn’t colossal in their world,” he said.

Acknowledging AI Urgency

“Apple has been investing a lot into AI, but the generative piece has been a surprise for them,” observed Gene Munster, co-founder of Loup Ventures, a venture capital firm in Minneapolis.

“They’ve got a huge opportunity around generative AI,” he told TechNewsWorld. “I think they’ll figure it out.”

Nevertheless, he added: “There is urgency. They have to move faster.”

“They’re talking about a billion-dollar investment,” he said. “With this kind of opportunity, it should be a three to five billion a year investment.”

While Apple hasn’t hopped on the generative AI bullet train yet, it’s no stranger to artificial intelligence technology.

• In 2018, it acquired a Seattle-based AI startup called Xnor.ai that specialized in efficient AI algorithms running on edge devices, like the iPhone.
• In 2019, Apple acquired the majority of Intel’s smartphone modem business, which gave it a team of engineers with expertise in wireless AI chips.
• In 2020, Apple announced the A14 chip, which contains a neural engine for accelerated machine learning tasks, and each new generation of their A-series chips has expanded AI capabilities.
• In 2022, Apple released Core ML 4, which simplifies deploying machine learning models on Apple devices, as well as announcing plans to transition Macs to homegrown AI chips.

Closing Gap With Leaders

“Apple has integrated AI into all of its products for decades,” observed Tim Bajarin, president of Creative Strategies, a technology advisory firm in San Jose, Calif.

“I saw their first AI demo in 1987 when they introduced the Knowledge Navigator concept. And AI is used in Siri and maps today,” he told TechNewsWorld.

“What they have not introduced is a dedicated app like ChatGPT,” he said. “Instead, they choose to make AI technology a part of everything they do in software.”

For Apple, he continued, catching up to the leaders in generative AI isn’t as much of a concern as resolving if they want to create their own ChatGPT-like architecture for a dedicated Apple search program.

“Given their deep history of using AI, I would think that a public-facing app like ChatGPT could be created in the near future,” he noted. “But given their deal with Google for search and the amount of money they make from that, I am not sure if they want to make their own search engine.”

Sag, though, believes the catch-up game will be determined by what generative AI model Apple adopts — open source or proprietary.

“I believe that Apple will likely stick to an in-house software architecture because it knows the rest of the industry will adhere to whatever Apple does, and focusing on performance and battery life will be pivotal for Apple since it will want most things to run on-device for cost, privacy and security reasons,” he said.

Getting It Right

In his newsletter, Gurman wrote that the stakes are high for Apple to get generative AI “right.”

“They don’t have to be perfect out of the gate in every area,” asserted Greg Sterling, co-founder of Near Media, a news, commentary, and analysis website.

“They can roll out more limited functionality and then expand from there,” he told TechNewsWorld. “But they will face high expectations, scrutiny, and competitive comparisons.”

“Apple Maps is a cautionary tale,” he added. “The product wasn’t ready when it was released, and it took them basically a decade to recover.”

Rob Enderle, president and principal analyst at the Enderle Group, an advisory services firm in Bend, Ore., noted that getting generative AI right will be important for Apple but not critical.

“We are still in the earliest phase of this tech wave, so they have time for a little trial and error,” he told TechNewsWorld.

“They don’t have to get it right out of the gate,” Munster added. “They have to get it right in the next three to five years.”