Microsoft is pushing hard on AI products and has released a new suite of laptops and tablets featuring an AI assistant called CoPilot+.
Before you dismiss it as just another AI chatbot, consider that this is a strategic move by Microsoft to establish a new way of using devices and capture market share if demand picks up.
For the last decade, computing has been moving to the cloud and becoming more device-agnostic, fueling the demand for servers and data centers. However, Microsoft has thrown a curveball by enabling CoPilot+ to process data locally on the device. This allows AI functions to be available offline, without a connection to the internet.
This is the first of its kind and comes with key advantages. First, it could accelerate AI speed since it is not dependent on server load and connection speed; the device is dedicated solely to the user. It should also be cheaper to run on device for services Second, privacy is a concern for some users. Having an AI on the device means that communications with it and the data will not need to be processed or stored on a server shared with other users and owned by corporations.
Having AI on a local device is made possible by Qualcomm’s Snapdragon X Elite and Plus processors. These chips include a CPU, GPU, and, most importantly for AI functionality, neural processing units (NPUs). Microsoft claims that CoPilot+ PCs are 58% faster than Apple’s M3 MacBook Air.
These chips are based on Arm architecture and are power-efficient, which is crucial since AI processing consumes a lot of power. CoPilot+ PCs can offer up to 22 hours of local video playback.
Microsoft’s strategy is not about creating a walled garden. Instead, it often embraces a model of “coopetition.”
In Satya Nadella’s book, Hit Refresh, he elucidates this core business strategy:
“Partnering is too often seen as a zero-sum game—whatever is gained by one participant is lost by another. I don’t see it that way. When done right, partnering grows the pie for everyone—for customers, yes, but also for each of the partners. Ultimately the consensus was that this partnership with Apple would help to ensure Office’s value was available to everyone, and Apple was committing to make its iOS really show off the great things Office can do, which would further solidify Microsoft as the top developer for Apple.
For me, partnerships—particularly with competitors—have to be about strengthening a company’s core businesses, which ultimately centers on creating additional value for the customer. For a platform company, that means doing new things with competitors that can accrue value back to one of the platforms.”
This philosophy explains why Microsoft made ChatGPT available on Bing rather than pushing its own solution. It also underpins Microsoft’s decision to incorporate OpenAI’s new GPT-4 model into Copilot+ PCs as a native app.
Furthermore, this approach is evident in Microsoft’s partnership with Qualcomm, instead of using its own chips like Apple, and its plan to feature Intel and AMD chips to power AI devices in the future.
Microsoft will also make CoPilot+ available to other PC manufacturers such as Lenovo, Dell, Acer, Asus, and HP, not just on Microsoft devices. This inclusive strategy aims to accelerate adoption and market reach, leveraging the fact that these manufacturers have been offering Windows OS.
Microsoft is attempting to be a first mover in enabling AI functions locally on devices and spreading them to other players as quickly as possible, thereby making CoPilot+ a ubiquitous digital personal assistant with network effects—much like Office software whereby the advantage is that it becomes hard to switch away from it because everyone else is using it.
I previously thought that Google was best positioned to do this, but Microsoft beat Google to it first. (It is not too late, Google.) Additionally, this might be a reason for Apple to kill off its EV ambitions and focus on delivering AI in its devices. Apple might have gotten wind of Microsoft’s intentions, and if Microsoft succeeds, it could pose an existential threat to Apple.
If Microsoft can succeed in defining AI PCs, it will gain significant mindshare as well as tangible market share in the years to come. However, I don’t expect adoption to be rapid, it feels like too much AI to take in right now; there’s some fatigue setting in, and I don’t foresee myself clamoring for an AI PC at the moment. Nevertheless, this is definitely a business strategy worth monitoring, along with Microsoft’s stock.
