….about AI PCs and why they could be revolutionary for the channel
Last month TD SYNNEX brought its Destination AI programme to the UK and Europe, providing a framework to support resellers on their AI journey.
The end-to-end programme has four pillars: awareness – to educate partners on the latest AI advancements and trends; enablement – to give partners the tools and knowledge, certifications and training they need to build solutions and incorporate AI seamlessly into their operations; support during the sales cycle – to provide guidance and expert advice to help partners design and create solutions to meet customers’ needs; and, lastly, support after the sale – to help partners with post-sale services that could come on the back of AI solutions, potentially stepping in where partners don’t have in-house capabilities or resources.
The bulk of our conversation with Brandon Lieberthal centred on AI PCs and their likely impact on PC sales, but to start with we asked him about Destination AI and what that end-to-end programme tells us about the changing role of channel partners.
Technology Reseller (TR): The four pillars of Destination AI suggest that AI will require a more joined up approach from partners. Is that how you see it?
Brandon Lieberthal (BL): I think that’s spot on. Looking at it purely from one piece of the puzzle, which sits within my area of responsibility, the emergence of AI PCs will revolutionise the role of the channel. Why will it revolutionise it? Because you’re altering the value proposition of the channel. Whether you’re a value-added reseller or focus purely on software, AI is all about the solution. You can buy the fastest PC with the highest TOPS count (trillions of operations per second), but until you have an AI app to go with it, it’s just like the PCs we sold before.
The value proposition both for vendors and resellers is going to change. No longer will you just be a vendor of hardware devices or a channel partner that has stock for next-day delivery at a good price, with a bit of technical support. You are going to need a more consultative mindset and use hardware devices to create a final product that delivers a solution enabled by AI.
The challenge for the channel will be developing the right skill set. If your value prop is to help your customers with AI trainings, you will need to make sure you have your own resource. As we saw with cybersecurity three or four years ago, there will be skill set shortages in the channel, and with the greater focus on services and solutions, there will need to be more training for sales reps and technical reps.
The benefit for the channel and for vendors will be an increase in margins. When you have the right skills with the right solutions, you’ll be able to command a higher price in the market. Prices will slowly start to increase because of technological advancement, but also because many partners will be able to charge for more services than they might have been able to in the past.
The final piece is you’ll probably see a higher turnover of equipment as demand increases and new use cases and new applications emerge. Typically, refresh cycles started with the move from Intel Pentium 1 to Pentium 2 and were then driven by changes to the operating system or by an OS end of life and finally by financial considerations – a PC was an asset and when that asset had depreciated to a certain point it was replaced. I think we’ll now see churn really driven by use cases and applications and the need for more computing speed – ‘I’m going to need TOPS of greater than 100 and the machine I bought a year ago only has 40’. That churn will be different to what we’ve experienced in the past 20 or 30 years.
TR: It’s what the PC industry needs at this point.
BL: I don’t like to celebrate COVID, but it did give us a real shot in the arm, albeit for the wrong reasons. The market wasn’t driven by a technological advancement but by the need to be able to work from home. You’ve got to go way back to 1999/2000 to find something that made everybody want to move forward.
AI PCs are revolutionary because when people start to really feel the benefits of AI, they will recognise that they need an AI-enabled computer to help them work better. It’s been a while since people have said I need a computer, but I don’t need just any computer.
TR: How would you define an AI PC?
BL: I made a video a couple of weeks ago to help answer something as simple as that because there are a lot of interpretations around it.
From a hardware perspective, the single biggest difference between an AI and non-AI PC is that an AI PC must have a neural processing unit (NPU) in the processor. After that, you can start to ask yourself, what is the best one.
NPU speed in measured in TOPS (trillions of operations per second). Last week Microsoft announced the Copilot+ PC powered by Qualcomm, the first so called next generation AI PC with a speed greater than 40 TOPS.
In its most simple form, that is what differentiates an AI PC from a non-AI PC.
TR: People are using those two letters, AI, at every opportunity. Is there a risk devices are going to be sold as AI PCs when they might not be?
BL: In our industry, we have different vendors with different strategies who may disagree on the terminology, but what everyone can agree on is the requirement for a neural processing unit.
Beyond that, Microsoft might say their view of an AI PC is that you have Windows 11, you have a Copilot button and you have a neural processing unit. Apple would probably disagree with that: one, because they are not going to put a Copilot button on their machine; and, two, because they’ve had neural processing units on their machines for some time without calling them AI.
When we talk about TOPS, I’m a little cautious in case we get ahead of ourselves and we start to see competing messages on the market, for example if Apple were to say ‘you’re talking about a technology we had three or four years ago and you didn’t notice it then so why do you care if it’s 40 TOPS today?’.
Vendors will have a view. Software manufacturers may have a view. Processor manufacturers will definitely have view. Qualcomm, Intel, AMD typically talk about NPU and TOPS as their current focus.
TR: Some of the projections for AI PC sales are very impressive. What are some of the challenges to be aware of?
BL: The first one, the most obvious one, is use case. Costs will go up and not only hardware costs. The use case is defined by the application and the application costs will be there, whether it’s CoPilot, whether it’s a Google application, whether it’s Adobe. Where you need the hardware plus the app, costs will go up.
That raises the question of who needs AI? Many CIOs and CFOs will be asking what they get in return for those costs. When we talk about the benefits of an AI PC, the number one point that comes up is productivity. Tools such as Copilot give a huge amount of productivity and allow people to become multi-skilled. For example, before, if you wanted to do copy or imagery work, you would often have to go to multiple people, which impacted productivity. Now you could potentially get Copilot to do it for you.
If you look at the growth projections, I don’t think we’re getting ahead of ourselves. The number I hear bandied about is that this year a maximum of 15% of machines will be new AI devices. Then, when we get into 2027, we talk about 60%-65%, maybe 70%. That tells me that a lot of people will still be transacting on the devices they use today and that’s probably because they won’t have the need for an AI-enabled PC.
That’s fine, but what we need to be wary of is everyone saying I’ll wait for the next generation to come out. The job of the channel, whether that be vendor, distributor or reseller, is to convey the real benefits of AI to their customers so that they see the value in investing in the applications and understand that those applications work better on hardware that’s AI-enabled. That’s our first step in this journey, education.
TR: Might people argue that they have been using AI satisfactorily on their existing devices?
BL: For me, it’s all about having the applications to provide a use case, and so far there are two names that people recognise. The first one, which woke us up to AI, is ChatGPT. If you ask people what AI means to them from a software perspective, most will say ChatGPT. That’s a web-based application.
Of course, we know there are security risks and data risks around what you type into ChatGPT because you’re sharing what you want with an external person, which brings you to this edge concept. How do you take the power of the cloud and bring it safely onto the edge device so you don’t need to share everything with everyone? You do that with Copilot.
The truth is you’ve got two versions. You’ve got Copilot, available on every Windows 11 machine now coming out, giving functionality similar to what people use ChatGPT for. That’s the cloud-based offering. Then you have Copilot for Microsoft 365, with subscription-based enterprise licensing, which is where you will experience the benefits of the machine and the ability to do a lot of computations on the edge.
Almost every person will have an AI assistant. We call it a personal computer and it really will become very personal. It will learn about you and get used to how you speak and how you think, to the point where it’ll start to answer emails for you, write them up, write a response. You’ll just say I agree or I don’t.
We’ll also get more used to collaborating between tools and apps. I always explain Copilot in terms of its ability to transform the way we work, really integrating all of the capabilities of AI into the things we do daily.
This is what CIOs need to think about. What model do they want? Do you invest in the machine if security is a big driver for you or if you’re a data scientist and you need huge computing power? In certain areas, do you replace human interaction? Today, we are all used to interacting with
a chatbot, but AI will take this to the next level where you could have something you think is a human being talking to you in any language you want. Interactions in our day to day lives will change as a result of AI solutions powered by AI PCs.
TR: You mention data security. Do AI PCs have additional levels of security?
BL: When people ask me about the advantages of an AI PC, the first thing I mention is greater efficiency. Number two on my list is security.
Today we tend to think about lock and key security – you might say my machine already has a software application on it managing security. AI PC will take security to another level. For example, threat detection will become managed on a machine and react more quickly than traditional security software that needs to see potential threats, see what’s turning into a real threat and then put patches out onto machines.
AI will learn and start to detect patterns. It will respond in real time to what it thinks is a threat and can probably identify unusual behaviour a lot faster because it’s now personalised to your machine, in addition to cloud AI looking at network traffic and unusual behaviour there. More importantly, it will be able to take preventative actions.
Many companies run simulations on a weekly or monthly basis, sending out emails to check the compliance of employees, and compliance is not always great because there’s a human interaction. Mention something everybody’s talking about, like Taylor Swift tickets, and your brain automatically lowers its defences. This machine will likely recognise that’s not what normally happens on this network and take a preventative action. Phishing is getting smarter and AI will be able to start challenging it and matching it in terms of speed of identification and resolution.
Another benefit of an AI PC is that the new processors and the new devices coming out are less power intensive, better with power management. One unique selling feature might simply be that your fleet of 20,000, 30,000 or 40,000 power- hungry PC devices will offer better power efficiency in the future. That may be the use case. It may be as simple as saying all I want an AI PC for is better power management and longer battery life. For many, a 20-hour or 22-hour battery life will be a game-changer.
TR: What needs to happen if AI PCs are to have the impact on PC sales many are hoping for?
BL: People in the channel have to position themselves as experts in AI technology from a hardware and from an application perspective. Investing in training and certification for sales teams and technical teams will enable them to have more valuable conversations with customers. Approaching this as a solution sell, rather than individual pieces, whether that’s an application or a device, will bring additional money and benefits through consultancy and integration.
For vendors in particular building a strong relationship early on, making people understand the benefits of AI PCs, providing training to the channel and helping them to be able to convey those benefits to their customers is going to get adoption going.
I really believe we’ve got a huge opportunity. It’s been a while since there’s been a killer reason to replace PCs and at the moment there are a few big things coming together. You have AI PCs and the benefits of AI. At the same time, there’s a big opportunity with Windows 10 going end of life next year. I’ve seen estimates that 67% of the hundreds of millions of devices out there are still not running Windows 11.
There is an opportunity to merge those conversations. You have to move to Windows 11 for future-proofing and there are also compelling reasons to move to AI PCs. Those two things will really elevate the industry starting in the second half of this year and put some turbos on it as we go into 2025. Together, they will get PC growth back to levels we last saw during COVID.
Be First to Comment