In our view, the technological revolution that is really driving this theme is "the cloud or distributed computing", not necessarily the models themselves. AI or Machine Learning knowledge and techniques have been around for a long time - but it was only when Amazon realized that they could rent out their spare datacenter infrastructure that we realized that we can scale computing resources much more rapidly than we ever thought. Since then, we've had Microsoft and Google enter the fray, and together, these companies are the infrastructure providers that enables it all. One can think of them as "digital rails" in some sense, analogous to the railroad companies in the 19th century which laid out the physical infrastructure powering physical networks of commerce.

These cycles of technological revolutions are often coupled with financial cycles, as new paradigms easily stoke the imaginations of investors. The first wave of Cloud adoption led to the boom in SaaS stocks (Software-as-a-Service) in the pre-pandemic days. But that was more of a false dawn in some sense. While investors were focused on software solutions for specific problems, behind the scenes, AI researchers were figuring out how to leverage computing and effectively make software itself capable of figuring out and solving any problem. Investments in the field will likely remain focused in two primary directions for the foreseeable future:

• building out the technical infrastructure needed and,

• training AI models.

The Internet can be a helpful proxy to think of a framework for AI - we didn't know how but we knew for certain it would change our lives. And how did we get to today? We first laid out the fiber optic cables and invested in the telecom infrastructure needed (3G to 5G and so on) before the smartphones and applications showed up that enable our highly digital lifestyles of today. When it comes to AI, it feels akin to the early days of the internet as we are still investing in the infrastructure (advanced chips, datacenters), and we are yet to reach that point where AI applications and use-cases are pervasive.

Strong investment performance in tech sector is making some investors nervous and "bubble" chatter is now fairly common. But are talks of a bubble premature?

Tech stocks tied to AI are booming and setting new all-time highs. Venture Capital dollars are flowing once again in Silicon Valley - and yet, there seems to be a certain skepticism or hesitance among investors. It may eventually become a bubble (normally identified after the fact), but we are nowhere close to that. It is tempting to look at Nvidia's stock price appreciation over the past year, compare it to stocks like Cisco in the dot-com era, and then dismiss it as evidence of overhype. But we know that the real value of any business is its earnings and cash generation potential. And, this time is truly different in many ways, as there's real earnings and cash flows that are powering this performance.

This is unlike the late 90's era when investors valued tech businesses all wrong - at the time, it was common to value tech stocks on the basis of clicks/eyeballs, rather than the inherent profitability of each click. Suffice to say, the market has learned from its past mistakes and case in point, while Nvidia's stock has appreciated 7x since its bottom in October 2022, it's valuation multiples have actually compressed and halved from peak levels - one may question why? And the answer is straightforward: Nvidia's earnings are projected to grow from $4.3 billion in 2022 to an astounding ~$56 billion in 2024. The market follows earnings and cash flows. Our view has remained steadfast all along and we stick to our thesis: We are in the early stages of an AI themed buildout of computing power that it is too early to even worry about a potential bubble.

In all likelihood, it's difficult to predict anything with reasonable conviction except that:

• we will see tremendous investment into computing capacity in the near term and,
• unprecedented change in the long-term.

Even from a real investment sense, we first need to get to the point of oversupply of computing power before we worry about this trend losing steam and the reality is there is not enough production capacity to meet the demand.

When framing what's next, a quote from the Great One (Wayne Gretzky) comes to mind - "Skate to where the puck is going, not where it has been." Said differently, a major reason some of these stocks have rallied is that the market consistently underestimated the fundamentals. It couldn't see where the puck (cash flows) was going. And understanding this is key to adding value for our clients. As long as we feel there is dislocation between our estimates of FCF (free cash flow) and consensus, a stock merits inclusion in our portfolios. For our portfolios if anything, the decision on "what we do with our Tech exposure today" is straightforward because "we made the right decision over a year ago". Among the range of Global and North American strategies at TD Asset Management Inc., we have been invested in this theme for some time, we will remain invested in our high-conviction ideas, and as such, we don't have to worry about playing catch-up.

本文所含信息由道明资产管理有限公司提供，仅供参考。信息出自我们认为可靠的来源。本信息并未提供财务、法律、税务或投资建议。具体的投资、税款或交易策略应根据每位投资者的目标和风险承受能力加以评估。

本文档中的部分陈述可能包含预测性的前瞻性陈述（“FLS”），其中包含“预计”、“预期”、“打算”、“认为”、“估计”和类似的前瞻性表述或其否定形式。前瞻性陈述基于当前对未来普遍的经济、政治、相关市场因素（例如利率和汇率、股票和资本市场）以及普遍经营环境的预计和预测，并假定不发生税法或其他法律或政府管制方面的任何变动或灾难事件。对于未来事件的预计和预测本身受无法预见的风险和不确定性的影响。此等预计和预测可能在未来并不准确。前瞻性陈述并非对未来表现的保证。实际发生的事件可能与前瞻性陈述明示或暗示的事件存在实质差异。包括上文所述各项因素在内的多个重要因素均可能造成这种背离。您不应在任何程度上依赖于前瞻性陈述。

道明资产管理有限公司 (TD Asset Management Inc.) 是道明银行 (The Toronto-Dominion Bank) 的全资拥有附属机构。

® TD标志和其他TD商标为道明银行或其子公司的产权。