For the first time in history, intelligence itself is becoming a resource we can scale.
Artificial intelligence has taken markets by storm, going from niche experiment to mainstream productivity engine. Adoption is accelerating, with a large share of workers now using AI in their daily workflow. AI is touching almost every function, from code-writing and customer support to inventory management and creative work. The result is a step change in productivity and a rebuilding of how organizations are structured from the ground up.
Key Takeaways from the year in AI:
-
Capital expenditures (CapEx) from megacap tech companies have laid the groundwork for the AI infrastructure buildout, with bottlenecks shifting from chip supply to electricity generation.
-
While bubble risks are real, several factors distinguish the AI boom from the dot-com era, including profitable core players and the steady use of data centres.
-
Leading companies will continue to consolidate and verticalize, with Microsoft, Google, and Amazon all positioning themselves across the AI stack.
We’re starting to see leaner teams and flatter org charts as AI handles repetitive, lower-value tasks. Large technology firms are already reporting substantial savings from cutting back-office headcount and embedding AI tools. It’s as if every knowledge worker suddenly gained a tireless digital deputy.
More importantly, AI is not just about efficiency; it’s enabling capabilities that weren’t feasible before. Modern models can draft legal documents, debug complex codebases, design campaigns, and reason through multi-step problems. Breakthroughs like GPT–4 opened the gates, and within two years, ChatGPT has scaled to hundreds of millions of weekly users. Enterprise co-pilots are becoming standard across productivity suites, but on this front, we’re only getting started.
Entire industries are being rewired. In retail, AI-driven systems are cutting operating costs and lifting sales through better personalization and inventory accuracy. Call centres, software development shops, banks, logistics operators, and healthcare providers are deploying AI to compress cycle times and raise throughput.
This is not a simple robots-versus-jobs story. History shows that general-purpose technologies change the type of work we do rather than eliminating work entirely. The agricultural revolution pushed labour from farms to factories. The computer pushed work from paper to screens. AI will push work from routine execution toward higher value creativity, judgment, and relationship building.
I believe those who embrace AI as a tool could see their productivity and compensation rise. Those who ignore it risk being left behind.
In that sense, AI is both the steam engine of the mind and the great reorganizer of the workplace. The advantage could belong to those who skate to where the puck is going by embracing AI-enabled productivity.
The CapEx Supercycle
This enthusiasm around AI has triggered the largest technology CapEx cycle in decades. We’re still early in the game, yet CapEx numbers already look like something out of the 1960s space race.
The global hyperscalers, including Microsoft, Amazon, Google, and Meta, were collectively projected to spend more than $450 billion in AI-related CapEx in 2025 alone, according to Evercore ISI Research. That investment has primarily gone into data centres, cutting-edge chips, networking, and supporting systems. Evercore forecasts suggest this could rise to $600 billion by 2026, as nobody wants to be the cloud platform that runs out of AI capacity.
The spend is building the digital factories of the AI age. Tens of billions of dollars are going into new campuses stuffed with advanced semiconductors, dense optical networking, sophisticated cooling, and high-voltage power infrastructure. Semiconductors are at the tip of the spear. Demand for AI accelerators is far ahead of supply.
NVIDIA has been the clear early winner. Its GPUs became the default “picks and shovels” name for AI, driving year-long order backlogs and extraordinary earnings growth. It’s rare to see a $1 trillion company grow at triple-digit rates, supported by gross margins in the 70s and robust free cash flow. That reflects not only silicon dominance but also a software ecosystem that locks developers in.
The CapEx wave is broader than one company. Equipment makers, networking vendors, specialty memory, data centre builders, and integrators are all seeing powerful tailwinds. Cloud providers are also designing their own silicon, such as TPUs at Google and Trainium at Amazon, to diversify supply and improve economics.
The less glamorous bottlenecks may ultimately matter most. AI data centres consume enormous electricity and create intense heat. They already account for a meaningful share of U.S. power demand, with credible scenarios where that share climbs materially by 2030. In several key regions, the local grid, not the chip supply, is becoming the gating factor on new builds. Transformers, switchgear, cooling towers, and backup generation are all needed, which is slowing projects and reshaping where new capacity can be sited.
In aggregate, AI-related infrastructure spend is likely to hit $3–4 trillion globally. Unlike prior manias, this is not being funded primarily with leverage. The megacaps are spending out of enormous cash flows and cash balances. Microsoft alone is pacing toward roughly $80 billion in annual CapEx spending, while still generating strong free cash flow. That self-funding provides a margin of safety versus debt-fueled booms.
With that said, the sheer size and speed of the investment cycle naturally raises the question: Is this the foundation of a new computing era, or are we laying track too fast?
Addressing Bubble Concerns
Any powerful tech boom attracts bubble talk, and AI is no exception. There are real yellow flags that deserve attention.
One concern is circularity. Large cloud platforms are investing heavily in AI startups, which then spend a significant portion of that capital back on cloud and chips from the same platforms. Revenue looks strong, but some of it is essentially capital recycling. The telecom bubble saw something similar with capacity swaps. That ended badly.
Another concern is the gap between investment and visible revenue. Estimates for U.S. AI infrastructure spend in a single year sit in the hundreds of billions of U.S. dollars, while clearly identifiable AI revenue is still a fraction of that. Surveys suggest most enterprises have yet to generate positive financial returns on AI programs. Consumer services like ChatGPT still monetize only a small portion of their user base.
There are also accounting questions. GPUs and related hardware become obsolete quickly, yet some firms are depreciating them over delayed time horizons, which flatter short-term earnings.
These are valid issues, and in smaller, more speculative names, we already see signs of bubble behaviour. However, several factors differentiate this cycle from the dot-com era.
First, the core players are some of the most profitable companies in history, not pre-revenue concepts. Microsoft, Google, Amazon, and NVIDIA are integrating AI into products that customers already pay for and are seeing real revenue uplift. Azure’s growth has re-accelerated with a meaningful contribution from AI workloads on an $80+ billion base. OpenAI is already at a multi-billion-dollar revenue run rate and is growing quickly. These are not speculative hopes; they’re sizable businesses.
Second, the capacity being built is getting used. The constraint over the past two years often wasn't demand, but access to GPUs. Many new data centres are effectively presold. That is very different from the empty fibre networks of the early 2000s.
Third, the productivity and cost-saving benefits are tangible. Companies are cutting costs, collapsing processes, and improving service levels with AI. Internal deployment of co-pilots and agents is allowing large firms to slow hiring or even reduce headcount while supporting more activity. That operating leverage is hard to ignore if you are a CFO.
Finally, the underlying technology is improving at an unparalleled pace. The cost to run a given level of AI capability has fallen roughly 1,000x in just a few years. When the marginal cost of intelligence collapses, entirely new use cases appear. That’s classic Jevons-style dynamics: as a resource like AI becomes more efficient, consumption of the resource rises. I believe this supports the view that this is a general-purpose technology with wide and durable demand, not a narrow fad.
None of this means there will not be drawdowns or disappointments. Valuations for the leaders already embed strong expectations. Some projects will fail, and cycles in CapEx and sentiment are inevitable. But as with the internet and cloud, even if parts of the trade prove bubbly, I believe the infrastructure and platforms built during the boom will power the next decade of digital growth.
The Winners in the Ecosystem: Verticalization and Scale
Looking to 2026 and beyond, leadership is likely to concentrate in players that control multiple layers of the stack.
The cloud giants are positioning themselves as full-stack AI platforms. Microsoft’s partnership with OpenAI has turned into a strategic flywheel, with Azure providing compute, OpenAI providing models, and the Microsoft 365 suite embedding AI deeply in daily workflows. Amazon and Google are following similar playbooks, pairing proprietary silicon and models with massive customer reach and integration.
On top of that, data and software platforms that can sit in the middle of enterprise data estates and orchestrate AI are emerging as important winners. Names like Snowflake and Databricks are evolving from databases into AI application fabrics that connect models, data and business logic. The more data and workflows they centralize, the more valuable they become.
A major shift is also underway in software monetization. As AI agents become capable of performing complex tasks, pricing tied to seats looks increasingly mismatched to value. Expect more usage-based and outcome-based pricing as vendors experiment with charging for tasks completed, insights delivered, or savings generated. Incumbents that already own systems of record and deep integration points should be able to bolt agents on and capture that value. Startups that are AI native will likely push from below.
Ultimately, scale will matter. Those with the most data, compute, distribution, and integration will be able to offer better AI at lower marginal cost, which attracts more users and data, which further improves the product. That feedback loop can produce new forms of concentration, just as we saw in search, social, and cloud.
Wrapping Up…
Heading into 2026, the technology and AI outlook is optimistic, but should be approached with clear eyes. We’re in the early stages of an AI-driven productivity boom that is reshaping business models and capital allocation. A historic CapEx cycle looks to be laying down the rails, power, and silicon for an AI-centric world.
There are genuine concerns around circularity, accounting, and overenthusiasm in pockets of the market, and investors should expect volatility. In my estimate, the core of this cycle is anchored in real products, real revenue, and real efficiency gains at some of the strongest franchises in global markets.
For institutional investors, I believe in staying focused on the secular trend while being selective on exposure. Look at vertically integrated, scale-driven platforms, and the critical infrastructure and data layers that support them. Be wary of pure narrative stories that lack cash flow and competitive moats.
The tide is shifting toward an economy where intelligence is cheap, abundant, and deeply embedded in every workflow. The companies that build, distribute, and harness that intelligence best are likely to drive superior returns through the back half of this decade. Our job is to own them, size them appropriately, and hold our nerve as we climb the next tech wall of worry.
We are the first generation to collaborate with non-human intelligence. How we choose to use that partnership will define the century.
Strong Convictions. Loosely Held.
— Nicholas Mersch, CFA
The content of this document is for informational purposes only and is not being provided in the context of an offering of any securities described herein, nor is it a recommendation or solicitation to buy, hold or sell any security. The information is not investment advice, nor is it tailored to the needs or circumstances of any investor. Information contained in this document is not, and under no circumstances is it to be construed as, an offering memorandum, prospectus, advertisement or public offering of securities. No securities commission or similar regulatory authority has reviewed this document, and any representation to the contrary is an offence. Information contained in this document is believed to be accurate and reliable; however, we cannot guarantee that it is complete or current at all times. The information provided is subject to change without notice.
Commissions, trailing commissions, management fees and expenses all may be associated with investment funds. Please read the prospectus before investing. If the securities are purchased or sold on a stock exchange, you may pay more or receive less than the current net asset value. Investment funds are not guaranteed, their values change frequently, and past performance may not be repeated. Certain statements in this document are forward-looking.
Forward-looking statements (“FLS”) are statements that are predictive in nature, depend on or refer to future events or conditions, or that include words such as “may,” “will,” “should,” “could,” “expect,” “anticipate,” intend,” “plan,” “believe,” “estimate” or other similar expressions. Statements that look forward in time or include anything other than historical information are subject to risks and uncertainties, and actual results, actions or events could differ materially from those set forth in the FLS. FLS are not guarantees of future performance and are, by their nature, based on numerous assumptions. Although the FLS contained in this document are based upon what Purpose Investments and the portfolio manager believe to be reasonable assumptions, Purpose Investments and the portfolio manager cannot assure that actual results will be consistent with these FLS. The reader is cautioned to consider the FLS carefully and not to place undue reliance on the FLS. Unless required by applicable law, it is not undertaken, and specifically disclaimed, that there is any intention or obligation to update or revise FLS, whether as a result of new information, future events or otherwise.