When Mark Baribeau talks about artificial intelligence, he does not sound like someone debating a theoretical future. For the head of global growth equities at Jennison Associates, AI is already reshaping capital flows, business models, and the competitive landscape at a pace he describes as “happening at the speed of light”.
Baribeau was one of the speakers at Curate’s first annual global investor day in South Africa, which was held in Cape Town last week. The event explored the global investment landscape and the forces shaping markets.
Curate is an independent global fund manager that specialises in offering a “curated” range of unit trusts by partnering with top-tier specialist investment managers rather than managing funds directly. The independent manager identifies specific investor needs and appoints expert managers to run those portfolios.
Who is Jennison – and why they ‘fish in the growth pool’
Baribeau describes Jennison as a pure growth-equity house. The firm is “a leader in growth equities in the United States”, he explains, with about $150 billion in growth assets overall and roughly $30bn in its global growth suite.
Jennison’s mandate is unconstrained: the global team “can go anywhere in the world” and is “pretty agnostic” about regions, countries, or sectors as long as they can find “real market leaders”.
What defines those leaders is consistent:
- They tend to be emerging market leaders with a unique business model.
- They are often founder-led, entrepreneurial, and disruptive.
- They are driving or benefiting from structural change in their industries.
- They build “really strong competitive advantages”.
Critically, they also exhibit “far superior unit economics”: every incremental unit of revenue drives unusually high free cash flow over time, even if reported earnings look depressed because the companies are reinvesting heavily to grow.
This is why Jennison deliberately fishes in the growth pool. Historically, the top quintile of earnings growers in the global market has delivered returns of about 16.6% a year, with the second-highest growth quintile also compounding at strong double digits.
Baribeau describes those tiers – the hyper-growth names and the more stable compounders – as the zones in which Jennison wants to live, typically with roughly two-thirds of the portfolio in the highest growers and one-third in the second tier.
In other words, the firm’s philosophy is simple: sustained growth drives long-term equity returns, and the job is to identify the businesses capable of delivering that growth before the market fully prices it in.
Phase one: massive AI cash outlays – and why he thinks they’re justified
From that vantage point, AI is not merely another theme; it is “the biggest technology shift in a generation”, one that Baribeau believes will “change everything we do, how we work, how we live… how we get access to information, how we buy things”.
He breaks the AI cycle into phases. Currently, investors are in Phase One: the build-out of data centres and accelerated computing infrastructure required for generative AI. That means:
- GPU suppliers such as Nvidia;
- Networking vendors;
- Cloud hyperscalers such as Microsoft Azure, Amazon Web Services, and Google Cloud; and
- Power infrastructure, given the enormous electricity needs of AI data centres.
The headline that unnerves many investors is the sheer scale of capital expenditure.
Capex in the AI industry is experiencing an unprecedented, massive surge, driven by hyperscalers and specialised firms building out infrastructure for AI model training and inferencing. These outlays are characterised by multi-billion-dollar investments in data centres, specialised computing hardware (GPUs/chips), and energy infrastructure.
By Baribeau’s estimate, the big four US hyperscalers – Microsoft, Amazon, Google, and Meta – are on track to spend about $650bn this year on Capex, an increase of about 80% year-on-year.
He acknowledges that “the market is very sceptical” of this level of spending, noting he has “never seen anything more controversial” in terms of technology debates.
Yet he argues that this time is different for one key reason: these Capex outlays are largely funded from cash flow, not speculative leverage. The “richest companies in the world” are paying for this build-out from their own operating cash, rather than by piling on debt.
From his perspective, the Capex is demand-driven, not speculative. He notes that AI compute usage – measured in tokens, the basic unit of AI computation – is “absolutely exploding”. For example, Google Cloud’s AI token volumes have grown twenty-fold in a year, and major cloud providers are effectively sold out of AI capacity. In his words, “Google’s cloud can’t grow faster than it is right now… AWS (Amazon Web Services) can’t grow faster… they’re sold out.”
That is why, he insists, “we’re not that frightened by the size of the Capex”. In his view, the hyperscalers “are spending this money because demand is so strong they can’t keep up with it”, and the market has that dynamic “dead wrong”.
Phases two and three: where Baribeau sees the real AI opportunity
If Phase One belongs to the infrastructure builders, Phase Two and Phase Three are where Baribeau sees the most compelling long-term equity stories.
Phase Two is the application layer – “the interesting one”, as he puts it. Private companies such as OpenAI and Anthropic are already, he says, among the fastest businesses in history to reach multi-billion-dollar revenue run-rates.
In a panel discussion, he pointed out that Anthropic is already at about a $14bn annual run-rate, roughly ten times the prior year, while OpenAI is near a $20bn run-rate, mostly on subscriptions – even though “most of their billion-person user base is free”.
He sees an oncoming wave of AI-native application businesses: search, AI agents, software development and coding tools, workflow and productivity tools, and vertical specialists in areas such as customer service or healthcare.
Many remain private today; Baribeau says Jennison is already evaluating them in anticipation of future initial public offerings.
Phase Three is AI “on the edge” – bringing intelligence directly onto devices such as smartphones, cars, and robots. He anticipates that smartphones will function as true personal digital assistants as soon as this year, driven by new software layers running on top of those large language models.
In all of this, Baribeau expects only a small handful of foundational large language models (LLMs) to dominate. He predicts that within a couple of years, there will be just “a couple of large language models” that matter, and that today Google’s Gemini and OpenAI’s ChatGPT already hold the lion’s share of that market.
The contest between them, in his view, will be determined less by the underlying model and more by how effectively they monetize through differentiated products and services.
Around those platforms, however, he expects a wide-open field for specialist AI application companies, particularly those attacking specific verticals such as customer service, healthcare, or drug discovery.
Risks: winners, losers, and the need for agility
For all his optimism, Baribeau is explicit that investors face a “horse race” dynamic in AI. In the panel session, he argued that although monetisation is clearly under way for leaders such as OpenAI and Anthropic, “there’ll be winners and losers”, and “not everybody’s going to win the horse race”.
He sees stock-selection risk rather than systemic risk as the central challenge. Some heavily funded efforts, he suggests, will quietly disappear. As he quipped, “X.AI is already out of the race because Elon [Musk] is stuffing it into SpaceX, so nobody can see it because they’re not winning.”
That is why he emphasises this is an environment in which investors “have to be very flexible and dynamic”, warning that “things are going to change a lot in the next few years”, with market leadership rotating rapidly.
On the infrastructure side, he acknowledges concerns about power and capacity but treats them as bottlenecks, not roadblocks. Electricity demand for AI data centres is surging, and he notes multi-year backlogs for leading-edge power equipment, but he frames this as a manufacturing and implementation challenge – “not a physical problem” of running out of energy.
What this means for financial advisers and their clients
The implications for advisers are two-fold: how markets behave, and how clients behave.
On the market side, Baribeau believes AI will intensify the dispersion between winners and losers, favouring truly innovative, competitively advantaged growth businesses. For growth investors, that creates opportunity – but only if they are willing to continually reassess where durable growth and economic moats reside, and to sell quickly when a growth thesis breaks.
On the client side, he expects AI to commoditise information and many routine computational tasks. LLMs, he points out, “embody every known information” in a way that far exceeds any individual’s knowledge, and he expects clients to have ubiquitous, low-cost access to sophisticated frameworks and answers via tools such as ChatGPT.
That does not mean advisers become obsolete. Baribeau argues that although clients will have better information, they will still struggle to:
- Frame the right questions;
- Interpret the trade-offs; and
- Execute a coherent long-term plan.
He believes “humans want interaction with humans”, noting that the failure of the metaverse was, in part, a reminder that people do not want to live in virtual environments without meaningful human contact.
In his view, the adviser’s “competitive edge” will increasingly lie in trust, judgement, and synthesis: taking commoditised information and turning it into personalised, behaviourally grounded advice that clients can stick with across cycles.
At the same time, he is blunt that advisers cannot opt out of AI. As access to tools becomes “ubiquitous and homogenous”, he argues, advisers will need to “be on top of that and be able to distil, in a very simple way”, what these technologies imply for portfolios and financial plans.
Those who refuse to engage may, find that opting out is technically possible – but commercially untenable.
A landscape moving ‘at the speed of light’
For Baribeau himself, keeping up with AI is already a full-time intellectual sprint. He describes his own routine as a “constant barrage of information”, requiring him to stay “very, very current on everything that’s happening in the industry, how it’s developing, who’s taking a lead, who might be falling behind”.
That pace, he believes, is the new normal. The combination of hyperscaler cash-funded Capex, explosive demand for compute, and an emerging ecosystem of AI-native applications suggests to him that AI is not a bubble waiting to burst, but a foundational shift that will compound over years – albeit with all the usual cycles of over-exuberance, disappointment, and selective failure that define past technology revolutions.
For financial advisers and their clients, the implications are becoming increasingly evident: AI will not only reshape which companies are worth owning, but also transform how clients access information, how advice is delivered, and what ultimately defines real value. Those who adapt early, Baribeau implies, will find that AI amplifies their ability to serve clients. Those who wait risk discovering, too late, that the language of money is increasingly being spoken in only a handful of AI “languages” – and that their clients have been listening to them for years.
Disclaimer: The information in this article does not constitute investment or financial planning advice.
