Canada Pension Plan Investment Board 13F Portfolio

The biggest dollar adds are almost embarrassingly one-sided: the AI stack from silicon to capital equipment. Broadcom alone saw roughly $1.03B of incremental capital as CPP lifted the position +26.6%, while Micron took another about $747.0M with shares up +24.8%.

They didn’t stop at memory and accelerators. Advanced Micro Devices was boosted +28.1%, Nvidia itself still got more capital despite already being a 5.69% anchor, and KLA’s share count jumped +102.0%. Layer that on top of sizable increases in Applied Materials and Lam Research, and you have a clear view: the bet is that fabs and tooling are the real bottleneck — and profit pool — in AI.

Beyond chips, there’s a parallel theme of reinforcing mission-critical platforms. Tesla was increased +23.4% despite only a modest gain vs cost, signaling belief in its long-run industrial and software optionality. JPMorgan, Goldman Sachs, and a pack of Canadian banks (Royal Bank of Canada, Bank of Nova Scotia, Bank of Montreal) all saw double-digit percentage adds, turning financials into a deliberate second pillar behind tech.

Health care quietly joins this inner circle. Eli Lilly (+18.5% shares), Johnson & Johnson (+16.3%), UnitedHealth (+21.3%), and AbbVie (+22.8%) all grew, suggesting CPP is willing to pay for durable cash flows in obesity drugs, specialty pharma, and managed care as a counterweight to cyclical AI sentiment.

If the buys say “AI fabs and banks,” the sells say “we’ve been paid well on reopening and defensives — thank you, next.” Viking Holdings was slashed -44.2% and Live Nation -46.7%, locking in triple-digit percentage gains vs their average costs as discretionary travel and entertainment look fully valued for a cautious allocator.

Constellation Energy stands out as a rare large losing trim: shares cut -34.7% with the position sitting below cost. That looks less like profit-taking and more like a decision that regulated power doesn’t justify capital in a world where every extra dollar can chase AI or high-ROE financials.

Elsewhere, they are pruning around the edges of strong performers. Union Pacific (-33.9%), GE Aerospace (-30.4%), Mastercard (-29.2%), and Linde (-21.0%) all remain material but smaller: classic partial monetizations of winners to fund higher-conviction themes. Digital Realty, Equinix, Amazon, Walmart, Cisco, Canadian Natural Resources, and Bank of America are all modestly downsized, suggesting CPP is tightening exposure to rate-sensitive, cyclical, or fully priced assets rather than exiting them outright.

Sector data makes the strategy explicit. Technology climbs to 54.86% from 51.95%, not through new tickers but by pressing existing winners up and down the stack: Nvidia, Broadcom, Micron, AMD, Intel, KLA, Applied, Lam, Meta, Alphabet, Shopify, and Palantir all got more capital.

Finance nudged up to 11.19% from 10.59% on the back of sizable increases in JPMorgan, Royal Bank of Canada, Bank of Nova Scotia, Bank of Montreal, and Goldman Sachs, partially offset by trims to Canadian Imperial Bank and Bank of America. Health care also grew, to 5.11% from 4.46%, via Lilly, Johnson & Johnson, UnitedHealth, and AbbVie.

The losers in this rebalancing are telling. Consumer discretionary falls to 7.71% from 10.09% after big cuts to Viking, Live Nation, and a smaller trim to Amazon, even as Costco’s stake grows. Real estate drops to 3.70% from 4.71% with reductions in Digital Realty, Equinix, and Mastercard (treated here as a separate “real asset” proxy). Utilities slide to 2.36% from 3.14% on the Constellation cut, and basic materials ticks down as Linde is partially harvested.

Net-net, the portfolio is more barbelled: a dominant AI/tech complex on one end and a thicker sleeve of high-quality banks and pharma on the other, funded by slimming rate-sensitive and discretionary exposures.

This is not a “chase the last AI winner” book; it’s a buildout thesis. The capital is flowing into companies that manufacture chips (Nvidia, AMD, Micron), enable connectivity and custom silicon (Broadcom, Marvell), and sell the lithography and process tools without which the AI story dies (KLA, Applied Materials, Lam Research).

At the same time, CPP is acting like a long-horizon balance-sheet investor, not a momentum fund. It is upgrading defensiveness via large-cap banks on both sides of the border and resilient cash compounders in health care, while shrinking exposure to travel, entertainment, utilities, and some interest-rate-sensitive real estate.

Going forward, expect two things if this playbook holds. First, any pullback in semis or semi-cap is more likely to be met with incremental buying than panic — their adds came with many positions already deeply in the green versus cost. Second, consumer and utility names now look more like funding sources than growth engines: they will probably be trimmed again if AI infrastructure, banks, or pharma offer better forward return per unit of risk.

For readers tracking institutional sentiment, CPP’s 2026-Q2 filing is a clear tell: among the world’s sober mega-pools of capital, the AI boom has graduated from narrative to long-duration capex cycle — and they’re positioning for that, not for a quick trade.