Bessemer Group 13F Portfolio

The biggest adds make their thesis explicit: Bessemer wants to own the capacity build and the challengers, not just the incumbent AI toll collectors.

• Tesla (1.04%, up +3087.9%): A massive step-up, taking the stake from negligible to a core industrial holding. They are buying below their average cost and leaning into EVs, autonomy and grid storage even as sentiment is still polarized.
• AMD (0.95%, up +2438.1%): A huge dollar add of $664.8M signals a bet that the GPU/accelerator duopoly will not remain a one-stock story. They are willing to chase strength here rather than treat it as a late-cycle trade.
• Lam Research (0.79%, up +67.6%) and Applied Materials (1.49%, up +23.0%): Classic “picks and shovels” for the AI and memory upcycle. These adds say Bessemer expects fabs to keep spending on deposition and etch tools well beyond the initial AI hype.
• Intel (0.78%, up +64.3%) and Micron (0.64%, up +35.6%): Backing the recovery of legacy US semiconductor champions, with Micron’s add especially telling given its high gain versus cost; that’s conviction, not averaging down.
• Cisco (0.74%, up +28.2%) and DoorDash (0.39%, up +35.7%): On the edges, they’re quietly upgrading AI-era networking capacity and consumer logistics rails.
• GE Vernova (0.39%, up +146.4%): A sharp increase in a pure-play grid and energy-transition name, reinforcing the electrification side of their bet.

Across these moves, the pattern is consistent: they’re using prior AI windfalls to build exposure to bottleneck assets — compute, memory, tools, networks, and power — that have more cyclical torque if the AI and electrification capex wave continues.

On the funding side, Bessemer is tapping precisely the pockets of the book that have already repriced on AI and rate cuts.

• Visa (0.88%, down -34.6%): The single largest trim by dollars, despite a very strong gain versus cost. This looks less like a call on Visa’s moat and more like an acknowledgment that card rails are ex-growth relative to the AI and industrial opportunities they’ve identified.
• Microsoft (3.15%, down -11.3%), Broadcom (2.40%, down -13.5%), Nvidia (6.59%, down -4.0%) and Alphabet Class C (4.69%, down -4.9%): These are classic profit-taking trims in mega-cap AI and cloud leaders with multi-bagger gains. Importantly, weights remain large; Bessemer is not abandoning the platforms, just right-sizing them.
• JPMorgan (1.83%, down -8.6%) and XPO (0.55%, down -19.9%): Here, they’re reducing cyclical financial and transport exposure that had been strong performers, freeing up capital from more mature or late-cycle plays.
• Keysight (0.37%, down -26.3%): A sizeable cut in a test-and-measurement name, arguably another way of saying they prefer to own direct semiconductor and power equipment exposure rather than upstream tools.
• Energy majors Chevron (0.66%, down -16.2%) and Exxon (0.46%, down -18.5%): Trims suggest they see less upside in old-economy hydrocarbons versus new-economy power and grid names like GE Vernova.

Taken together, the sells read as deliberate: high-gain, highly owned AI winners and mature cash-flow engines are the ATM feeding a more aggressive stance on chips, electrification, and select cyclicals.

Headline sector weights barely budged — technology sits at 55.07% versus 55.09% prior — but the composition of that tech exposure is changing meaningfully.

Within tech, Bessemer has subtly rotated from front-end software and mega-cap platforms toward semiconductors and related capital equipment. Adds in AMD, Intel, Micron, Applied Materials, Lam Research and Cisco stand against trims in Microsoft, Alphabet, Nvidia and Broadcom, implying more sensitivity to capacity cycles and less to pure cloud ad and SaaS growth.

Industrials climbed from 4.58% to 5.90%, driven by the outsized escalation in Tesla and increases in Boeing, while Keysight and Howmet were cut. This is not a generic industrials bet; it is concentrated in aerospace and EVs, i.e., big-ticket beneficiaries of capex and energy transition.

Finance edged down from 8.01% to 7.65%, as JPMorgan and Bank of America were trimmed while Capital One and Citigroup saw modest adds — a slight shift from global wholesale banks toward more credit and consumer finance exposure. Energy fell from 2.40% to 1.94% on the Chevron and Exxon trims, and real-estate-classified Visa was cut enough to drag that bucket from 2.37% to 1.52%.

Unclassified exposure — mostly broad ETFs plus GE Vernova and junior gold miners — ticked up from 5.34% to 5.72%. That combination suggests a barbell: concentrated single-name bets for alpha, plus passive and gold miners as ballast against macro or market shocks.

Bessemer’s 2026-Q2 book reads like a house that believes the easy AI platform money has been made but that the capex and infrastructure wave it unleashed is still in the early innings.

Keeping technology above 55% while pushing harder into semis and equipment says they see AI as a multi-year, not quarter-to-quarter, theme — yet they prefer to own the supply-constrained nodes (compute, memory, tools, networks) instead of just the demand aggregators. The simultaneous build in Tesla and GE Vernova indicates they’re linking AI and data-center growth directly to power, grid and EV demand rather than treating these as disconnected trades.

The trims in Visa, energy majors and some big banks are the tell: capital is coming out of cash-flow stalwarts with limited incremental growth levers and into more cyclical, more volatile names where operating leverage to AI and electrification is maximal. That is a conscious move up the risk spectrum, even if tempered by the steady presence of S&P 500 and EAFE ETFs and a static gold-miners position.

If their thesis is right, the next leg of returns will come less from multiple expansion in already-loved megacaps and more from earnings power in the enablers — the fabs, toolmakers, grid builders and EV complex. If it’s wrong, these adds will have increased the portfolio’s cyclicality exactly as the capex cycle rolls over. The 11.41% latest-quarter gain suggests they’ve had the wind at their back so far; this quarter’s reshuffle shows they’re willing to press that advantage rather than coast on past winners.