Cantor Fitzgerald L P Rotates From Crowded AI Trades Into Broader Beta

Rising conviction clusters around three themes: scalable AI/software platforms, crypto ecosystem picks, and broad market beta.

On the AI/software side, Cantor is clearly shifting away from only owning the chips to owning the workflows sitting on top of them:

• AMD: A 61.4% add and a 5.54% position say they like the challenger’s upside versus more fully priced peers.
• ServiceNow (NOW): A massive 626.6% increase to 2.26% of the book is a loud statement that AI-driven workflow software is a core pillar, even though the position sits -30.5% versus their average cost.
• Microsoft (MSFT), Adobe (ADBE), Salesforce (CRM), Palantir (PLTR), and AppLovin (APP): Broad-based increases across these names — from an 892.3% ramp in MSFT to a new 1.68% stake in AppLovin — show a tilt toward platforms that monetize AI via productivity, data, and advertising rather than just silicon.

They also push deeper into crypto-adjacent and digital trading infrastructure:

• BMNR and HOOD: A 123.1% ramp in BMNR and 25.0% add in Robinhood, both still under water vs cost, reveal they’re averaging into the weaker parts of the ecosystem instead of cutting and running.
• Circle (CRCL) and modest adds to IBIT sit alongside a steady FBTC stake, reinforcing that this is an intentional crypto/Bitcoin complex, not a legacy mistake.

Finally, IVV is a new 3.25% anchor position — nearly as large as AMZN. That’s a clear decision to buy broad market earnings at scale after a drawdown, using the index as the liquidity and volatility release valve for an otherwise high-octane book.

The funding leg of this rotation is unambiguous: they are monetizing their AI hardware and speculative space/energy wins, and recycling that cash into software and beta.

The biggest trims are a who’s-who of 2024–2025 AI darlings:

• NVIDIA (NVDA): Still the top holding at 11.24%, but a -40.7% cut freed up about $213.1M at a hefty 74.2% gain vs cost. That’s classic risk management: stay exposed, stop letting one name own the portfolio.
• MicroStrategy (MSTR): A -25.0% trim after a -26.1% mark versus their cost looks less like victory, more like risk control on a volatile, Bitcoin-geared position.
• Intel (INTC) and Broadcom (AVGO): Big reductions of -38.8% and -23.7% respectively, both at strong gains, signal a willingness to harvest profits across the semiconductor stack, not just NVDA.

They are also reining in some of their more speculative or earlier-stage tech:

• Rocket Lab (RKLB) and IonQ (IONQ): Cuts of -66.3% and -45.2% are telling; both had delivered striking percentage gains from cost in RKLB’s case and near-flat economics in IONQ. These were the optionality chips they’re now cashing to fund higher-conviction areas.
• Satellogic (SATL), OKLO, NVO, and HIMS: Broad trims across space imaging, advanced nuclear, GLP-1, and telehealth imply less patience for long-duration, loss-making or politically sensitive themes when the core AI and crypto complex already drives plenty of risk.

The net effect: they’re trading down the hype curve – less frontier hardware and space, more scaled platforms and liquid benchmarks – without abandoning growth altogether.

On a sector level, Cantor is delicately deconcentrating. Technology’s share falls from 68.93% to 61.1%, but that 7.8-point drop hides a more meaningful internal reshuffle from semis and speculative hardware toward cloud, workflow, and consumer internet platforms.

Finance climbs from 10.61% to 13.36%, driven by BMNR, HOOD, CRCL and SPAC exposure (IACO, CGCT). This is effectively a levered bet on trading volumes, alternative assets, and capital markets complexity.

The “Unclassified” bucket — structurally ETFs and macro tools — jumps from 4.72% to 8.6% via IVV, BRK.B, and GDX. That’s how they’re buying macro optionality and liquidity without giving up stock-picking freedom elsewhere.

Consumer exposure edges up from 4.9% to 6.18% with adds to AMZN and a new WMT stake, a straightforward expression of resilient U.S. consumption. Health care nudges up to 5.72% via adds in UNH and LLY even as they slash NVO, signalling selective belief in healthcare earnings durability rather than a blind GLP-1 chase.

Industrials drop hard from 3.55% to 1.88% after the RKLB trim and a modestly sized BA position, while Energy goes from 0% to 1.61% on the new SHEL stake. In other words, they’re engineering a barbell: AI/crypto on one side, energy and gold on the other, with S&P 500 beta in the middle.

For a fund with a mixed 5-year weighted track record and a -12.15% latest quarter, this portfolio looks like a manager who got the big secular call (AI and crypto) mostly right but is now admitting that concentration and timing cuts both ways.

The new shape of the book suggests several forward-looking bets:

• AI remains the spine of the strategy. NVDA, AMD, AVGO, ARM, MSFT, NOW, and a suite of data/software names show they still expect AI capex and monetization to compound — just with less reliance on any single ticker.
• Crypto and trading infrastructure aren’t going away. Maintaining FBTC, IBIT, CRCL, GCMG, BMNR, HOOD, CME, and even SPACs argues they see volatility and alternative assets as structural, not cyclical, features of the market.
• Macro hedging is now part of the toolkit. The step-up in IVV, GDX, AU, SHEL, and BRK.B points to a manager who wants to survive more than one scenario for inflation, growth, and rates.

If this pattern persists, expect Cantor Fitzgerald L P to keep using strength in headline AI and speculative names as a funding source, recycling into scalable software, fintech, and macro hedges. The portfolio is still aggressive and thematically narrow by sector, but it is no longer all-or-nothing on a handful of AI chips and moonshots, which should modestly reduce drawdown risk even if it caps some upside torque.