Goldman Adds Microsoft, Drops Broadcom and ServiceNow-That 38% Upside Call Is the Real Hook

Goldman's Conviction List rebalance points to a broader market

Goldman Sachs' latest Conviction List change matters because it is a portfolio-timing signal, not a corporate announcement. By adding Microsoft and five other names while removing Broadcom and three others, the bank is signaling a preference for a broader market rather than a narrow AI infrastructure trade.

What changed in the rebalance

The more important detail is not just the rotation itself, but what Goldman still believes about Microsoft. The firm kept its Buy rating on Microsoft stock and maintained a $640 price target, implying nearly 38% upside from current levels. In Goldman's framing, Microsoft is positioned as the AI cycle shifts from training and infrastructure toward enterprise deployment.

That makes the core thesis easier to state: platform companies with enterprise relationships, cloud distribution, and software monetization may be becoming more valuable than pure infrastructure exposures. Microsoft fits that view because Goldman highlighted accelerating Azure growth, improving AI economics, and Copilot monetization, with EPS growth expected to move from 12% in fiscal 2027 to more than 20% by fiscal 2029.

Microsoft's AI case depends on monetization, not just demand

Recent results show scale, and the AI run rate is accelerating

Microsoft's latest reported quarter was revenue of $90.0 billion, up 18%. In the prior quarter, management said its AI business had reached a $37 billion annual revenue run rate, up 123%. Those figures support the idea that AI demand is already feeding into Microsoft's cloud and productivity businesses, not staying confined to headlines helping businesses deploy AI at scale.

Management also said Microsoft Cloud gross margin percentage decreased to 66% as AI investments and usage rose. That does not invalidate the demand story, but it does make the key question clearer: Microsoft needs to convert AI interest into better unit economics, not just higher gross usage.

Why the margin dip matters

Cost of revenue rose 22%, faster than the 18% revenue increase, and Microsoft said that was tied to continued AI infrastructure investment and growing AI product usage. Goldman's bull case still works, but only if today's heavier spend starts producing proportionally stronger earnings over time.

That is the real tension. Demand is already visible. The harder test is whether Microsoft can improve AI economics before investors decide the margin pressure is lasting rather than transitional.

Broadcom's and ServiceNow's exits underline a wider rotation

The broader portfolio signal matters just as much as the Microsoft addition. Goldman added six names and removed four, including Broadcom and ServiceNow. That mix suggests a tilt toward businesses Goldman sees as better positioned for a broader market.

Why the new basket looks wider

Alongside Microsoft, the updated list includes companies tied to travel, logistics, retail resilience, and semiconductor equipment. The common thread is exposure to businesses and consumers that are still spending, not just companies selling the build-out of AI infrastructure.

What the exits likely mean

Dropping Broadcom and ServiceNow does not mean Goldman thinks those companies are broken. A more measured reading is that the bank now prefers broader-market leaders and defensive-growth names with durable cash-flow profiles. For Broadcom, AI enthusiasm may already have pushed expectations higher. For ServiceNow, the market may simply want cleaner execution at this point in the cycle.

What would confirm or challenge Goldman's Microsoft call

The next clear checkpoint is Microsoft's next earnings release. That is the best near-term opportunity to test Goldman's call on Azure acceleration, improving AI unit economics and Copilot monetization.

If those operating trends strengthen, Goldman's rebalance may still look early. If continued investments in AI infrastructure and growing AI product usage keep pressuring margins without a clearer payoff in earnings, the thesis becomes harder to defend.