Truist reiterates Buy on Microsoft stock, $675 target amid OpenAI shift

Microsoft’s revised OpenAI partnership gives the company six-year IP control and more flexibility, while limiting access to future OpenAI technology. Analyst views remain mixed: Truist kept a Buy/$675 target, Wedbush and Evercore reiterated Outperform with $575/$580 targets, while Oppenheimer cut its target to $515 from $630. The main market focus is whether Microsoft’s 20% revenue-share cap with OpenAI will be reached sooner as OpenAI’s revenue scales rapidly and competitors like Oracle, Google Cloud, and AWS gain more of the AI workload.

Analysis

The immediate market read-through is not really about Microsoft; it is about the AI infrastructure stack becoming more modular, which lowers the cost of switching among clouds and weakens the “single-winner” Azure narrative. That is constructive for adoption, but it also means the largest hyperscalers are moving from exclusivity rents toward a more competitive, capacity-constrained utility model, where power, chips, and networking become the true bottlenecks rather than model access alone.

For NVDA, the second-order effect is mixed but ultimately supportive over the medium term. More cloud optionality for frontier AI customers likely increases aggregate compute consumption because it removes deployment friction, but it also raises the risk that inference and workflow spend get optimized across vendors, pressuring gross margin capture at the platform layer while amplifying demand for accelerators across all clouds. In other words, this is less a Microsoft-specific loss than a redistribution of AI capex across multiple balance sheets.

The underappreciated risk is timeline. The next 1-2 quarters could show headline noise around OpenAI-related revenue contributions and partner mix, but the larger catalyst path is 6-18 months: whether Microsoft’s in-house models materially reduce external dependency, and whether Oracle/Google/Amazon can actually monetize incremental AI workloads without destroying price discipline. If investors conclude that the AI stack is becoming commoditized faster than usage is expanding, multiples for the whole basket can compress even as unit volumes rise.

Consensus may be underpricing the durability of demand, but overpricing the durability of moat. The market is likely right that this change reduces near-term key-man risk for Microsoft; it is less clear that it improves the long-run economics of any one hyperscaler. The best setup is probably relative value rather than outright directional exposure: the winners are the players with the cheapest incremental power and the strongest distribution, not necessarily the ones with the best model relationships.

Key Decisions for Investors

• Maintain/trim short-dated MSFT upside exposure into event-driven strength; the immediate de-risking is probably already in the tape, while the real downside only emerges if Azure AI share or margin commentary softens over the next 1-2 earnings cycles.
• Go long NVDA versus MSFT on a 3-6 month horizon: the deal structure is incrementally bullish for total AI compute spend, but MSFT’s relative narrative loses exclusivity premium while NVDA benefits from broader multi-cloud deployment.
• Initiate a paired long ORCL / short AMZN or GOOG trade for 6-12 months: the market may continue rewarding “AI capacity optionality,” but Oracle has the cleanest incremental monetization story if multi-cloud AI workflows expand without a commensurate capex arms race.
• Buy 6-12 month MSFT downside protection via put spreads if positioned long: the key risk is not immediate revenue loss but multiple compression if investors start discounting AI commoditization and competitive pressure on M365/Azure economics.
• Watch for confirmation in hyperscaler capex guidance over the next two quarters; if capex re-accelerates while cloud mix broadens, add to NVDA and reduce exposure to the weakest cloud-margin names.