5 big analyst AI moves: Upgrades for TXN and DDOG; Further upside seen in SNDK

Analyst moves were mixed across AI and semis, with Bernstein raising SanDisk’s target to $1,250 from $1,000 and outlining a $3,000 blue-sky case on strong NAND pricing and FY2027 EPS of $224. Stifel upgraded Texas Instruments to Buy with a $250 target on improving free cash flow, while UBS cut ServiceNow to Neutral and Guggenheim downgraded GitLab to Neutral on AI disruption and budget pressure. Guggenheim also upgraded Datadog to Buy, forecasting 27% 2026 revenue growth to $4.36B, helped by Anthropic and broader AI-native demand.

Analysis

The clearest second-order read is that AI infrastructure spend is now becoming a redistribution game inside software and semis, not just an incremental growth story. Datadog and Texas Instruments look like beneficiaries of AI capex intensity rising faster than general IT budgets, while ServiceNow and GitLab are the canaries for where that budget is coming from: workflow and developer seats are the easiest line items to pressure when CFOs need funding for model training, inference, and power/compute. That makes the current market reaction less about near-term beats/misses and more about which vendors can prove they are attached to non-optional infrastructure rather than discretionary software usage.

On the semiconductor side, SanDisk’s setup suggests the market is still underpricing the duration of the NAND cycle, but the real opportunity may be in the read-through to the broader storage/power chain rather than the single name. If memory pricing stays firm into mid-year, expect contract renegotiations and capex discipline across adjacent suppliers, which can extend margins for the group even if unit growth normalizes. Texas Instruments is interesting because AI data center buildout creates a bottleneck in power management before it becomes visible in revenue; that usually gives the analog complex a lagged but more durable earnings uplift than the headline AI GPU names.

The contrarian risk is that the market is currently treating AI spend as structurally secular, while end customers may still force a budget reallocation cycle in the next 2-4 quarters. That is the key asymmetry: infrastructure names can keep compounding even if application software multiple compression continues, because the winners are selling capacity-constrained inputs; the losers are selling seats and workflow layers that can be partially substituted by agentic tools. If that narrative hardens, ServiceNow and GitLab could see multiple and estimate compression well before revenue growth breaks, making them cleaner shorts than the broader software index.

Key Decisions for Investors

• Long DDOG vs short NOW for the next 3-6 months: DDOG has visible AI-native customer expansion and lower budget sensitivity; NOW is exposed to enterprise software spending pressure. Target a 2:1 upside/downside skew, with NOW downside likely front-loaded if budget anecdotes turn into guidance cuts.
• Buy TXN on pullbacks over the next 4-8 weeks ahead of the next two earnings cycles: AI-related power management demand plus capex normalization should support a multi-quarter FCF inflection. Use a medium-term thesis with a 15-20% upside case if the market starts capitalizing 2027-28 FCF earlier.
• Maintain a tactical long SNDK position into the April 30 print, but hedge with short-dated downside protection: the setup is favorable if NAND pricing stays firm, yet the stock is vulnerable to any sign that contract pricing is peaking. Structure as stock plus puts or a call spread to preserve convexity.
• Avoid GTLB until there is evidence that AI coding tools are monetized rather than merely defended against: the risk/reward is asymmetric to the downside over the next 2 quarters because seat growth can decelerate before headline ARR growth breaks.
• Watch for a pair trade long semicap/short application software basket (TXN, DDOG, SNDK vs NOW, GTLB): this isolates the AI-capex beneficiaries from the budget-pressure losers and should work if enterprise spending remains constrained through mid-year.