Piper Sandler names 5 software stocks cutting AI token costs
Piper Sandler analyst Rob Owens told clients that five infrastructure software firms are positioned to reduce AI token costs by using proprietary customer data to cut token usage for AI agents. The note named Elastic (ESTC), GitLab (GTLB), MongoDB (MDB), Snowflake (SNOW), and Atlassian (TEAM), citing early 50% to 75% token savings and consumption-based pricing.
How this was made
The 30-second read
Why it matters
Piper Sandler argues that stored customer data and context layers can reduce AI agent token usage, enabling consumption-based revenue growth even if token unit prices fall. The article also highlights Snowflake’s separate AI consumption meter and the broader shift away from “Tokenmaxxing” toward model routing.
Market read
This is a multi-stock analyst thesis that could influence positioning in enterprise infrastructure software, but it is not a new company disclosure; traders will likely wait for earnings consumption metrics to confirm.
What to watch
The 50% to 75% token savings are early use-case figures, and none of the five companies discloses context-layer revenue as a separate line item, so validation relies on analyst estimates and management commentary.
Background
Enterprise software has lagged as investors assumed LLMs would eventually make the category redundant, while token pricing and consumption economics have been shifting.
Ticker impact
Piper Sandler’s note names Elastic as a primary beneficiary of using stored customer data to cut AI agent token costs.
Mild positive bias into the next earnings cycle as investors look for consumption metrics and net revenue retention tied to AI usage.
The article is an analyst thesis with specific token-cost mechanism, but it does not provide new Elastic financial disclosures or audited results.
GitLab is included in Piper Sandler’s five-name list positioned to reduce AI agent token costs using proprietary customer data.
Limited near-term upside unless management confirms context-layer-driven consumption growth.
The article provides no new GitLab datapoint beyond analyst framing and mentions it has been the weakest of the five.
MongoDB is cited as a primary beneficiary, with early deployments reportedly cutting token usage by 50% to 75% when clean organizational context is provided.
Potentially stronger relative performance versus peers if upcoming reports show consumption metrics and improving net revenue retention.
The article includes a standout performance reference for MongoDB and a concrete token-savings range, but still lacks audited, company-specific revenue disclosure.
Snowflake is named, including discussion that it splits AI usage onto a separate consumption meter so customers can track token spend.
Moderate positive bias into earnings as investors verify whether context layers produce revenue via consumption.
The article ties Snowflake’s pricing mechanics to the token-cost thesis, but does not disclose new Snowflake results in this piece.
Atlassian (TEAM) is included on the list, with the thesis that context layers can reduce token usage costs and support consumption-based growth.
Neutral to slightly positive, contingent on management commentary and consumption metrics in the next earnings cycle.
The article flags TEAM as more seat-dependent and notes no context-layer revenue line item, increasing execution uncertainty.
Market effects
Reframes enterprise software AI monetization from per-seat risk to consumption and context-layer efficiency, potentially shifting valuation focus across infrastructure software.
Primarily US-listed software names; no explicit regional catalyst beyond sector sentiment.
Supports a global enterprise AI deployment narrative where token economics and retrieval/context layers influence vendor demand.
Counterpoint
Frontier model providers may build retrieval and memory directly, reducing the need for third-party context layers and limiting monetization for these vendors.
Key entities
- analyst_firmPiper Sandler
Issued a client note naming five infrastructure software stocks as beneficiaries of lower AI token costs via context layers.
- analystRob Owens
Led the note and provided the token-cost mechanism and validation checklist for upcoming earnings.
- software_companyElastic
Named as a primary beneficiary in the context-layer token-cost thesis.
- software_companyGitLab
Named as a primary beneficiary in the context-layer token-cost thesis.
- software_companyMongoDB
Named as a primary beneficiary, with early deployments cited for large token usage reductions.
