Goldman sees more than another wave of AI spending
Palantir Technologies (NASDAQ: PLTR) is approaching its record high again, but Goldman Sachs thinks the more important move may be happening inside the business.
Goldman analyst Gabriela Borges upgraded Palantir from Neutral to Buy on Thursday and raised her 12-month target to $230, roughly 18% above Wednesday’s close. Shares rose about 3% following the call even as the broader market weakened.
Another bullish analyst call is not especially remarkable for one of AI’s best-known stocks. Goldman’s reasoning is.
Rather than assuming Palantir simply wins because companies spend more on AI, Borges argues its addressable market may be undergoing another significant expansion. Sovereign AI, bespoke applications and a more focused industry strategy are giving Palantir more ways to become embedded inside the organizations already using its software.
That creates a different growth model. Palantir does not necessarily need thousands of new customers buying the same standardized product. Increasing the number of critical workflows running through its platform can make each existing relationship considerably larger.
Palantir’s unusual sales model may be becoming its moat
Palantir has always approached enterprise software differently.
Instead of handing customers a software package and leaving implementation largely to internal IT teams, Palantir deploys engineers directly into organizations. Those forward-deployed engineers work alongside customers, connect fragmented data and build applications around real operational problems.
The approach can look expensive compared with traditional software because it requires highly skilled people in the field. Yet Goldman believes Palantir has built an unusually effective feedback loop between those engineers and its product teams. Problems discovered inside one customer can eventually become tools that solve similar problems elsewhere.
AI agents could make that system far more scalable.
An engineer who once had to manually configure large portions of an application can increasingly use Palantir’s AI tools to automate parts of the process. Better automation means the company can potentially deliver more custom software without growing its engineering workforce at the same rate.
Competitors appear to recognize the appeal. Microsoft (NASDAQ: MSFT), Salesforce (NYSE: CRM) and Snowflake (NYSE: SNOW) have all been building out field-oriented engineering teams of their own.
When large software companies begin copying a model that once looked unusually labor-intensive, it is worth asking whether Palantir figured something out early.
Sovereign AI could open an entirely different market
Corporate AI gets most of the attention, but Palantir’s government roots may become increasingly valuable as countries build their own AI infrastructure.
Governments are becoming more sensitive to where data is stored, which models process it and who controls the underlying technology. Defense, intelligence, healthcare and other sensitive workloads cannot always be pushed into a generic public AI platform.
Palantir already operates in environments where data security, access controls and mission-specific applications are fundamental requirements.
Sovereign AI extends those same needs into a much larger market. Countries building national AI capabilities may want models and computing infrastructure, but they also need software capable of connecting those systems to sensitive government data without giving up control.
Goldman’s thesis suggests Palantir can occupy that layer between raw AI capability and actual government operations.
Verticalization creates a similar opportunity commercially. Healthcare, manufacturing, energy, financial services and defense each operate differently. Generic AI tools can answer questions, while Palantir is trying to build applications capable of making AI useful inside the specific workflows that run those industries.
The distinction matters because software becomes much harder to replace once it is tied directly into how an organization makes decisions and operates day to day.
A record stock price raises the standard
Palantir’s opportunity may be expanding, but investors are hardly discovering the company at a bargain valuation.
Shares are trading just below their previous record above $207, with Goldman arguing another move toward $230 is possible. The stock’s valuation already assumes exceptional growth, leaving less room for disappointing execution than most enterprise software companies.
Goldman’s upgrade therefore puts an important question at the center of the next phase: can Palantir scale its highly customized model without losing the economics that made software businesses attractive in the first place?
Watch U.S. commercial growth, expansion inside existing customers and the efficiency of the forward-deployed engineering model. If AI agents allow Palantir to build increasingly sophisticated applications without a proportional increase in labor, margins and revenue growth could improve together.
Failure to achieve that leverage would revive one of the oldest criticisms of the company, that highly customized deployments eventually start resembling consulting rather than scalable software.
Goldman is betting Palantir has already moved beyond that problem.
The next chapter of Palantir’s growth may depend less on selling AI to more companies and more on becoming increasingly difficult to remove from the companies and governments already using it.
That would make the addressable market much larger without requiring Palantir to reinvent what made it successful in the first place.
