What if a company could predict crimes before they happen?
Palantir, a data mining company founded in 2004 by Peter Thiel, has sought to do just that, taking an early investment from the CIA’s VC fund In-Q-Tel and working with government agencies from the CIA, FBI, and the Department of Defense to the Department of Justice and the Securities and Exchange Commission. The company focuses on combining massive troves of information from both public and proprietary sources and simplifying the integration and analysis process, allowing the people using the system to use their judgment to answer pressing problems quickly. That has included predicting the locations of improvised explosive devices in Afghanistan and Iraq, and finding Osama bin Laden.
Since its start working in intelligence and national security, Palantir has sought to expand its business in the US, looking toward both the private sector and state and local governments for growth opportunities. Its work with the New Orleans Police Department came to light recently in an investigative piece by the Verge, which found that Palantir had contracted to do predictive policing work there through a nonprofit that avoided the city’s standard procurement rules. While the capabilities Palantir has developed on that front could lead to good outcomes—James Carville claimed that the company’s work reduced the murder rate by a third—it could also have significant unintended consequences if either the technology, its programmers, or the analysts who use it enter bias into the process. A 2016 study of a similar predictive policing program in California found that the data used to train the machine learning algorithms replicated systemic bias, meaning that “predictive policing of drug crimes results in increasingly disproportionate policing of historically over-policed communities.”
While caution is certainly necessary and better care and oversight from government users of Palantir’s services are warranted, it’s clear that Palantir has built a powerful set of capabilities—which has been valued at $20 billion.
Palantir’s strategy of working through government is a major differentiator. While government contracts can be notoriously hard to get at first, they can also be very sticky, with solid long-term revenue from agencies that are generally loath to switch vendors, and the credibility that can come from solving tough problems. That model can be hard to translate to the corporate world, though. In 2016, Buzzfeed reported that Palantir was having trouble retaining corporate clients at its high rates of sometimes over $1 million per month, losing several including Home Depot as the companies failed to see value after the initial implementation. But Palantir’s clear ownership of the intelligence market, its growing presence with state and local governments, and the impressive array of financial services and corporate clients it has amassed, from JPMorgan Chase and Morgan Stanley to American Express and Walmart, speak to the company’s ability to make the transition out of government contracts into a more diversified set of offerings and clients.
As the company grows, it will need to face the increased pressure and scrutiny of a more visible company, as well as the challenges of a mission-focused startup that grows and expands its work. The company was historically able to pay below market rates for talent as a result of its ability to pitch the impact of government work. In 2016, however, increased turnover led the company to offer a 20% raise to all employees who had been there longer than 18 months, indicating a softening in the draw of that mission-for-money tradeoff. Palantir has also benefited from an ability to fly under the radar with many of its projects by staying private, but as it nears its 15th anniversary and is among the highest valued start-ups in the country, its investors are likely to push it to go public.
 William Alden, “Inside Palantir, Silicon Valley’s Most Secretive Company,” Buzzfeed, May 6, 2016, accessed April 9 2018, https://www.buzzfeed.com/williamalden/inside-palantir-silicon-valleys-most-secretive-company?utm_term=.ej8b7m41V7#.jrnaVDYWeV
 Andy Greenberg, “How A ‘Deviant’ Philosopher Built Palantir, A CIA-Funded Data-Mining Juggernaut,” Forbes, Aug 14, 2013, accessed April 9, 2018, https://www.forbes.com/sites/andygreenberg/2013/08/14/agent-of-intelligence-how-a-deviant-philosopher-built-palantir-a-cia-funded-data-mining-juggernaut/#679ba0ff7785
 Kristian Lum and William Isaac, “To predict and serve?” Significance Magazine, October 2016, accessed April 9, 2018, https://rss.onlinelibrary.wiley.com/doi/pdf/10.1111/j.1740-9713.2016.00960.x
 Ingrid Lunden, “Big data company Palantir quietly raised another $20M in November,” Tech Cruch, Nov 24, 2016, accessed April 9, 2018, https://techcrunch.com/2016/11/24/big-data-company-palantir-quietly-raised-another-20m-in-november/