Marc Frankel arrived at Palantir with a background in finance and found himself sitting beside someone who knew rather more about computers. Daniel Bardenstein had studied symbolic systems at Stanford. Frankel, by his own cheerful account, was borrowing a look at his neighbor’s screen during orientation. They had started on the same day. The patient neighbor became a friend, then a co-founder, then the person who would succeed him as chief executive of Manifest.
It is an agreeable beginning for a company concerned with knowing what is inside things. Before the software inventories, the government contracts, and the AI models, there was two neighboring seats and an admission that somebody else knew the answer. Frankel tells the story without dressing it up as destiny. A decade of friendship followed. The business came later.
A Russian major takes the long way round
Frankel studied Russian Studies at Colgate University, graduating in 2006. A Thomas J. Watson Fellowship sent him out to study isolated Jewish communities in places including Tahiti, Namibia, northern Morocco, rural Portugal, Peru, and Costa Rica. The itinerary belonged to a young researcher following people across a map, rather than a future software executive following customers through a sales funnel.
When he returned to Colgate for a celebration of Jewish life, he recalled fresh tuna for Shabbat in French Polynesia and hiking a Chilean volcano with Israeli backpackers. His gratitude to the university was explicit: the experience there had encouraged him to pursue Jewish life around the world. These were encounters with communities, traditions, and ways of belonging. He had gone to see how people lived.
A separate research trip to Siberia, during his senior year, supplied another lesson he would later describe in writing. Before leaving, he had heard warnings about crime and corruption. Meeting local people changed his understanding. When a graduate-school application asked for something beyond his grades and résumé, he chose that trip. He wanted the admissions committee to understand what seeing a place firsthand had meant to him.
The interviewer who was also applying
At Tufts University’s Fletcher School, Frankel studied international security and earned a Master of Arts in Law and Diplomacy. In 2010, he was playing intramural soccer, participating in the Fletcher Business Club, contributing occasionally to the humor newspaper, the Fletcher Ledger, and volunteering to interview prospective students.
He was also applying to business schools himself. That gave him an unusually useful view of the admissions table: he occupied both sides. His advice to applicants was practical. Be candid. Learn the basics of the program. Arrive prepared. Relax enough to tell your own story. He understood how difficult ordinary common sense could become when someone was judging you.
Later that month, he offered an equally concrete test for an application essay. Remove the prompt, give the essay to a friend, and ask the friend to identify the question. If the answer did not match, revise. It is a pleasingly inexpensive diagnostic. No dashboard required; just a reader who can tell whether you answered what was asked.
From finance to the first day at Palantir
His route into technology ran through a quantitative trading firm and graduate study. At Palantir, he found a way to work on national-security missions without serving in the military. He held leadership positions there and later at Expanse, the cybersecurity company acquired by Palo Alto Networks. His career placed him close to the institutions that would eventually buy Manifest’s software.
The friendship with Bardenstein began before any founder pitch. In a 2024 conversation, Frankel recalled their contrasting preparation for Palantir orientation: his finance background, Bardenstein’s technical education, and the help he received. They stayed in touch through work involving federal civilian agencies, intelligence, and defense. By the time they started Manifest, they had years of experience of each other.
The question that ruined December
In December 2021, the Log4Shell vulnerability exposed a problem that Frankel and his colleagues found hard to accept. Large enterprises, including their own employers, struggled to locate the affected library across their networks and products. Finding the answer meant calls to vendors, messages to engineers, meetings, and spreadsheets. The underlying question was elementary. The work needed to answer it was anything but.
Frankel’s objection extended beyond that incident. After all the searching, organizations could still face the same exercise when another vulnerable component came along. An emergency investigation produced an answer for the emergency. It did not necessarily leave behind a useful, continuing inventory. The next incident could send everyone back to the beginning.
Manifest launched in 2022 to make software bills of materials useful in everyday work. An SBOM is a machine-readable inventory of software components. Frankel saw the next problem immediately: what should an organization do with hundreds or thousands of these files? Where should it keep them? How could their contents guide a purchase or help someone respond?
“Software is the only thing that we buy that we don’t know what’s in it.”
Marc Frankel, Crossing the Valley, 2024
A schematic of the workflow Manifest describes, rather than a performance claim.
A bottle of ketchup enters the argument
Frankel’s explanation of transparency often starts in the grocery aisle. In July 2023, he wrote an essay about the history of food labeling and the objections raised against software ingredient lists. He had been reading about resistance to disclosure in another industry, in another century. The complaints sounded familiar: officials would not understand the information, compliance would cost too much, and ingredients could be proprietary.
He used H. J. Heinz as an example of an industrialist who saw a commercial reason to support transparency. In Frankel’s telling, clear standards could help trustworthy producers earn confidence. The argument gave his software proposal a historical setting. Companies had argued over what buyers deserved to know long before anyone had a Java dependency.
Ketchup is a useful visitor to a cybersecurity discussion. It gives the reader something to hold in mind while the acronyms gather. Frankel uses the comparison to argue for cooperation between government and industry, and for treating disclosure as part of a product’s value. The bottle makes its appearance; the buying decision remains the point.
The customer’s smallest question
In August 2023, Frankel described a government demonstration that had lasted an hour. The team had discussed advanced features. At the end, an analyst asked whether Manifest supported CycloneDX or SPDX, two formats for software bills of materials. Frankel was struck by the assumption embedded in the question: the customer expected to have to choose.
Manifest supported both. His essay announced support for OpenVEX and CSAF VEX too, alongside JSON and XML. He compared the compatibility problem to VHS and Betamax. Whatever the merits of a format, a practitioner still needed to exchange information with a counterpart who could read it.
For a founder, the exchange was useful feedback. A demonstration can be full of features while a customer is still wondering whether the basic plumbing fits. Frankel’s response was to make compatibility explicit. The analyst’s final question had supplied a better starting point for explaining the product.
Taking the inventory abroad, and into government
In December 2023, Frankel announced that Manifest’s interface was available in Japanese and offered translation into any of the European Union’s 24 official languages at no additional cost. He connected the decision to the growth of software-transparency efforts internationally. Users, he argued, needed to understand the words in their security tools.
That same month, an Air Force Research Laboratory innovation contract took Manifest’s work into artificial intelligence bills of materials. The project, undertaken with the Fletcher School, would examine whether AIBOMs could be generated automatically at scale for Air Force components. Frankel wanted organizations to record where models were deployed, what they were trained on, and which software they used before another scramble began.
In December 2024, he announced FedRAMP High authorization, enabled through Palantir’s FedStart program; a January 2025 announcement followed. The milestone allowed Manifest to deliver its inventory-management capabilities in an authorized environment for government customers. His former employer had become a partner in reaching the institutions he had spent years learning to serve.
The black box acquires another compartment
By August 2025, Frankel was writing about executives deploying AI systems without clear knowledge of their origins. Manifest launched its AI Risk product that month. His attention had expanded from software dependencies to model provenance, training data, licenses, and the chains of services connecting AI tools.
The inventory question had become more complicated. A model could arrive with software dependencies as well as its own history of training and modification. Frankel argued for continuing visibility into that combination. A list created once would have limited value if the system changed and nobody kept track.
Earlier that year, Manifest announced a $15 million Series A. The company identified investors including Ensemble VC, AE Industrial Partners, Homebrew, First Round Capital, Overmatch Ventures, and XYZ Venture Capital. The funding was a company milestone under Frankel’s leadership, supporting a business whose remit now included both software and AI transparency.
Passing the chair to the patient neighbor
On October 23, 2025, Manifest announced that Bardenstein would become CEO on November 1. Frankel would remain on the board and become a strategic advisor, continuing work on partnerships, government engagement, and enterprise initiatives. The founding chief executive was changing his responsibilities within the company he had helped build.
In his own announcement, Frankel called growing Manifest the highlight of his professional career. He described Bardenstein as his friend and co-founder, and placed long-term strategy and customer relationships among his continuing interests. Their collaboration, which began at orientation, had reached a new division of labor.
The opening anecdote now has a second scene. The technically prepared neighbor leads the company. Frankel stays involved, bringing the institutional and customer experience accumulated along his winding route. Russian studies, field research, finance, diplomacy, and software have all found a place in that route. The question he helped turn into a business remains wonderfully plain: what have we actually bought?