a16z's First Employee on the History of a16z, Venture Capital & Fixing the Government | Scott Kupor
Pioneers
Scott Kupor
Summary
- a16z built its differentiation by turning an investor’s personal network into a service the whole firm could deliver. Kupor describes the ambition as “institutionalize John Doerr”: surround investing partners with specialists in recruiting, business development and government relations. Marketing that offering let a newcomer compete without an established investment record, while the staff grew from three to nearly 100 in the first year.
- Skype’s rapid exit financed the service model and accelerated a16z’s fundraising. Despite pitching early-stage venture, the firm joined a buyout of Skype, then sold to Microsoft 18 months later for four times what it paid. Kupor says the transaction returned half the fund to LPs; the second fund supplied additional fees needed to keep hiring. “Momentum begets momentum.”
- Scaling a16z eventually required dismantling the centralized organization Kupor had built. Around ten years and roughly 500 employees in, investing teams and supporting services moved into specialized verticals. Kupor initially resisted losing scope, then accepted that “the needs of the business have changed”; his own investing shifted toward later-stage healthcare, where operating performance mattered more than technical judgments he felt unqualified to make.
- Kupor sees government performance as an incentive problem that recruiting alone cannot solve. He reports that 99.7% of employees meet expectations and roughly 90% of senior executives receive the top two ratings, leaving little differentiation in accountability or bonuses. His response combines “permissioned risk,” a 40% ceiling on top ratings, and directing at least 60% of the compensation pool toward that group.
- The federal workforce’s age profile makes a stronger early-career proposition urgent. Kupor says only 7% are under 30 versus 23% across the US workforce, while 44% are over 50. Tech Force aims to recruit 1,000 technical specialists for two-year terms, with skills assessments and development from 40 private-company partners: “do something that’s actually good for yourself,” alongside serving the country.
- Kupor argues that DOGE’s disruptive approach was needed, while acknowledging that its tactics remain debatable. He reports civilian headcount falling from roughly 2.3 million to about two million and says that reversing government’s direction required a “big bang strategy.” He says the federal workforce is now at its lowest share of the overall workforce since about 1960. His stated task now is to “institutionalize this,” including deciding which activities agencies should stop doing.
- The retirement operation supplies the episode’s clearest example of modernization’s potential and its execution burden. Kupor describes 400 million paper records stored in a Pennsylvania mine and processing that historically took three to 12-plus months. He says the first iteration can reduce processing time by more than 50%, targets a 15-day turnaround for complete electronic applications, and separately forecasts reducing annual HR IT spending from $2 billion to $40 million by consolidating 120 systems.
Deep dive
1. The first employee had to build everything around the investors
- After selling their last company to Hewlett-Packard in 2007, Kupor and colleagues remained there for another 12–18 months. The founders discussed institutionalizing their angel investing in summer 2008, between Bear Stearns’ collapse and Lehman’s bankruptcy. Kupor hesitated over leaving a stable job and moving his family from North Carolina; around Christmas, he committed. His official designation was first employee, not co-founder.
- The assignment was “we’re gonna invest, and you need to figure out everything else we need to do to build a firm.” That included furniture, bank accounts and capital calls against LP commitments. Without a firm email address, Kupor requested money from roughly 50 LPs through Gmail. One replied, “Who the fuck are you?” and warned the founders someone was impersonating them.
- Kupor’s original downside calculation was modest: they could probably raise a first fund, ride through the financial crisis and, if necessary, he would find another job. “I never imagined we’d get to where we were,” especially so quickly. Senra distinguishes entrepreneurial confidence from foresight: founders may expect success, but cannot predict its eventual degree.
2. a16z turned personal networks into a firmwide service
- Michael Ovitz, whom Kupor knew through the Loudcloud and Opsware boards, supplied the organizational analogy. CAA preserved the relationship between actor and agent while surrounding that relationship with specialists in books, radio and international deals. Senra calls it a “phalanx”: the client could draw on the whole firm rather than depend on one person’s connections. Kupor recalls that CAA had about 90% market share when he thinks this model was in place in 1990.
- Kupor describes a16z’s early LP pitch as “trying to institutionalize John Doerr.” An exceptional investor might know every relevant CFO, salesperson and political contact, but that model hits board-seat capacity limits and requires each partner to rebuild similar networks. As software expanded into more industries, the necessary expertise also expanded; the firm wanted professionals maintaining those relationships continuously.
- The application was a “reverse pitch”: after founders presented, a16z introduced the team available to help them. Entrepreneurs would get their board partner plus specialists whose daily work was building useful networks. Much of Kupor’s early job was hiring those experts. He estimates the firm grew from three people to nearly 100 in year one, and says he thinks it has about 700 by the conversation.
3. Skype supplied the returns and fees that accelerated the model
- Within months of closing its first fund, a16z joined Silver Lake and Michael Dell in buying Skype from eBay. Kupor says the neglected business still had powerful network effects and growing subscriber numbers. Ownership of its underlying intellectual property worried prospective buyers: there was a dispute involving the original founders, Janus and Nicholas, and eBay. Egon helped work with them to address that question. The buyout departed substantially from a16z’s stated early-stage strategy.
- Critics called the transaction crazy. Eighteen months later, Kupor says, the investors sold Skype to Microsoft for four times what they paid, returning “literally… half of our fund” to LPs. The result changed the narrative and helped a16z raise its second fund ahead of schedule. Kupor acknowledges the early outcomes could have gone the other way; instead, they dramatically accelerated the firm. Another early win was Nicira, a networking company that sold to Cisco during the first fund.
- Those returns mattered to the operating model. The founders initially took no salary so fees could support hiring; the second fund added a “stacking fee stream.” Launching in 2009 also helped: established firms were protecting existing portfolios after the crisis, while a16z could announce, “we’re open for business,” with fresh capital and attention available for new companies.
4. Marketing changed the basis of competition
- Kupor connects venture’s historical reluctance to market with capital scarcity: entrepreneurs approached firms because those firms controlled the money. a16z lacked competitors’ accumulated investment records—he points to Sequoia’s conference-room displays of Cisco, Apple and Google. The founders’ response was to “change the rules of the road” and compete around “capital plus”: money accompanied by an organization serving entrepreneurs.
- The enterprise-software lesson was that a good product still required explanation. Other VCs complained that a16z talked about services before delivering them; Senra responds, “That sounds like an entrepreneur to me.” Kupor agrees that creating a category required telling its story, while accepting the obligation behind the pitch: “we had to deliver on it at the end of the day.”
5. Scaling required smaller teams and less central control
- Kupor initially felt slighted that the founders would invest while he operated the firm. He later accepted a limitation: he was not a deep technologist. Hearing ten database pitches without understanding databases leaves you with “no fucking clue” which company deserves investment. His conclusion was to concentrate on strengths rather than treat investing as a status entitlement.
- For roughly ten years, the firm remained centralized, reaching around 500 employees with operating functions and effectively investors reporting to Kupor. Then specialization became necessary. Investing teams and their services moved into verticals: a gaming founder’s business-development needs differed from a life-sciences founder’s. Legal, finance, compliance and LP relations stayed centralized where a common structure still made sense.
- Kupor resisted losing responsibilities. The founders’ answer was that “the needs of the business have changed.” The reorganization also opened room for investing: the firm launched Later Stage Venture in 2019 with David George from General Atlantic. Kupor focused on healthcare and operating performance—“one to infinity” rather than the “zero-to-one” side. Before government, he had invested in about 15 companies and held roughly ten board seats.
6. Government’s problem starts with incentives to avoid risk
- The policy move had roots predating venture: Kupor worked on healthcare research in high school, then with Professor Alan Garber at Stanford on prostate cancer, and later with Jay Bhattacharya at Stanford Law School. The timing also fit his circumstances—children out of the house, financial flexibility and support from the founders. At the Office of Personnel Management, he describes his remit as talent and HR policy for roughly two million civilian federal employees.
- His organizational model is simple: people aligned around objectives, supported by incentives that reward desired behavior and correct poor performance. Government’s compliance culture instead encouraged avoiding exposure. One report described the job as keeping him out of the news and away from congressional scrutiny. Senra’s response: “let me just sit here and do nothing.” Kupor says that is precisely the problem.
- His alternative is “permissioned risk,” deliberately distinguished from Silicon Valley’s permissionless risk. Venture can absorb failures because a successful investment can generate enormous upside; government lacks that payoff structure. Still, Kupor argues that risk tolerance should depend on the activity. Life-or-death decisions warrant great caution, while improving administrative services requires allowing appropriately sized experiments and accepting that some will fail.
7. Strong performers need differentiated rewards
- Kupor reports that 99.7% of federal employees receive ratings of three or above, leaving 0.3% below expectations. The system is generally described on a five-point scale, though some organizations use four points. Among roughly 7,000 senior executives, about 90% receive fours or fives. Senra questions whether those distributions are possible; Kupor calls them evidence that “we’ve gamed the system.” His concern is that strong employees leave when colleagues who coast face no consequences.
- At the top, inflated ratings spread bonuses too thinly to motivate or retain people. Kupor describes a rule limiting the highest ratings to no more than 40% of the population and allocating at least 60% of the compensation pool to that group. He personally wants stronger differentiation, but says, “we’ve got to crawl before we walk and run.”
- Promotion rules create another barrier: employees generally must spend at least a year in grade before advancing, then face another year after promotion. A software developer who proves exceptional within six months remains stuck. Kupor says they are eliminating those “time in grade” constraints so demonstrated ability can determine progression rather than elapsed tenure.
8. Recruiting younger talent requires a credible career bargain
- When Senra suggests government clearly lacks the right people, Kupor qualifies the diagnosis: many employees are very good, and much of the problem lies in systems and incentives. The demographic gap is separate: 7% of federal workers are under 30 versus 23% nationally; 44% are over 50 versus 33%. He warns that, if nothing changes, half the workforce could retire over the next five or ten years, making a replacement pipeline necessary.
- A manager objected that his approach weakened government’s traditional lifetime-employment pitch. Kupor calls that “the worst pitch I’ve ever heard.” His alternative is two or three years working on difficult problems—at NASA or the Department of Energy, for example—followed by support moving into private industry. Senra endorses appealing to personal interest alongside patriotism: the experience should help both the participant and the country.
- Tech Force aims to bring in 1,000 software engineers, AI specialists and data scientists for two-year terms at GS-14, which Kupor describes as the second-highest individual-contributor level. The program uses third-party technical assessments rather than self-rated questionnaires. Forty private companies, including Alphabet, Meta and Databricks, will provide development support. Kupor says the plan is to hold a large job fair at the end of the two years. He invokes Israel’s IDF cyber corps: service should produce people whom investors and employers compete to recruit.
- His hiring example is Ed, known as “Big Balls.” Kupor says he thinks Ed was 19, attended Northwestern and dropped out as a sophomore; he now works with Joe Gebbia’s National Design Studio redesigning government websites. A government employee reportedly said Ed would never qualify without a degree and five years’ experience. Kupor rejects those proxies: demonstrated skill should justify GS-14 or GS-15 placement, rather than entry-level pay around $40,000. “Merit matters. Take a test, take an assessment.”
9. DOGE’s disruption now requires lasting institutions
- Kupor developed plans with Elon Musk’s team, including Steve Davis, before inauguration, but his nomination process kept him from executing them on day one. He reports civilian headcount declining from roughly 2.3 million to about two million, which puts the federal government at its lowest share of the overall workforce since about 1960.
- Turning a 2.3-million-person “oil tanker,” he argues, required a “big bang strategy.” He allows that tactics deserve debate and says history will judge the episode. His own role is to “institutionalize this”: examine what agencies actually do, identify priorities and remove activities. “We never take stuff off the plate in government. We just keep adding.”
10. The retirement mine shows what decades of paper cost
- OPM handles retirement for roughly 100,000–200,000 people annually. Kupor describes a foundational gap: government does not assemble the canonical history of an employee’s service until that person retires. Employees move among agencies with separate HR and payroll systems, so reconstructing decades of service becomes part of processing the retirement itself, rather than a record already available when needed.
- Applications historically traveled by stamped mail through HR, payroll and finally an underground former limestone mine in Boyers, Pennsylvania. Eight hundred retirement-services employees read paperwork and performed manual calculations. Processing took three months at the low end to more than 12 months; a missing signature could trigger another mailed exchange and another month’s delay. Staff drove golf carts through the facility’s limestone corridors.
- Kupor credits Musk and Airbnb co-founder Joe Gebbia with starting automation before he arrived. OPM declared June 30 the last day for incoming paper; Kupor says the process is now online. He says the first iteration can reduce processing time by more than 50% and targets a 90% reduction. His commitment is conditional: a complete electronic application should produce a full retirement check within 15 days.
- The legacy stock remains: 400 million pieces of paper, correcting Senra’s initial recollection of four million. A caller’s request can require a “runner” to retrieve a physical file, sometimes dating to the 1950s. OPM has purchased two room-sized scanning machines to digitize that stock, with the eventual aim of closing the mine’s storage operation and giving service staff immediate access.
11. Modernization depends on changing who bears the risk
- A separate project addresses the fragmented employment record directly: Kupor says OPM is consolidating 120 HR IT systems into one. He projects annual spending falling from $2 billion to $40 million—a reduction he describes as 95%—while creating an employment history available throughout a person’s career. The intended result connects lower administrative spending with eliminating reconstruction work at retirement.
- The mine’s problems were documented in the 1970s and 1980s, Kupor says; identifying them was never the breakthrough. The barrier was that poor service carried less personal risk than attempting change and attracting congressional criticism. His rejection of the idea that Silicon Valley celebrates failure is blunt: “Nobody fucking likes failure.” Innovation nevertheless requires accepting that some reasonable attempts will go wrong.
- Kupor told the president during his December interview that he was not pursuing another job through this appointment. That independence, he argues, makes scrutiny easier to accept: “if it turns out I did something stupid, that’s fine. Fire me.” His sequence remains recruitment, a workplace that rewards performance and then modernization; bringing ambitious people into the old incentive system would leave the underlying problem intact.