Ryan Akkina on MIT’s Venture Portfolio and Direct Co-Investing

Ryan Akkina of MIT Investment Management Company tells Harry Stebbings that selecting venture funds has become harder as the market has grown more complex and competitive. He discusses manager selection, liquidity and direct co-investments from the perspective of a university endowment.

Akkina first expected to be an engineer and worked on chip design at HP and Intel while young. At Stanford he became involved in entrepreneurship activities and met founders and investors, then briefly worked in management consulting before joining MIT’s investment team in 2009.

He had considered venture capital directly but did not yet feel he had an edge as a young graduate. An endowment offered a way to learn about many types of investing while supporting an institution. He describes chance and timing as part of how he found the job.

Historically, he says, large endowments could concentrate on a few established US Series A firms and rely on their long-standing relationships and institutional reputations. Today more stages, geographies, specialties and fund models require broader coverage, while access to proven managers is more competitive.

Even MIT can miss strong firms, Akkina says. Once a manager is established, existing LPs often receive priority, leaving a new investor with a small allocation or none. He thinks smaller commitments can sometimes still be worthwhile for the relationship.

He sees difficult fundraising periods as a way to observe which emerging managers are committed to building a firm. Less competition and lower prices may create opportunities, although hard conditions also make it harder for talented managers to raise.

Akkina evaluates managers through “see, pick, win and service”: access to opportunities, selection, the ability to win an allocation and support for founders. He spends particular attention on winning and service because founder references can affect access to subsequent deals.

Track record matters differently by stage. Angel investments can reveal a new manager’s taste, but a tiny check into a successful company does not necessarily demonstrate an ability to lead a larger investment. In early cases the assessment must rely more heavily on qualitative judgment.

He also considers whether a founder will want to work with a manager for years. Personal rapport can distinguish credible investors competing for the same deal. He recalls that different partners at Sequoia appealed to different kinds of founders, broadening the firm’s reach.

Akkina says it can take five to seven years for a venture investor to build a reinforcing cycle of wins, reputation and deal flow, and luck affects that timing. In the first three years of a fund, he often cannot tell whether the strategy is working; later he looks for fundamental company progress rather than just higher round prices.

Successful firms can falter if they grow too quickly, lose humility or lose motivation, he says. MIT seeks to understand what drives managers and whether they can continue competing. These are his observations, not a mechanical predictor of any particular fund’s performance.

MIT’s default is loyalty to existing managers, but fund growth still requires scrutiny. Akkina considers how many strong partners a firm has and how much capital they have historically been able to deploy. He says he tries to communicate a decision not to reinvest early, rather than surprise a manager.

MIT nearly passed on Green Oaks’ first institutional fund under an earlier preference for established brands. Akkina says he reconsidered overnight and proposed a small commitment. He cites unusually strong founder references and the team’s initial e-commerce thesis as reasons for the decision.

A manager may need to evolve beyond an original strategy. Akkina accepts that tension if a change is purposeful and communicated early, while wanting managers to be clear about what their LPs are underwriting.

He describes three US venture buckets at the time: approximately six to eight core manager relationships with $50 million to $150 million per fund, a middle group of roughly 20 managers with smaller checks, and a newer experimental bucket with about $1 million commitments. Sizes and counts are his contemporaneous estimates.

Commitments vary with the cycle. Akkina describes a peak of roughly $3 billion in total annual private commitments and a more recent level of about one-third as much; he is not certain of the venture-only amount. Those figures should not be interpreted as a precise venture allocation.

He wishes managers had been more disciplined about compressed deployment in 2021. At the same time, he says completely sitting out a period can itself amount to market timing. He wants managers to acknowledge mistakes, explain lessons and maintain discipline across cycles.

Improved public markets had eased the denominator and liquidity pressure, he says, but many large private companies had not yet produced liquidity. Rapid fundraising and larger funds had also increased commitments. These comments describe conditions at the time, not a current liquidity assessment.

MIT has to support university spending, which Akkina describes as roughly 5% of assets annually under a smoothing formula. The team models those needs, but unusually fast fund deployment and delayed exits made the recent period more challenging.

He says an LP has limited control over when private companies exit, though a manager may occasionally consider a secondary sale. MIT rarely sells fund positions because secondary buyers generally seek a discount; he says the institution had not been forced to sell for liquidity.

Direct co-investments add another tool. Akkina finds it easier in some cases to make a substantial commitment to a known company than to a blind-pool fund, provided the valuation, structure and diligence support conviction. Co-investing still involves risk and depends on access.

He describes a Coupang investment in a senior structured note as an example of seeking downside protection while retaining upside. Seniority and a view of enterprise value informed the decision; neither structure nor seniority guarantees recovery in every scenario.

MIT first learned about Coupang through a founder reference on Green Oaks. Akkina says the team later visited South Korea, toured warehouses and studied the business and financing terms. Prior experience with e-commerce companies elsewhere helped frame the analysis.

He reports a $120 million Coupang co-investment and a distribution-price multiple he recalls as roughly seven to eight times, adding that the value was lower after the stock declined. This is his approximate account; distribution price, subsequent value and realized cash return are different measures.

On endowment incentives, Akkina sees a tension: conventional structures limit incentives to take risk, but governance also needs to prevent staff from having a one-sided option on institutional capital. For a hypothetical family office, he would consider personal investment alongside performance-linked incentives.

Price discipline remains important in direct deals. He says MIT declined some co-investments during the market peak despite strong supporters, while also acknowledging that the team itself was influenced by high public-market valuations and made mistakes.

Stebbings asks whether saying no could cost access to the next deal. Akkina says relationships matter but an LP must reject marginal uses of capital. In some circumstances a small commitment may support a valuable relationship, while the decision still needs to make economic sense.

Akkina describes deciding quickly on a co-investment during a bank collapse after already knowing the company and founder and trusting the manager. He says MIT did not provide the whole round. Prior knowledge, relationships and financing structure were important to that compressed decision.

His worst direct investment involved oil and gas acreage. He describes an initial rise in value followed by a merger, falling oil prices, disappointing wells and debt. A high mark before exit did not become a realized gain.

The lesson he draws is to examine businesses that must repeatedly reinvest much of their balance sheet to maintain production. Even with good individual decisions, repeated exposure to operational and commodity risk can compound. He estimates MIT ultimately recovered about one-third of its capital in that investment.

Over time Akkina moved from more public-market work toward private investments, where he feels his own skills are better suited. He says a long feedback loop makes it important to trust judgment while remaining open to evidence.

He identifies passing on OpenAI as a significant omission. At the time, MIT knew Sam Altman through Y Combinator but viewed the technical outcome and unusual structure as speculative. In hindsight he thinks a small position might have been justified; this is retrospective judgment, not proof the outcome was predictable.

At the overall portfolio level, he describes a target of roughly half public and half private investments, with 5–10% in cash or short-duration government securities. He also monitors how much the portfolio might decline relative to the market. These are the targets he described at the time.

Akkina says fewer managers were returning to market and some fund-size expectations had fallen. He does not consider LPs closed for business, but expects smaller and emerging managers to face more difficulty than firms with established reputations.

He advises fundraising managers to build a pipeline and relationships over time, while understanding each LP’s preferred stage. Pushing an institution into a formal evaluation too early can make a later conversation harder. His suggested 10–20% limit on any one LP is a guideline with possible exceptions.

Investor-relations teams can be helpful if they protect investors’ time while keeping LPs informed and providing access when needed. Akkina says LPs also have a responsibility not to consume excessive time that managers could spend investing.

On China, he says MIT had reduced activity after a successful earlier period and was considering regulatory and sensitive-sector risks. He does not rule out all future investment. These comments reflect the institution’s stance at the interview’s time.

Generational transition is hard because exceptional investors are scarce. He looks for mentorship, deliberate skill development and timely sharing of economics. Large fund sizes also reduce the probability of a given multiple, in his view, so MIT may scale down if it no longer sees exceptional return potential.

Akkina has become less skeptical of hard technology after seeing companies such as Tesla, SpaceX and Anduril. He warns against procyclical enthusiasm followed by blanket rejection of entire categories, and tries to remain open to unusual ideas.

His favorite investment advice is to remain humble in both success and disappointment. He values independent thinking, names time as a scarce resource and says delivering a rejection—especially to a long-standing partner—is among the hardest parts of the job.

Looking ahead, Akkina imagines continuing similar investment work at MIT or a family office. Stebbings thanks him for discussing an area of the industry that receives less attention.