Before James Kim evaluated education startups, he used their products in his classroom.
The Reach Capital partner began his career as a high school math teacher with Teach For America after studying astrophysics at Yale. He later worked in college access and Yale admissions, and earned graduate degrees in business and education from Stanford before joining Reach. As a teacher, he was an early user of ClassDojo, an experience he describes as “game-changing.”
His investment firm, Reach Capital, has also evolved. Founded in 2015 after spinning out of NewSchools Venture Fund, the San Francisco-based firm initially focused on technology for K-12 schools and districts. Its mandate has since expanded to learning, health and work, with investments in companies including ClassDojo, Cartwheel, Stepful and Replit.
In August, Reach announced its fifth fund, a $265 million vehicle that brought its assets under management to nearly $1 billion. The firm, which has backed more than 180 companies, plans to invest in approximately 50 additional teams through the new fund, targeting pre-seed, seed and Series A rounds.
Crunchbase News spoke with Kim about how Reach’s education roots inform its broader strategy, what it saw in AI detection company GPTZero, and why some of its best investments initially divided its own team. We also discussed how ownership stakes and secondary transactions have shaped the firm’s approach to returning capital.
This interview has been edited for length and clarity.
Crunchbase News: You started your career as a high school math teacher. How did that shape your view of education technology?
Kim: I taught ninth through 12th grade, mostly Algebra I and geometry. I was very early to ClassDojo. Back then, it was a very simple classroom management tool, and it was game-changing for me. It has since evolved into a sprawling suite, but I already knew technology could move the needle.
Reach began as an education investor. How did the firm’s focus evolve into learning, health and work?
Kim: Reach started in 2015, and our original focus was exclusively on learning and education technology — more specifically, technology sold into schools and districts.
Our founding partners were seeing the growth of devices in schools, one-to-one device ratios and broadband penetration. The belief was that you would start to see a proliferation of technology serving students, teachers and administrators and moving the needle on learning outcomes. That is what ended up playing out. Some of our early bets are now household names in edtech, such as Newsela, Nearpod and ClassDojo.
Reach spun out of NewSchools Venture Fund, where three of our four founding partners had started doing for-profit seed investing before 2015. Over time, we branched out from K-12 into higher education, and then into consumer learning.
We also started getting pulled toward healthcare and work. Health and wellbeing affect a learner’s readiness for the classroom, so we were being pulled toward mental health, behavioral health and developmental health. On the work side, formal learning typically ends in college, but you still need the bridge to a job. Once you’re on the job, you need to continue learning, especially in the current environment.
A few years ago, we officially put a stake in the ground and said we invest across learning, health and work. The common thread is that these are the most important systems underlying human development and flourishing.
So far, the breakdown has been roughly 50% learning, 25% health and 25% work.
These sectors have different buyers, regulations and business models. What did investing in education teach you that translates to healthcare and work?
Kim: Education is highly regulated, and institutional buying is complex. Many stakeholders must provide input, and decisions often require board-level approval. We were familiar with and comfortable with regulated products and complex go-to-market motions.
Selling to a university or a K-12 district, in terms of the complexity and the need to intersect with regulation and policy, is quite comparable to selling to a health system, hospital or insurance carrier.
Cartwheel is a good example of where our experience gave us an advantage. It provides mental health teletherapy, but the customer is the school district and the care setting is in schools.
While we were relatively new to behavioral health when we made that investment, we were very familiar with selling to school districts and implementing an intervention within a school setting. That gave us an advantage both in winning the investment opportunity and supporting the company afterward.
Reach talks about “elevating human potential.” What makes an investment fit that mandate?
Kim: We look for products and services that increase access to quality education, quality care or quality earnings, or some combination of those. We really have to see the mechanism for that happening.
It’s not good enough to be pitched a SaaS or AI product and have the founders say, “This will make people more productive at work, and therefore maybe they will earn more or enjoy their jobs more.” That’s plausible, but it isn’t a good enough theory of how the product actually increases access to quality earnings.
Several teammates, including myself, are former educators. We care about meaningful, needle-moving impact.
We address access from both the demand side and the supply side. On the demand side, we build products that create delightful, impactful experiences for consumers: Brilliant, Stepful’s allied healthcare training, and Replit for what its founder calls “citizen developers.”
On the supply side, it’s helping schools, health systems and companies function more efficiently and effectively. WorkWhile, for example, helps logistics and hospitality companies hire hourly workers better and treat them better. We think about the problem from both sides of the market.
Reach backed GPTZero, which agreed to join Superhuman. Was that an investment in solving a problem created by AI?
Kim: That wasn’t actually our investment thesis. In my opinion, they have the best AI detector on the market. But their thesis was that AI use in education should not be banned. It shouldn’t be punitive or become a cat-and-mouse game between the educator and the learner.
Since all of us as professionals use AI in our jobs, students should learn how to use it responsibly and thoughtfully. It was the first AI detector company I had seen with that positioning, as opposed to helping schools, districts and universities clamp down on AI usage.
That’s why I liked it. I don’t see it as a problem for students to use AI for certain purposes, such as brainstorming, organizing thoughts and editing. GPTZero helps educators work with students and teach them how to use AI responsibly.
Do you see other opportunities to build companies around the problems AI creates?
Kim: Yes, most notably in AI governance and evaluations.
Going back to regulated categories, you need to make sure AI is not irresponsibly using personally identifiable information in education or healthcare. Those categories have very strict regulations, and infrastructure companies are emerging to help ensure AI products interact with that data responsibly.
I do think there are opportunities for large, successful, profitable companies to be created in response to the problems AI creates.
What does GPTZero’s sale tell you about which AI products can become durable standalone companies and which belong within larger platforms?
Kim: I don’t know whether GPTZero is an example of either. They had been fielding acquisition offers from pretty early on. This particular opportunity arose earlier this year, and there was a compelling synergy story with Superhuman, formerly Grammarly.
The founders and the board agreed that it made sense to combine the companies. But they could just as easily have remained standalone. There is another AI detection company called Pangram that’s quite good and is still a standalone, venture-backed company.
It’s not clear to me that this category can’t exist on its own. We have to be very case-by-case. We don’t know until we see the trajectory, financial sustainability and evolution of the market, whether it makes sense to explore a strategic acquisition or grow independently and perhaps one day become a publicly traded company.
Reach has backed more than 180 companies. What patterns have you noticed among those that exceeded expectations?
Kim: The reality of venture is that a very small share of investments yields the overwhelming majority of returns. Given the outlier nature of those successes, it’s hard to identify unifying principles that always predict a successful exit. If we could, we would only make those investments.
But there are a few things we think are somewhat common across our wins. One is a very smart, gritty, genuinely mission-driven founder or founding team, often with a strong personal connection to the problem.
Second is some sort of nonconsensus insight about the market. That often comes from experiencing the problem themselves, sometimes as practitioners in the industry — a former educator, for example.
Third is a defensible and scalable business model.
The nonconsensus insight is funny because it’s what makes these companies look weird and maybe uninvestable in the beginning. Even within our team, they’re often the most polarizing opportunities. You look at the votes, and a bunch of people vote no, but it gets just enough yeses to pass. Then, 10 years later, it’s one of our best investments.
People lose sight of this in the current environment, but the best opportunities are often the most nonconsensus. Those can be the most creative and interesting founders, and you can get in at reasonable valuations early and help them become consensus bets.
Reach says its first two funds have returned investors’ capital. What produced those returns?
Kim: We’re no exception to the power-law distribution in venture. Our first fund has been returned 2.2x by literally two investments. Our second fund has been returned 1.2x by one investment. These are cash distributions, by the way.
Especially in the second fund, there’s still remaining value, so we’ll see how high that number gets. One fund-returning exit would be great; two or three would be amazing. That’s what we’re striving for.
How has that experience changed your approach to ownership and follow-on investing?
Kim: We’ve had some great exits where, given how prominent the company is, you might be surprised that our share of the returns wasn’t that large. That’s purely an ownership calculation.
From our first fund to today, we’ve meaningfully increased our initial ownership target. We now try to get as close to 15% ownership upfront as possible.
That means writing slightly bigger checks and reserving relatively less for follow-on investments and more for the initial check. It’s important for the exit math and distribution math.
What role do secondary transactions play in getting liquidity?
We’ve become very proactive and conscientious about liquidity opportunities. Much of our distributions to investors come from secondary sales — selling our shares in a subsequent round, before an actual exit.
Whenever there’s an opportunity, we evaluate it. We consider how long we’ve been in the company, how much we can get out, and how much upside we think remains.
As companies stay private longer, early-stage investors are achieving liquidity much more often through secondary transactions. I like to think we were early to that trend, but it’s certainly mainstream now.
Sometimes, when we see enough upside, we’ll buy secondary shares. GPTZero was an example: After our Series A investment, we bought secondary shares, and I’m really glad we did because it worked out well. For us, we sell more often, but occasionally we buy.
Related Crunchbase query:
Related reading:
Illustration: Dom Guzman