Corporate Venture Capital: Innovation as Investor

In June 2016, at CB Insights’ Future of Fintech conference, Fred Wilson — the Union Square Ventures partner with one of the best track records in venture — told a room full of corporate development people that what they did for a living was dumb. Not misguided, not difficult: dumb. “Corporate investing is dumb. I think corporations should buy companies. Investing in companies makes no sense.” He compared taking corporate money to “doing business with the devil.” The clip made the rounds, corporate VCs got defensive on Twitter for a week, and then everyone went back to work — which, at the time, meant corporates participating in something like a quarter of all venture deals globally, per CB Insights’ own tracking.

Both things are true at once, and that’s the subject of this post. Wilson’s critique lands squarely on most corporate venture funds ever raised, and the industry’s own boom-bust history is the evidence. And yet a handful of corporate funds — Intel Capital across three decades, GV since 2009, Salesforce Ventures in its ecosystem — have compounded value for their parents in ways no internal program could replicate. The difference between the two populations isn’t sincerity or strategy decks. It’s structure, and the structure is knowable.

Here’s the frame I’ll defend: corporate venture capital is not a growth engine, and the moment a company treats it as one, it fails. It’s a sensing instrument — purchased optionality on futures you can’t build internally. Everything in the previous six parts of this series has been about innovating inside your own walls or just outside them. CVC is the admission that some of the futures that matter to you will be built by other people, and that a minority equity check is the cheapest legitimate way to sit close enough to watch. It works when the fund is built like an actual venture firm with a strategic sensor bolted on. It fails when it’s a strategy department with a checkbook.

The forgotten history: this is the fourth boom

CVC feels like a 2010s invention if you’re under forty, which is exactly the problem — the industry has no memory, so it re-learns the same lessons every cycle at full price. Gompers and Lerner, who did the foundational academic work on corporate venturing, document distinct waves, and the shape of each one is identical: bull market, corporate FOMO, a rush of new funds, a downturn, and a corporate retreat far faster and more total than anything the independent VCs do.

The first wave came in the 1960s conglomerate era, when a large share of the biggest US industrial companies — DuPont, GE, Exxon among them — set up internal venturing and investment arms, mostly imitating the early success of the institutional venture funds. The 1973 market collapse ended it. A second wave rode the 1980s technology boom and died with the 1987 crash. The third wave was the dot-com era, when corporate venture money grew to a meaningful slice of all venture investment at the 2000 peak — and then came the wipeout. When the bubble burst, corporates didn’t trim; they fled. Units were shuttered wholesale, portfolios written down by billions, and Henry Chesbrough opened his 2002 HBR article on the subject by cataloguing the carnage. The pattern to notice: independent VC firms had a brutal 2001–2003 too, but the firms survived, because a ten-year fund structure forces patience. Corporate units investing off the balance sheet had no such structure, so a bad year for the core business became an execution order for the venture arm — regardless of how the portfolio itself was doing.

The fourth wave, the one we’re living in, began in the early 2010s and dwarfs the others. By the late 2010s and early 2020s, CB Insights’ industry reporting had corporate venture participation at roughly a quarter of global venture deals, with hundreds of new corporate funds launched every year — many of them, statistically, by first-time corporate investors who will not exist as investors by the bottom of the next cycle. That’s not cynicism; it’s base rates. The historical average lifespan of a corporate venturing program has been measured in single-digit years, against the decade-plus horizons that venture returns actually require. Most corporate funds get shut down before their first fund’s vintage could possibly have matured.

The standard-bearer for the alternative is Intel Capital, founded in 1991, which has invested billions of dollars into well over a thousand companies across every cycle since — including straight through the dot-com crash that killed most of its peers. Keep Intel Capital in mind, because it’s not just a longevity story. It’s the cleanest example of the one quadrant of Chesbrough’s map where strategy and investing genuinely reinforce each other. We’ll get there.

What Fred Wilson actually said, and what answers him

The pop version of Wilson’s 2016 broadside is “famous VC hates competition.” The actual argument deserves better, because it’s a precise structural diagnosis, and any corporate fund that can’t answer it point by point shouldn’t exist.

His case has three parts. First, conflicted motives: a strategic investor wants something from the company other than the company winning — market intelligence, product alignment, an option to acquire — and when those interests diverge from maximizing the startup’s outcome, the strategic investor is a bad shareholder by construction. “You’re not a good investor if you have a strategic motive,” in his words: the entire discipline of investing is pricing and supporting outcomes, and a strategic agenda corrupts both. Second, no speed and no conviction: corporate investment decisions route through committees of people who don’t own the outcome, which means corporates rarely lead rounds, rarely price them, and show up late asking for a “look.” A lead investor who can’t move in days is a lead investor who loses every competitive deal, and venture returns are concentrated in exactly the competitive deals. Third — and this is the one founders learn the hard way — strategic strings poison exits. Rights of first refusal on a sale, blanket information rights flowing to a potential acquirer’s corp-dev team, board seats occupied by a competitor of your future acquirer: each of these makes the startup less buyable by anyone other than the investor’s parent, which caps the auction, which is where startup outcomes are made. Wilson’s conclusion followed cleanly: if a corporation wants a company, it should buy the company. Minority investing with strategic intent is the worst of both worlds.

Every word of that is right — about the median corporate fund. The counter-evidence is the funds engineered specifically so that none of it applies.

GV is the purest counterexample because it was designed as one. When Bill Maris set up Google Ventures in 2009, the founding decisions were all repudiations of standard CVC practice: a financial-first mandate (returns are the goal, strategic benefit is exhaust), independent investment decision-making without sign-off from Google product units, compensation for partners tied to portfolio performance the way a real firm’s is, and no strategic rights in term sheets. The result reads like an independent firm’s record — Uber, Nest, Flatiron Health, Slack — and founders take the money precisely because it doesn’t behave like corporate money. The strategic value Alphabet gets is real but indirect: a decade-plus of privileged visibility into where technology is going, bought with a fund that would justify itself on returns alone.

Salesforce Ventures, launched the same year, answers Wilson from the opposite direction: it’s unapologetically strategic, but the strategy is ecosystem growth, which happens to align with founders rather than against them. Salesforce invests in companies building on and around its platform; every portfolio company that wins makes the platform more valuable, and Salesforce’s distribution help makes portfolio companies more likely to win. The conflict Wilson describes largely dissolves because the strategic goal isn’t optionality to acquire — it’s a bigger ecosystem, which both sides want. And Intel Capital ran a version of this logic at industrial scale for thirty years: investing in companies whose success expanded demand for computing — and therefore for Intel silicon — which meant Intel needed its portfolio to win in the market, not to be captured cheaply.

Notice what all three have in common. In none of them does the fund’s success require the startup to subordinate its outcome to the parent’s roadmap. Wilson’s critique is fatal to funds whose strategic logic needs the startup to serve the parent. It’s toothless against funds whose strategic logic is served by the startup winning on its own terms. That distinction is exactly what Chesbrough’s framework formalizes.

Chesbrough’s map: four kinds of corporate investment

Henry Chesbrough’s “Making Sense of Corporate Venture Capital” (Harvard Business Review, 2002) remains the best organizing tool for this decision, and it’s a simple 2x2: what’s the objective of the investment (strategic vs. purely financial), and how tight is the link between the startup and the corporation’s own operational capabilities — does the startup use, strengthen, or connect to what the parent actually does?

Driving investments — strategic objective, tight operational link — advance the current strategy directly: you invest in companies that plug into your platform, your standards, your supply chain. This is Salesforce Ventures’ home quadrant, and much of Microsoft’s and SAP’s ecosystem investing. It works, with a known ceiling: driving investments deepen the current strategy but will never tell you the current strategy is wrong, because everything you fund is selected for fitting it.

Enabling investments — strategic objective, loose operational link — fund complements: businesses whose success stimulates demand for yours without touching your operations. Classic Intel Capital: money into video, networking, and software companies that made powerful chips more necessary. The test for this quadrant is honest arithmetic about spillover — you need to capture enough of the demand you’re stimulating for the strategic logic to beat just indexing the market.

Emergent investments — financial objective today, tight operational link — are the sensing quadrant, and to my mind the most underused. These are investments in startups exploring markets adjacent to your capabilities that your strategy currently says no to. Financially they must stand on their own; strategically they’re an option that becomes valuable precisely when your strategy changes. This is the disciplined answer to the Innovator’s Dilemma mechanics from part one of this series: your resource-allocation process will rationally starve an internal team chasing a small weird market, but a minority check into someone else’s team chasing it survives, because it isn’t competing for your engineers or your margin thresholds. You’re paying a startup to hold a position your own P&L won’t let you hold.

Passive investments — financial objective, loose link — are just conglomerate diversification wearing a fleece vest, and Chesbrough’s verdict was blunt: this is a questionable use of shareholders’ money, since shareholders can diversify themselves without paying your corp-dev team to do it badly. If your fund’s portfolio can’t be sorted into the first three quadrants, you don’t have a strategy; you have a hobby.

The map’s practical use is as a portfolio audit. Run every investment your fund has made through the 2x2. Funds that work cluster deliberately — Salesforce in driving, Intel in enabling, GV effectively running emergent-plus-financial at scale. Funds that fail scatter across all four quadrants, because “strategic” was never defined tightly enough to say no to anything.

The design choices that actually decide it

Chesbrough tells you when CVC adds value. Whether your fund can capture that value is decided earlier, by a handful of unglamorous structural choices — and I’ve watched corporates get every one of them wrong while sincerely believing they were building the next GV.

Fund structure versus balance sheet. A committed, ring-fenced fund with a multi-year horizon is what lets a venture portfolio survive the parent’s bad quarter. Balance-sheet investing — deal-by-deal approvals from corporate treasury — means your portfolio’s life expectancy equals the time until your CFO needs a cost story. The history section above is one long argument for this: the 2000-era units died not because their portfolios were uniquely bad but because there was no structure forcing anyone to wait out the trough.

Carry, and the refusal to pay it. Venture investing is a talent business, and the talent is priced in carried interest. Corporates almost universally refuse to pay it — HR can’t stomach an investment manager out-earning the CEO in a good exit year — and the consequence is mechanical: anyone on your team who turns out to be good leaves for a firm that will pay them what their track record is worth, and you’re left with the people who couldn’t. GV’s willingness to compensate like a real firm isn’t a detail; it’s arguably the design decision. The pattern rhymes with Xerox Technology Ventures, which Chesbrough studied closely: XTV produced outstanding financial returns in the mid-1990s and was shut down anyway, in part because its success — and its partners’ venture-style payouts — created more internal friction at Xerox than its returns bought goodwill. Read that twice: a corporate venture unit can be killed for succeeding on the wrong pay scale.

Investment committee speed. If your process can’t produce a signed term sheet inside a couple of weeks, you will systematically see only the deals that faster investors passed on. Adverse selection isn’t a risk of slow committees; it’s their guaranteed output.

The strategic-rights trap. This is Wilson’s third point, and it’s where well-meaning corporate lawyers do the most damage. Every right your legal team is proud of negotiating — right of first refusal on acquisition, expansive information rights, a board seat for your business unit — is a discount on the startup’s future exit and a warning flare to every quality co-investor in the next round. Sophisticated founders now treat ROFRs as near-disqualifying, and sophisticated corporate funds have learned to take standard investor terms and win access through being useful instead. The uncomfortable rule: the more contractual strategic protection you demand, the less strategic value you’ll ever receive, because the companies that matter can refuse your terms, and will.

Sponsor mortality. Every corporate fund has a champion — usually the CEO or a powerful CSO — and the empirical lifespan of corporate venture programs tracks the tenure of champions, not the maturity of portfolios. This is the same lesson Denmark’s MindLab taught about innovation labs, which I covered earlier in this series: sixteen admired years, dead in one budget cycle when the sponsors moved on. A venture portfolio is even more exposed, because its returns arrive on a seven-to-ten-year clock while corporate sponsorship turns over faster than that. The defenses are structural, not rhetorical: committed capital the next CEO can’t cheaply claw back, an external LP or fund-of-funds structure if you can get it, and a reporting line that makes the fund’s results legible on their own terms rather than as a line item in someone’s strategy budget.

Where this sits in the toolkit

Step back to the series arc, because CVC only makes sense in relation to the tools it complements. Part five argued for metered funding — funding internal ventures like a VC funds startups, tranche by tranche, priced against evidence. CVC is the same discipline pointed outward: instead of metering money into your own teams’ hypotheses, you’re buying stakes in other people’s, and the market — not your innovation board — runs the metering. Part six covered spin-outs, where an idea born inside leaves to grow outside your processes and values. CVC is the mirror image: capability born outside, with a wire running back in. Internal metered funding, spin-outs, and CVC form one continuum — how far from your own P&L does this bet need to live to survive? — and the honest answer differs per bet.

What CVC uniquely buys, when the structure is right, is the thing no internal mechanism can: exposure to futures your own strategy has already rejected. Your labs explore what you can imagine; your spin-outs carry what you invented but couldn’t keep; your venture fund watches what other people believe strongly enough to bet their careers on. That last signal — real teams, real capital, real customers, in markets your planning process filed under “not material” — is the most honest strategic intelligence money can buy, and a minority check is the price of a front-row seat.

But notice the assumption buried in everything above: that the companies you back are worth backing. The best-structured fund in the world, sitting in the right Chesbrough quadrant, with clean terms and committed capital and a fast IC, still produces nothing if it picks badly — and picking is a craft with its own evidence base, its own failure modes, and its own pop mythology in need of correction. That’s part eight: how to back the right startup.

Further reading

• Henry Chesbrough, “Making Sense of Corporate Venture Capital” (Harvard Business Review, 2002) — the 2x2, plus the post-dot-com carnage that motivated it.
• Paul Gompers & Josh Lerner, “The Determinants of Corporate Venture Capital Success” (NBER, 1998) — the evidence that strategically aligned corporate investments perform as well as independent VC, and misaligned ones don’t.
• Fred Wilson at CB Insights’ Future of Fintech (2016) — the “corporate investing is dumb” remarks, worth watching in full rather than in quote-tweet form.
• CB Insights, State of CVC reports — the running census of the fourth wave, including corporate participation rates in global venture deals.