Blackbird Raised a Record Billion. Then Wrote Down Canva by $7 Billion. Here's Why It Makes Sense.

On one side of the headline, Morgan Stanley, Schroders, and a U.S. private equity firm just handed Blackbird Ventures A$1.05 billion ($750 million) — a record for an Australian venture capital fund. On the other side, also that week, Blackbird wrote down the value of its biggest-ever bet, Canva, by $7.1 billion, from $42 billion to $34.9 billion.

This is weird, unless you think about what a venture capital fund actually is.

Blackbird doesn't sell stocks. It sells something closer to a lottery ticket factory. The firm has raised six funds since 2012, starting with $29 million and now sitting above $1 billion. What the investors in that new fund — pension funds, global asset managers, sovereign wealth pools — are buying is a commitment: Blackbird will take their money and deploy it into roughly 40 to 50 Australian and New Zealand startups. Pre-revenue. Pre-product. Sometimes, as the firm put it, "pre-idea."

About 96% of Blackbird's initial investments from its last early-stage fund went in at the pre-seed or seed stage. The business model is: most of them will fail. A handful will matter enough to cover everything else. Canva was that handful.

So when Blackbird's marketing leads with Canva, and then Canva gets marked down 17%, it reads like a problem. But the fund isn't a Canva position. The fund is a portfolio of dozens of pre-seed companies, most of which nobody has heard of, and the investors know the math.

The question isn't whether Canva is still valuable — it probably is. The question is what the valuation cut tells us about the economics Blackbird's new fund is walking toward, and whether the old playbook still works.

To understand the write-down, you have to look at what changed under Canva. It's not a product failure. Canva hit about $4 billion in annual recurring revenue at the end of 2025, with roughly 31 million paid subscribers out of 265 million monthly active users. The company is profitable at scale and dominates online design for non-designers.

What broke was the AI math.

Canva spent years building its case on the idea that software has near-zero marginal cost — once you write the code, serving one more user costs almost nothing. That was the whole SaaS pitch. Then AI features arrived, and each AI-generated design now costs real money in compute and model API calls. As one analyst put it, it's like discovering your car has a fuel bill.

Demand for the new AI tools "significantly exceeded expectations". Usage tripled. But inference costs blew through the company's projections, and CEO Melanie Perkins told investors the company had to cut its 2026 revenue growth forecast from 30% to 20% — a one-third reduction. The company slowed its AI rollout to rebuild architecture and reduce unit costs. Canva says the cost per AI task has dropped 90% since April, but with three times the volume, the savings are partially offset.

That's the same pattern we saw when Figma, the closest public-market comparator, disclosed its free-cash-flow margin halved from 27% to 14% in one quarter. The stock dropped 15%.

Both companies are learning the same thing: AI doesn't just add features. It adds a compute bill to every single task a user performs, and that bill grows with adoption instead of shrinking.

The valuation cut was the mechanical consequence. Canva's external backers (Blackbird and fellow early investor Airtree Ventures) marked the company down from $42 billion to $34.9 billion. A separate independent internal valuation — used for employee share sales — went even lower, from $38.9 billion to $31 billion. Neither figure represents an actual trade; no shares changed hands. These are book-keeping adjustments, the private-market equivalent of a re-pricing, but they send a signal about what the market is starting to price into AI-dependent software.

Here's where the Blackbird fund raise starts to look less like a contradiction and more like a deliberate trade.

Blackbird's own numbers tell a story of heavy concentration in the highlights, but diversification in the plumbing. The firm has invested more than $3 billion across about 190 companies since 2013. Its portfolio is valued at over $12.5 billion and has returned $2.25 billion in cash to investors, with a net internal rate of return of 32%. Canva accounts for a large portion of the paper value, but Blackbird already sold down a slice of its Canva stake in 2023, offloading about 3% of its shareholding for $150 million. That sale returned real cash to Blackbird's early-fund investors — actual distributions, not just marks on a spreadsheet.

Blackbird's model of going in early and riding companies through to late stage means it sometimes becomes the largest shareholder — which is how it ended up as Canva's biggest backer. But the model also means the new fund isn't buying into existing companies at their current valuations. It's buying into companies that don't exist yet, at prices that haven't been set, in a market that will look different in three to five years.

The investors who wrote checks — Morgan Stanley's investment management arm, Schroders, U.S. private equity firm Adams Street Partners, and Australian pension funds like Hostplus, HESTA, Aware Super, and the Future Fund — aren't buying a Canva proxy. They're buying Blackbird's ability to find the next outlier, wherever it appears, across AI, fintech, biotech, space, or something nobody has named yet. Blackbird's sixth fund is structured to scatter widely at the earliest stage and concentrate capital later, backing winners through the growth gap when check sizes jump.

The argument in favor is simple: Blackbird already did it once, and the Australian tech ecosystem is growing about 50% faster than the broader economy. The argument against is equally simple: Canva was an extraordinary outlier, and extraordinary outliers don't replicate on demand. The 32% IRR is impressive, but it's a rate of return, not a cash return — and the industry is wrestling right now with the gap between paper gains and actual distributions.

For a regular investor watching this from the sidelines, there's no direct play. You can't buy Blackbird. You can't buy Canva — it's still private, and the anticipated 2026 IPO has reportedly slipped to 2027.

What you can do is look at what this tells you about the market that Will matter when Canva eventually lists, or when other private companies with Blackbird-like backers go public.

The valuation cut from $42 billion to $34.9 billion, on roughly $4 billion in ARR, takes Canva's price-to-revenue multiple from about 10.5x down to roughly 8.7x. That's closer to what public SaaS companies actually trade at, and it's a reminder that private valuations set during hot rounds tend to come back to earth. The $65 billion peak valuation from late 2025 is now a headline from another era.

More broadly, the AI cost problem is a real structural shift, not a one-company issue. Software companies that build on frontier AI models are going to have compute bills that scale with user adoption. That changes the unit economics of the entire category. Companies that can reduce those costs — through proprietary models, caching, smarter routing, or pricing that passes costs to users — will win the margin war. Companies that can't will look like they're printing money to customers while losing it on the backend.

Blackbird's fund is a bet that the next Canva-sized success is still coming, and that the founders building it are in Australia and New Zealand. The valuation haircut to Canva itself doesn't kill that bet. But it does put the economics of AI-powered software squarely on the table, and the new fund will be tested on whether it can find companies that grow big without getting eaten alive by inference costs.

That's a hard filter. The investors who just wrote a billion-dollar check think Blackbird can clear it. The Canva write-down suggests they'd better be right about the next one, not just the last one.