The Venture Apocalypse
Gil Dibner

At the end of August, I returned from a few weeks break in the US. This past weekend, as I emerged from the depths of post-break catch up, the venture world seemed gripped by a wave of apocalyptic thinking. Over the weekend, I whatsapped a VC friend to see how they are doing. “Weird times,” they wrote back, “and apparently all funds will die except for two big ones.” at first, I didn’t understand what they were talking about. That, plus realizing that I had completely missed the “founder mode” moment, suggested to me that maybe I’d taken my week-long hiatus from social media a bit too far.

Reluctantly, I dove back in and found the conversation. To my relief, the “founder mode” thing is a nothing-burger. Just a buzz-wordy rebranding of some very old and perennially true ideas about agility and grit.

But the hand-wringing about the venture apocalypse seems much more significant. My friend was referencing an old debate that seems to have been roiling the VC world over the past week or so. The slowdown in funding for emerging managers has been going on for some time, and is not a surprise given the excesses of recent years. A lot has been made of a recent report that two firms (A16Z and General Catalyst) captured nearly 50% of US VC fundraising in 2024. One of the first to connect the dots in a thoughtful way was Sam Lessin who’s been making the point that the days of “venture capital on easy mode” and the Factory Model of Venture are over. On August 25, while I was enjoying some downtime with my wife and daughter, Josh Wolfe of Lux Capital, released his 2Q Letter to LPs which (as part of talking his book) made the prediction that LPs will “adjust their allocations…leading to widespread contraction in the number of venture firms and consolidation of LP dollars into fewer, name-brand managers with reputations for both stewardship in up and down markets as well as mentorship of an emerging generation of investors who have earned the right to partner with the most ambitious entrepreneurs attracted to the firm’s and the individual partner’s reputations. Far from being a democratized asset class, VC is and will remain a rarified ecosystem where only a select cadre of firms consistently access the most promising opportunities. The vast majority of new participants engage in what amounts to a financial fool's errand. We continue to expect the extinction of as many as 30-50% of VC firms.”

Over the past week, the temperature of this conversation rose even further. Rick Zullo at Equal Ventures wrote about the Extinction of Venture Capital, expressing his concern for the current generation of emerging venture managers and what this means for the diversity (in every sense) of the venture capital ecosystem. The always brilliant Frank Rotman of QED argued in a longform post that Zullo’s fears and Wolfe’s predictions are directionally correct. Frank’s conclusion is that “emerging Managers will do best if they position themselves firmly within one of two key boxes: The Solo VC or the Non-Consensus Alpha seeker.”

Beezer Clarkson, an experienced and outspoken LP with a long track record of supporting emerging managers, joined the discussion in a Twitter (X) thread. She sounded a more optimistic note, arguing that neither venture nor emerging managers are going extinct. Instead, she believes that a cyclical slowdown in exit markets is causing LP capital to contract, which - of course - is putting a healthy evolutionary pressure on venture managers. She cited the launch of Nick Chirls; new fund, Asylum Capital, as an example of next-generation back-to-basics venture. Chirls waded into the debate himself on the 20VC podcast, and as of yesterday - the discussion was still going on, with some spirited Twitter (X) discussion between Chirls, Keith Rabois, and Jack Altman.

Taking a step back, the state of play right now is that the entire venture world is wrapped up in a deep conversation about the future of venture capital and entrepreneurship itself. This conversation has been triggered by a crash that many (including Angular) saw coming long ago. We called it relatively early and have been very vocal about practicing a back to basics approach to venture capital. This global conversation that the venture community is having with itself has intensified over the past few weeks, and I expect it to get even more vigorous as capital continues to dry up and the pressure rises. At the very same time, however, megafunds will continue to pump ever larger amounts of capital into perceived blockbuster companies. The dichotomy of a raging party going on alongside this level of deep introspection is remarkable.

My own suspicion here is that while some players seem to have changed lanes, the fundamental basics of venture capital and entrepreneurship remain the same. Frank Rotman, for example, is correct that “non-consensus Alpha-seekers” will be the most performant (and, therefore, sought-after) VCs. But I don’t think this dynamic is new. The ZIRP era may have allowed some VCs with questionable strategies to post temporarily inflated returns. But “non-consensus Alpha” was always the name of the game in venture. The fact that many VCs (and many LPs) forgot (or ignored) this truth is anomalous. Similarly, the essence of early-stage non-consensus entrepreneurship has also remained fundamentally unchanged: five to fifteen people, working for two years, at reasonable salaries, building something of value that will trigger another investment and/or generate customer revenue. The fact that some “venture” firms are now managing billions and writing $100M seed checks doesn’t mean the nature of seed has changed. As my partner David commented on the recent $1B investment into Ilya Sutskever’s ‘safe’ AI start-up, “a16z and Sequoia are full of very smart investors but…whatever that round is, it’s not a seed round.”

Travel update:

• Gil will be in Copenhagen this week at TechBBQ (Sept 10-11), leading a panel on founder-led sales.

• David & Gil will both be in Malmo next week (Sep 16-17) for the Drop, a conference on advanced climate tech. David will be continuing on to Stockholm for the remainder of the week. Hit him up if you’re around!

• David will be in Dublin for SaaStock on Oct 15-16 where he will be speaking on the ever-changing milestones needed to raise venture financing.

No Sleepwalk to Success
Engineering success in a technical startup.

Revenue Durability in the LLM World
Everything about LLMs seems to make revenue durability more challenging than ever.

A Robotic Future for Retail Grocery, Finally.
Why we invested in Finally.

A Digital Fabric for Maritime Trade
Why we invested in Portchain.

More blogs

Israel / SaaS. Own was acquired for $1.9B by Salesforce for its tools for backing up data in cloud-based applications.

UK / SaaS. Doccla raised $46M, led by Lakestar, to fund the European expansion of its vertically integrated (software plus hardware) “hospital at home” platform.

Netherlands / ClimateTech. Carbyon raised $17M, led by Siemens Financial Services, Omnes Capital and Global Cleantech Capital, for its direct air capture technology.

Germany / SaaS. Reonic raised $14.5M, led by Northzone, for its sales platform for solar and heat pump installers

Sweden / Fintech. Palm raised $6.1M, led by Speedinvest and Target Global, for its enterprise cash management platform.

Belgium / Hardware. Cowboy raised $5.5M, led by Cypress Capital, to provide bridge financing for the electric bike manufacture, ith plans to reach profitability next year.

Italy / SaaS. Skillvue raised $2.8M, led by KFund and Italian Founders Fund, for its AI-based skill assessment platform.

ENTERPRISE/TECH NEWS

Foundational model companies pivoting. Foundational model AI startups are throwing in the towel, pivoting to other AI-related ideas, amid the massive investments by Big Tech giants. As Aleph Alpha (a German LLM company that raised $500M last year) CEO said: ““The world changed. Just having a European LLM is not sufficient as a business model. It doesn’t justify the investment.” Training a competitive model costs $100s of millions now. That number is only going up. And when you’re facing competitors that can invest billions of dollars (which, by the way, is just a small fraction of the cash they’re spinning off from their core businesses), you have to ask yourself if you’re in a game you can ever, ever win.

Is SaaS too expensive? Klarna announced this week that it has already shutdown its Salesforce instance, and plans to shut down Workday within a few weeks. “We are shutting down a lot of our SaaS providers,” Sebastian Siemiatkowski (Klarna CEO) said. With the help of AI, the company has been able to create a more lightweight tech stack which has enabled them to operate “more effectively with higher quality.” David Heinemeier Hansson made a similar point this past week when he wrote about how, at 37Signals, they’re rolling off Datadog and replacing it with a combination of Prometheus, Grafana and Kibana. “Our renewal bill for Datadog came to ~$83,000/year before we canceled. There's no way, no how the time we've spent replacing the key usage amounts to 1/3 or 1/2 of a full-time employee. Not even close. Enterprise SaaS pricing is bonkers.”

Llama ascendant. This past week, Meta announced that Llama models have been downloaded 350 million times (up 10x in the past year) and Llama usage doubled from May to July 2024. There are more than 60,000 derivative models on Hugging Face, suggesting there is indeed a vibrant community of developers building on top of Llama for their own use cases. This is a press release, so read it as such, but Meta also includes a handful of case studies from the likes of AT&T, DoorDash, Goldman Sachs, Shopify, Spotify and Zoom. (Even more bullish is the fact that both Infosys and KPMG are reportedly using Llama internally. In my opinion, the way these models gain ubiquity in the enterprise in the near term is via consultants, so buy-in from Infosys and KPMG is a big win for Meta.)

DOJ vs. Google. The DOJs second antitrust case against Google begins this coming week. The last case, which Google lost, was about alleged monopolist behavior in Google Search. This case is all about Google’s advertising technology, and whether or not Google used its advertising technology to create a monopoly that ultimately led to higher prices for customers.

Time to build. If you haven’t used Replit’s new coding agent, I’d give it a shot. It’s pretty magical. See this tweet for just one example of thousands from the past week post-launch: 

This is exactly what David was writing about a few weeks back in his piece “The Age of Artisanal Software is Over.” As David wrote: “As non-technical people are empowered to write more of their own code at the margins, their willingness to pay for new software will go down. They’ll only pay for something that they couldn’t build themselves.”

HOW TO STARTUP

Founder mode. If you weren’t still on vacation or on a self-imposed social media hiatus (both of which I would applaud you for!), you’ve probably already read about Paul Graham’s new essay on founder mode. It’s worth a read, if for no other reason than to know what everybody is talking about. I think Chamath Palihapitiya had a good reflection on the matter: “There is first principles management and stupid management and I've seen founders and managers, alike, exhibit both. When your company isn't working, the only solution is to take the time, as laborious as it sounds, to break your business down to its core essentials and rebuild it from the ground up with zero nostalgia or loyalty to people or technology. There is a certain ruthlessness that is required and it isn't borne in a title but in a psychological makeup. Some have it and some don't.” That take won’t get you the same amount of virality, but it sounds about right to me.

A vertical future. OnlyCFO makes the case for why the future of software will look even more vertical than it has over the past few years. They include some great data in the piece to back up the claim (e.g. lower churn, higher GTM efficiency), but to me the most significant point is the fact that generative AI, which makes development that much faster, may make new verticals more interesting than before. The faster and cheaper you can build, the more a niche vertical, that previously might have seemed too small to build a solution for, might make sense as a target. Long live vertical SaaS!

To build pipeline, look to the past. A useful roundup from Kyle Poyar of OpenView on ideas for building pipeline in 2024. The top themes were as follows: throw events, run account-based plays, and get creative with outbound. What’s the theme beneath the theme here? All these ideas are in response to the massive (negative) impact that generative AI is having on outbound. Any standard outbound playbook doesn’t work anymore, as it can be replicated with one of countless “AI SDR” tools. What works? Either getting in-person or getting hyper personalized. Building pipeline in 2024 won’t be easy, and it might end up looking way more like 2004 than you might have thought.

HOW TO VENTURE

Super seed. Safe Superintelligence, OpenAI co-founder Ilya Sutskever’s new safety focused AI startup, raised $1B at a $5B valuation. This was its first round of financing (though I think you’d be hard-pressed to call it a seed round!). The round was financed by a consortium of investors including Andreessen Horowitz, Sequoia Capital, DST Global and SV Angel.

Another reverse acquihire. Amazon hired the founders of Covariant in another “reverse acquihire” this past week. As part of the deal, Amazon will receive a non-exclusive license to Covariant’s AI models. Covariant will continue to operate, but with a heavily hobbled staff. The three cofounders are all leaving to join Amazon, as well as a quarter of the company’s employees.

Root Signals raised $2.8M for GenAI auditing, led by Angular Ventures, with participation from Business Finland.

Aquant’s CEO Shahar Chen shared how service organizations can navigate the complicated AI landscape.

Aquant
Senior Sales Engineer, EMEA (Germany Remote)

Candu
Back End Developer (London)

CruxOCM
Senior Software Engineer - Platform (Remote)

Vault Platform
Full Stack Software Engineer (UK Remote)

More jobs