|
WK651: We wrote a book. Want an early copy?
It’s been eighteen weeks since we wrote WK633. Where have we been? Heads-down building several things you’ll read about below. The irony of going quiet for four months while building tools meant to make us faster is not lost on us. Here’s what we’ve been thinking about since April. Why we wrote a book Airtable spent years as a fixture on every most-valuable-private-company list, last priced somewhere around $11 billion. It was just acquired for less than the total venture capital it raised, while sitting on roughly a billion dollars of cash, with comparable ARR, still growing about 20% a year. Huh? Unicorn status is supposed to be a signal: a rare and great outcome for everyone holding paper. Increasingly it signals larger funds buying late stage options versus indicating price. And the share of unicorns that ever reach an exit worth the paper is a falling well below a coin toss. Unicorn no longer means what you think means. Here’s the generalist VC standard right now, as far as we can tell: a story indicating zero to $20–30 million of revenue inside a year. Not “strong traction.” Not “great retention.” Narrative. And if you’re not able to tell that story, there is effectively no Series A market for a company that is merely doing well. Capital is going to stories with a shot at infinity, or to whatever a handful of well-followed people posted about yesterday. The end of the Series A and B as dependable milestones and the unmasking of the unicorn are both symptoms of one thing: the financing path most founders were taught to plan around no longer exists outside of the narrow AI stack. So we wrote a book; Speedstrapping, which is is our answer on how to build so that Plan A isn’t series A. In it we’ve refined thirteen years of learning about building for profitable growth, leveraging alternative capital, and building faster with AI, into something you can use to adapt to this new ecosystem. Hit reply and we’ll send you a draft copy in exchange for your notes. Learnings from our AI lunch and learns Since February we’ve been running AI show-and-tell sessions for the portfolio. The goal is to mutually demo what we’ve built, share workflows, and give each other practical tips to try. Here are two of our favorites so far. Brendan Hermalyn at Thalo Labs shared a live view of cash and plan-versus-actual stitched together from Ramp, QuickBooks, Gmail and Slack, an internal assistant, a DocSend replacement, and a lead-gen tool that reads legacy solar roof-mapping data to find HVAC units on rooftops. That last one is our favorite: someone else’s dataset, repurposed into your funnel. His advice for starting is to build whatever personally annoys you – he calls it “rage building” – and to write an AI usage policy early. Helena Merk at Lurk Research built a daily 7am job that reads her calendar and writes a prep brief for every meeting before she’s awake; a webhook fires when a meeting ends and handles the follow-up. Her sequencing advice is to do it manually first, add scheduled jobs when manual gets annoying, graduate to real-time triggers as needed. Here are four more lessons from the sessions:
If you’re building with AI and want to compare notes, reply. We’re collecting these for the book. Electrification is just better faster cheaper now In April, wind and solar generated more electricity than gas globally for the first time. Solar surpassed coal in the US. The binding constraint has moved to scaling: in California and New York the problems are permitting, interconnection, installation, and maintenance, not incentives. We made the security case for electrification back in WK633, and then it got tested. China absorbed a fuel supply shock by leaning harder on domestic renewables. We think batteries will continue to prove their value too, as Australia has shown. EV batteries are outlasting the predictions by a wide margin. Shaun provides a test case. He recently built a Lemons EV out of a Tesla Model 3 Performance with over 200k miles and an accident behind it. He put it back on the road and already it’s being tested at the track. The battery has lost some capacity, but even under extreme abuse it held up just fine. The hidden consensus on climate risk During the LA fires, $1.2M changed hands on Polymarket betting on how far they would spread and when they would be contained. A California-only market launched this year with the tagline “You can’t predict wildfire. But you can trade on it.” The Forest Service and CAL FIRE both say they use none of it. Kalshi won’t list fires but it takes bets on earthquakes and hurricanes. Fire is the one hazard a single person can start. Nine senators asked the CFTC to ban wildfire contracts in August, citing arson and insider trading. Meanwhile the World Risk Poll asked 140,000 people two questions: how worried are you about climate change, and how worried do you think everyone else is. In high-income countries about half call it a very serious threat. Only one in five thinks their neighbors agree. Betting on the risk doesn’t reduce it. The poll is the more useful number: the agreement already exists, and if you’re selling into this market you are probably underestimating it too. Portfolio Updates It was a heavy quarter. The short version of what we can share: Skycatch was acquired by Caterpillar, which is folding its spatial data capture and analysis into its mining technology stack. Mill was in Fast Company on using AI to attack restaurant food waste, and recently announced a commercial deal spanning Compass Group, Amazon and Whole Foods. Plentify is running South Africa’s first residential battery virtual power plant pilot. EvoLoh launched a 2.5 MW hydrogen pilot with 3M. Sonic Fire Tech got the Ars Technica treatment. If nobody credible is skeptical, you’re probably not doing anything hard. They also recently closed their seed round. Opportunities Shellworks wants warm intros to founder-led CPG brands in the US. Wasted is looking for its first paid WastedCan pilot and wants intros to multinational general contractors, military and defense infrastructure, and data-center developers. Multi-site operators especially. WastedCan is a waterless, climate-controlled sanitation hub for construction and remote sites that runs on power alone, no sewer or water, at well under the cost of a conventional restroom trailer Swellcycle is looking to fill three positions: Revenue Growth Lead, Operations Lead, Marketing Lead. Thalo opened up a new ML engineer position based in NYC. If you know someone wanting to bring ML + AI to the trades and HVAC, please let them know. See all open roles at jobs.thirdsphere.com. From us New York Climate Week is September 20-27. We’re not hosting anything this year as our AGM and SF Tech Week event is in San Francisco a few weeks later, but several of us will be around and we’ll share the events we’re going to closer to the date. Say hello if you’re there. Stanford GSB recently wrote a case study on how we invest called Third Sphere: Impact in Transition. Stonly is now Chair of the Impact Capital Managers Institute Board, which oversees ICM’s affiliated 501(c)(3) focused on research, education, and field building. The Institute is home to the Mosaic Fellowship and Turner MIINT (co-produced with Wharton Impact), and publishes research designed to strengthen the impact investing field. Capital Stack Compass has grown since we launched it in April with 300+ lenders and alternative capital providers. We’re now taking public submissions, and we’re using it to run real credit matches for startups beyond our portfolio. If you know a lender who should be in there and isn’t, add them. There’s still no time to waste. Best, P.S. WK651 = 651 weeks since we sent our first Third Sphere newsletter. P.P.S. Know founders we should meet?Forward this to them. P.P.P.S. Want an early look at the Speedstrapping book? Hit reply and we’ll send you a draft copy in exchange for your feedback. |
|
|