“Congratulations. You are being rescued.” – K-2SO, Rogue One

 

Over the past year, we’ve embedded AI into our core operations, launching a suite of internal robots that make us faster, smarter, and more scalable. These include automated systems for generating custom questions to help us evaluate startups (QUEST), tracking portfolio performance in real time (PULSE), sourcing non-dilutive capital (MATCHA), matching talent (NOVA), and building co-investor lists (ALIGN). These workflows let us give more feedback to founders, support more teams, and surface insights that sharpen our decision-making. More on our current bots here.

More robots are on the way, including marketing and sales tools, plus guides on how to implement AI in company building in our forthcoming Seedstrap Playbook – the 4th version of our fundraising playbook that considers what happens when teams can be 50% smaller and grow 2x faster (hint – growth VC should be Plan B).

Growth VC is plan B

It’s been 588 weeks since our first newsletter. In most updates, we encourage founders to focus on metrics that would allow them to raise funds from growth VC – series A, B, C, etc. But our guidance is changing with the times.  We believe what was previously a fallback, “seedstrapping” is now the default path. Plan A. Seedstrapping emphasizes early stronger customer validation with shorter time to revenue. This unlocks non-dilutive capital, and minimal equity fundraising.

This approach grants founders greater control and resilience, especially in a world where AI is enabling multiple time and cost savings – lower operating costs, unlocking new paths to growth, and making discovery of alternative capital easier. In our portfolio, some of the most successful companies already followed this model to profitable, rapid growth and maximum climate impact, proving it can work even in “capital intensive” sectors.

Why now? The venture landscape has bifurcated into mega-funds and sub-$100M players, with growth VC becoming a specialized tool, not a default route. Mega-funds remain largely unproven, while growth VCshave increasingly exposed founders to punitive dilution during flat or down rounds. Meanwhile, private credit now rivals VC in size and scale, yet remains massively underutilized by startups.

Our revised playbook positions growth VC as a strategic option, useful in some cases, like financing roll-ups, or even FOAKs, but it’s no longer the presumed destination. For founders seeking control, optionality, durable outcomes and higher probability of success, seedstrapping is now Plan A.

 

Some Obsession Required

It’s tempting to imagine an “AI-first” future where every workflow is optimized and every team is ten times faster. But the data so far is much messier. It mirrors prior tech eras like early Internet adoption. AI adoption successes look more like Norway’s sovereign wealth fund or this solo founder or a 24 Hours of Lemons Projects and have something in common: obsession.

It’s not clear where and how much AI will work, so large scale testing is required. And then given new features and capabilities, those outcomes need to be retested. All the while, there is your regular day of work not getting done, as you play around – it’s a classic explore versus exploit problem. But experimenting can and should become a performance metric. CEOs, like Shopify’s Tobi Lütke, are baking AI usage into employee reviews. But just shipping .AI features is not a differentiator, and in some cases amplifies weaknesses more than strengths. And yes, you must beware the patterns of AI usage that make you dumb. But there is value in understanding how AI is changing business functions and how companies get built.

At our most recent portfolio roundtable, we asked founders to share results from AI deployed in sales and marketing workflows. LLMs are reshaping GTM strategies, though response to AI voice tests was more tepid. It’s not about ranking for search keywords anymore, but owning the “answer surfaces” that matter. Welcome to the era of AEO: Answer Engine Optimization. What was notable was the feedback – after 10 years of running these types of sessions, this was the most people who reached out and said “I learned a lot and we have some new things to test.”

 

What About AI & Emissions? 

A critical lesson from the Dot Bomb era was that adoption rates were overestimated in the short term and underestimated in the longer term. There were technical limits like connection speeds but there were also human limits, like the number of people willing to trust strangers to ship them something they’d purchased online.

It’s one thing to call AI revolutionary. It’s another to prove it, especially when the gap between valuation and utility keeps growing. In a scathing, data-rich critique, Ed Zitron dissects the fragile logic behind OpenAI’s $125B projection and questions whether this boom is already showing signs of burst.

This is very much a climate story too. As generative AI consumes ever-larger chunks of the energy grid, the tension between hype and infrastructure planning is becoming harder to ignore. As BNEF points out, one of the interesting stories is that renewables are often cheapest, can be deployed independent of the grid and have the shortest lead times – specifically it’s faster to get batteries, solar and wind online for datacenters, versus the next best option that uses dead dinosaurs, gas.

Whether we overbuild capacity, or barely keep up, it seems likely that we’ll get more renewables and battery storage.

 

Portfolio News

Near Space Labs secured $20M to scale its geospatial monitoring stack. Renewell continues its push to redefine energy storage, winning support through DOE’s GFO-23-317 program. Wild is turning heads with a radically new take on enabling mobility decarbonization. Singularity is now bringing real-time carbon transparency to MISO markets. Sention announced a $4.3M raise and is using imaging to better design and control batteries. Meanwhile, Gradient’s hybrid HVAC tech made it to NPR’s spotlight as a pragmatic, future-proof alternative to gas heating. Congrats to Kiwibot on becoming Robot.com’s new stewards.

 

Jobs and Opportunities

Ohm is looking for an Operations Leader to help scale nationally to ensure reliable EV charging experiences. Arbor has an opportunity to support their work in pricing in retail energy as they help more people save money on their electricity and gas around the US. ChargeLab is looking for a Full-Stack engineer to expand the leading operating system for leading EV chargers in the North America.

CYCLE is looking for a Sales Manager in the UK to help more corporate customers and couriers make use of e-bikes and e-scooters. Robot.com is looking for multiple robotics engineers to deliver on their promise of helping more corporate customers  make use of robots in a wide range of everyday jobs.

Near Space Labs is rapidly expanding their service for the highest resolution, highest frequency aerial imagery and is looking for an Imagery Production Supervisor. Pallon is bringing on an AI Sales Account Executive to bring automated infrastructure inspection to more cities in the Germany.

 

There are 193 opportunities available across Third Sphere portfolio companies. Please do forward and share.

 

Test drive our robots

Want to test some of our robots?

Send a company deck and we’ll get you back an ideal investor profile and our own initial questions for the company.

Send a resume and we’ll identify opportunities within the portfolio where you can help.

 

Best.
Stonly, Yana, Miela, Roscel and Shaun at Third Sphere

 

P.S. If you think of any founders we should meet, please hit reply and let us know.

Did someone forward this email to you? Here’s how you make sure not to miss the next one.

You can find our previous updates here.