Hunter Powers

New startups today mostly don’t need VC money. Build costs collapsed...Distribution is accessible...AI turned small teams into highly leveraged operators. Reaching revenue no longer requires raising first. This creates a structural shift in venture itself. Mid-size VC funds ($100M–$1.5B funds) are the most exposed. Too large to write small experimental checks early. Too small to consistently lead or win competitive later rounds. They sit in the middle while the market moves around them. Meanwhile, two ends of the spectrum are strengthening. Large funds ($5B–$20B+ AUM) are the real winners. They provide scale, signaling, and the ability to deploy $50M–$300M checks once companies prove momentum. Micro VCs ($5M–$100M funds) will thrive for the opposite reason. Fast decisions. $50K–$1M first checks. Access before consensus forms. That said, venture math at the micro level gets tricky. In many cases, ownership, management fees, and outcome distributions make the economics challenging for GPs unless they consistently access outlier winners. The hardest place to be right now is the mid-tier. Too big for experimentation. Too small to dominate late-stage.

123

25 Comments

San Francisco-based DVC has launched DVC AI Fund I, a $75 million fund focused on Series A and B rounds. DVC is rewriting the venture playbook by replacing its analyst team with proprietary AI agents that handle deal sourcing and due diligence, analyzing 120 signals to build investment memos in minutes. The firm invests in core AI infrastructure, vertical model stacks, and application-layer solutions, leveraging a network of over 170 founders-turned-investors from companies like OpenAI, Google, and Tesla as LPs to mentor portfolio companies. The new fund is backed by notable LPs including Mike Arrington (Arrington Capital), Denis Yarats (Perplexity AI), and Andrew Filev (Wrike). The fund will be used to make early, low-friction investments in breakout AI startups, often providing capital months before a formal raise. Congratulations to founders Marina Davidova, Nick Davidov, and the DVC team. Tech Funding News: https://lnkd.in/dQKxCCrr Deals like this are reshaping the AI landscape. For weekly analysis on the funding, valuations, and strategic moves that matter, subscribe to the Charted Growth newsletter: chartedgrowth.beehiiv.com #AI #VentureCapital #VC #Funding #Startups #Investment #Tech #Innovation #SeriesA #SeriesB

17

Anthropic’s Claude Code just crossed a line most people are not ready to acknowledge. Yesterday, Boris Cherny, Head of Claude Code, shared that Claude Cowork’s entire codebase was written by Claude Code itself. Not assisted. Not reviewed line by line. Written. This comes right after he shared his own setup: multiple Claude agents running in parallel, coordinated through stop hooks, working for minutes, hours, sometimes days. And, according to another one of his posts, his numbers are striking: 259 PRs in 30 days. 497 commits. 40k lines added. 38k lines removed. Every single line written by Claude Code + Opus 4.5. That’s not autocomplete. That’s AI doing the work of an entire engineering team. What’s important here is not the speed, it’s the mode of operation. Claude isn’t producing snippets. It’s running continuously, holding context, iterating, shipping. This is what agentic development actually looks like in practice. And it raises an uncomfortable question for every CTO and founder: if an internal product can be built this way, what does “software development” really mean going forward? Of course, this only works in very specific conditions. Greenfield projects. Clear intent. Strong guardrails. Humans setting direction and constraints. But the direction is clear. We are (already) moving from humans writing code with AI assistance to humans designing systems, while AI does the execution. Claude Cowork is not a gimmick. It’s a preview. And it’s coming much faster than most organisations are prepared for.

63

52 Comments

The Google and NextEra partnership makes perfect sense. For years, the model has been: 1. Find a site. 2. Ask the utility for power. 3. Wait for them to deliver or build it. 4. Complain when they're late and over budget. That model is broken. Utilities can't keep up. (and it's not their fault... "keeping up" with demand is an impossible task in this case). This partnership flips the script. Google isn't asking for power. They're bringing their own. They're partnering with the largest renewable energy developer in the country to co-develop entire campuses. Land. Power generation. Interconnection. All bundled together. This is vertical integration. It's Google taking control of their own destiny. And it solves the biggest problem in the industry right now: ...Who pays for the infrastructure? NextEra's CEO said it himself: hyperscalers need to "bring and pay for their own power generation." That's exactly what this deal does. It's faster. It's more certain. And it doesn't stick ratepayers with the bill. This isn't just another PPA. This is the new model for the future of data center development. My question is... Will this just be a behind the meter design? Or will they connect their power generation to the grid? Comment below!

89

40 Comments

America cannot power its next century with the same grid it relied on last century.    This is very good news for the right energy founders. The increasing complexity of the grid financially, technically, and physically creates opportunities for cutting-edge software companies to break out, meet the moment, and build huge businesses.     Below are sixteen top startups leading the new energy economy from the front lines, divided into three broad categories.    -Utility Data: Arcadia, UtilityAPI, Bayou Energy, and Deck   -B2B Data and Simulation: Pearl Street Technologies, Texture, Bidgely, Halcyon, Innowatts, and Leap   -Electrification and Rebates: Commonweal portfolio company Pearl Edison (shoutout to Evan and Jake), 257, Sealed, Eli Technologies, Rock Rabbit, and David Energy   These three categories share the DNA of using software in novel ways to disrupt the traditional ways business was done in energy, whether by making siloed utility data more easily accessible to the long list of stakeholders who need it, by giving the long tail of energy companies ways to communicate and operate more productively, or by bringing powerful analysis to the residential level as America electrifies and policymakers roll out complex rebate programs. #AmTech   These startups have raised capital from the best climate and generalist investors including:   Aligned Climate Capital , Pear VC, Powerhouse Ventures, F2 Venture Capital, Cathay Innovation, Union Square Ventures, Keyframe, Equal Ventures, BoxGroup, Fifth Wall, Greycroft, MCJ, Voyager Ventures, New System Ventures, Abstract Ventures Inc., Lerer Hippeau, J.P. Morgan, Valor Equity Partners, Energy Impact Partners   This is part one in a series on #AmTech leaders shaping the future of the grid.    #AmTech #MeetTheMoment

122

21 Comments

In this week’s Capital Efficient, I unpack: # What it means for Vertical AI when foundation model providers start acting like enterprise software vendors. # Why agentic scaffolding might matter more than AGI timelines. # How founders can think about the “rebuild vs. retrofit” tradeoff when deploying agents. Link to the full post in the comments below.

28

3 Comments

Honestly, I have had enough of placid BS podcast episodes where people sit on the fence. “The 2:20 fee and carry model of venture is broken”. “VCs are spreadsheet monkeys”. “It is likely Putin will want more and to expand”. We cannot trust US tech, we need European sovereignty more than ever. My word, this does not lack opinion. Top 7 lessons from Taavet Hinrikus 👇 1. Why We Prefer Serial Entrepreneurs - They usually go for a bigger goal the second time around. - Torsten Reil built a gaming company before Helsing.  - Daniel Ek built an e-commerce company before Spotify. - Even if the likelihood of success is the same the outcome is so much bigger. 2. Why the Fee Structure of Venture Is Broken and VCs Should Not Get Rich From Deploying Capital  - The idea of collecting a 2-2.5% management fee does not align us with the outcomes. - We charge half of that at Plural & it enables us to make a few more investments. - You don’t want to hire people who are here for the biggest salaries. 3. The Fund That Makes Every Partner Invest Their Own Money With Every Check  - Every deal we do, the lead partner writes a personal check. - The size of it depends on the individual’s position. - We don’t like the idea of “playing with the house money”. 4. How This $500M Fund Makes Decisions - Every investor can only do a limited amount of deals a year, they have to commit their personal money & write a memo about the deal. - The memo has to include: - Why is it important to them? - Why is it important to the founder? - How can it be 100x from here? - Every deal we do, the partner has to be excited enough to be a co-founder of the company. 5. If It Is Not 100X We Do Not Want to Do the Deal - There are many great investments that are guaranteed to 5x. - They are not for Plural. - If we cannot imagine a 100x outcome we do not do the deal. 6. Change the GDP of Europe or No More Money for You - We tell our founders we’re here to have GDP levels of impact. - If your company is not doing it, we should not continue to fund it. - You can keep working on it, we’ll do our best to help you but it does not guarantee the next checks. 7. Why Europe Needs to Build Its Own Tech Now More Than Ever - Two things have happened: - The Russia/Ukraine war - Zelensky’s grilling at the White House - The US cannot be trusted to be the protector of Europe anymore.  - We need our own defence, space, & energy sector. Watch the full episode now by searching ‘20VC with Taarvet Hinrikus’ on YouTube, or click the links in the comments. #Founder #funding #business #investing #VC #entrepreneur #startup #investment #20VC

92

13 Comments

With RL techniques like GRPO, it’s become easy to fine-tune small models to outperform foundation models on vertical tasks. We’re open sourcing Osmosis-Apply-1.7B: a small model that merges code (similar to Cursor’s instant apply) better than foundation models. Links in comments! Using foundation models for high specificity, low complexity tasks like code merging can be overkill - instead, it’s better to use a specialized model instead for higher performance and latency. We fine-tuned the model on 100K real-world code commits where the model needs to merge a code edit snippet into an original code snippet. On the test dataset (subset of CommitPackFT), foundation model performance ranged between 0.77-0.93 reward score. In comparison, Osmosis-Apply-1.7B achieved a 0.98 reward score - while also being 3X-5X more affordable than the next cheaper model AND ~10X faster! Why this matters: the best AI agents use multiple models to ensure task specialization and reliability at scale - on top of cost and latency benefits as well. Links to the model and full write-up (with reward function, hyperparameters, etc.) are in the comments - reach out if you’re interested in using reinforcement learning!

239

18 Comments

An engaging read from Olivia Levine on why software engineering is no longer becoming the safe career bet from Greenfield Partners: What happens when 30% of production code is AI-generated? The review layer becomes the new battleground. As AI absorbs code production, the bottleneck has moved to human judgment, and a new category of "integrity platforms" is emerging to own this control point.

31

3 Comments

Vibe coding products like v0 (Vercel) / Bolt / Lovable have created ~$100M ARR in 6 months. Prediction: this will grow to $1B in the next year for this suite of tools. There are two market dynamics happening simultaneously giving us a glimpse into the future: 1) Consumer phenomenon: vibe coding is trending on TikTok, and a brand new market of creators is emerging. This is akin to the number of microsoft excel users around the world, but way cooler. You can produce a site or app in seconds with no need to write code. 2) Future of enterprise software: with the ability to seamlessly prompt a user interface, enterprise platforms like Salesforce and ServiceNow will see their rigid UIs open up and be personalized for each users own preferences. My bet is 30%+ of UI of each of these will be user generated.

157

31 Comments

Two trends in fundraising that I’m seeing on a daily basis: 1/ The "we'll monetize later" era for infrastructure startups is dead "We'll figure out revenue once we have users" / "we're focused on adoption first"/ "monetization comes later" This worked in 2021. It doesn't work now. Today’s market won’t reward research organizations or companies that believe in perfection over getting it in users' hands. If you don't have a clear path to revenue… you won't get funded. period. VCs aren't writing checks hoping you'll figure it out later. 2/ "Vibe ARR" is everywhere Remember Jasper? hyper growth, but the model companies ate them alive. We're seeing this again. Companies growing really quickly on point-in-time technology that isn't defensible These companies look amazing on paper. Growth charts going up and to the right. But if you think models are getting better and becoming platform companies, many solutions become obsolete. 

162

19 Comments

Check out the latest funding news from across the Techstars portfolio. → Zipline (Techstars 2011), a drone delivery and logistics startup, raises more than $600M in new funding, boosting its valuation to $7.6B, as it expands commercial deployments. → Language learning startup Preply (Techstars 2015) raises $150M and becomes the 23rd known unicorn from Techstars accelerators. → Hydrosat (Techstars 2019) raises $60M in Series B funding → Tive (Techstars 2017) receives a $20M investment, valuing the shipment-tracking systems developer at slightly more than $500M. → Parambil (Techstars 2023) raises an additional $6M in seed funding ($8M to date). → Seasats (Techstars 2021) receives $24M in Department of War APFIT funding to accelerate the fielding of its autonomous surface vessels. → Hawaiian Airlines’ newly announced $600M Kahuʻewai Hawai‘i Investment Plan features a strategic investment in Ampaire (Techstars 2018) to support hybrid-electric aviation in the islands. → Fintalo (Techstars 2025) secures $500K+ in oversubscribed Pre-Seed funding. → PraxisPro (Techstars 2024) closes an oversubscribed $6M seed round. → Samara Aerospace (Techstars 2024) closes $10M seed round to help bring more stability to sats in orbit. → MyARC (Techstars 2022), a platform that enables fitness creators to train their communities at scale, secures £1.5M investment. Massive congratulations to these incredible portfolio companies!   🔗  Read more at https://lnkd.in/dxtCTvrg 

154

6 Comments