A Figma advisor who joined after the seed round might have just made $65 million from their IPO for maybe a few hours of work per month. The math is brutal - advisors typically get 0.11% equity at seed stage, vesting over 2 years instead of the standard 4 for employees, plus single-trigger acceleration on exits. Most CEOs hire advisors then forget about them. 90% of advisors aren't worth it - they take similar dilution to a VP-level hire but only give you an hour per week.
Big Tech is spending $102.5 billion per quarter on AI data centers - more than all consumer spending contributed to GDP growth last quarter - and the scary part is how they're financing it. This looks eerily similar to the telecom bubble of 2000, except with one crucial difference: this time, banks are involved. Private credit funds, which barely existed a decade ago, are borrowing from banks to lend to data center companies, creating the exact kind of opaque debt chain that triggered 2008. If AI revenue doesn't materialize fast enough to pay back these loans, the correlations are high - they'll all default at once.
AI companies are hitting $100M ARR in under a year, but some of those stories unravel just as fast - growth spikes, customers churn, the business deflates.
The classic definition of risk treats it as volatility, but imagine being offered a stock that either gains you $100 or $1,000,000 - technically high volatility, but zero actual risk since you win either way.
It's time to ditch the system silos and slow workarounds. See how API integration can help your business work smarter—not harder. From cutting out manual work to accessing real-time insights across your systems, this guide explains how APIs help build a backbone for lean, agile operations.
Following its massive IPO debut, Figma's value surged to nearly $68 billion, with initial investors like Index Ventures, Greylock, Kleiner Perkins, and Sequoia holding $24 billion in stock. The IPO raised $1.2 billion, with most proceeds going to existing investors, although a lock-up period restricts major sales until January. This marks a shift for the tech IPO market, previously slow due to inflation and market conditions, with Figma standing out as a significant success story.
AI infrastructure company fal raised $125 million in a Series C round, led by Meritech and joined by Salesforce Ventures, Shopify Ventures, and Google AI Futures fund, valuing it at $1.5 billion. Fal specializes in running multimodal models for enterprises, such as image generation AI for ecommerce and advertising. Enterprise clients use fal to create infinite ad iterations and different product images, providing significant value through diverse asset customization.
After 2021's ZIRP era, tech executives swore they'd never miss another liquidity window - yet here we are with barely any IPOs despite markets at all-time highs. Since the election, speculative stocks like ARKK have outperformed the Mag7 by 5x. Companies keep waiting for the "perfect" IPO window while multiples march back toward 2021 levels, and recent IPOs like Rubrik and Circle show strong performance. The few who prepared early are capitalizing - Figma's is massively oversubscribed. In an industry where barriers to competition are falling and public markets are begging for new names to bid on, waiting for the perfect moment is the exact mistake everyone promised they wouldn't repeat.
David Cummings shares the memory of securing office space with a foosball table early in his startup journey. He recounts signing up for a credit card to purchase a foosball table after relocating to a new office space due to cash constraints. The story highlights the evolving nature of startup culture and the importance of cherishing early entrepreneurial experiences.
Executing a JTBD (Jobs to be Done) interview process revealed that interview difficulty stems from the need for ethnographic research skills and overcoming human memory limitations. The recommendation is to focus on recent buyers and conduct partner interviews to improve accuracy. JTBD insights at Commoncog discovered two customer paths to membership, revealing a predominant conversion via paywalled articles within series, challenging prior assumptions and guiding strategic focus.
Everyone's panicking about how to measure ARR with AI's usage-based pricing, but ARR was never that clean. Snowflake is entirely usage-based yet valued as recurring revenue. Datadog mixes flat fees with variable metrics. Even traditional SaaS has messy pricing - customers can be charged by seats, active users, and usage tiers all at once. What makes revenue "recurring" isn't whether it's fixed monthly, it's whether you can predict it with confidence. Your job is to define ARR honestly for your business model, not game it to fit some perfect definition.
The first time they tried stand-up comedy, this author strategically picked an open mic in a neighborhood where nobody knew them - failure would be anonymous. The set went well. The second time was different - now people had expectations, and expectation brings its own terror. Success creates a special hell where you're not afraid of being a fraud, but afraid you're genuine and still can't summon that magic again. The antidote isn't caring less but caring more about truthseeking - whatever that meant before anyone was watching.
If vibe coding is creating software by talking to AI, vibe analysis is getting insights the same way - and it might be even bigger. There are 2 million software engineers in the US, but at least 5 million people who pull data to answer questions every day. That's 10 billion hours annually spent on reports, dashboards, and deciding which experiments to ship.
iCapital raised $820 million, raising its valuation to $7.5 billion. The company plans strategic acquisitions and tech investments. Co-led by SurgoCap Partners and T. Rowe Price, the round included existing backers like Temasek and UBS. iCapital's platform manages $945 billion and supports private market investments and asset management.
Value used to be about what software could offer (think Microsoft Office licenses). Then it was about what it enabled (Adobe Creative Cloud). With AI, we are entering a new era where THE VALUE OF SOFTWARE IS WHAT IT ACTUALLY DOES (think OpenAI charging by tokens used to generate content).
Fireworks AI surged to a $4 billion valuation with its high-performance AI inference technology, optimized for PyTorch and open-source models. The company provides faster and more cost-effective deployments than traditional cloud providers, with partnerships including Google Cloud and NVIDIA bolstering its position. Upcoming funding will support further development of "compound AI systems" and global expansion, positioning Fireworks for a potential IPO.
Quince, an affordable luxury DTC brand, raised $200 million at a $4.5 billion valuation, doubling in six months. Iconiq Capital led the round, boosting Quince's total funding to $461.5 million. Despite challenges in the D2C sector, Quince stands out with fast-growing revenue and popularity among Gen Z and millennials due to its affordable pricing.
Dario Amodei, CEO of Anthropic, has been vocal about AI's rapid progression and its potential risks and opportunities, predicting significant job impacts while advocating for AI safety measures and regulations. Anthropic, valued at $61 billion, has shifted focus from consumer applications like ChatGPT to supplying AI models to businesses, primarily through APIs, which has driven fast revenue growth. Amodei's dedication to AI safety stems from personal tragedy, and he navigates industry skepticism while pursuing rapid AI advancements, believing in the transformative potential of AI, particularly in medical and scientific fields.
Karri Saarinen hated Jira so much at Airbnb that he built a Chrome extension to redesign it - 100 coworkers installed it. That hack became Linear, now worth $1.25B. The Finnish founders spent a year validating through bar conversations before building. They gated access obsessively, handpicking 10 users weekly from 10,000 waitlist signups. While competitors chase flexibility, Linear ships opinionated defaults. It hit profitability in 2021 with just 80 employees by choosing quality over growth metrics.
You abandon writing a $5,000 sales page to answer a dinner text because your brain prioritizes time-sensitive tasks over important ones. Researchers proved this by offering people five Hershey's Kisses or three - when they added a 10-minute deadline to the smaller reward, participants choosing it jumped from 13% to 31%.
Most developers didn't even know GitHub had paid features - they thought it was just open source. When Thibault Imbert watched users sign up, they picked free by default because GitHub hid what paid plans offered. Inspired by Grammarly's transparent signup flow, his team redesigned everything. Instead of hiding paid features, they showed exactly what was free versus paid, segmented by user type, and wrote copy like developers talking to developers. The results shocked them - 35% more paid conversions, worth over $1 million in ARR. They were so surprised they double-checked the analytics. CEO Nat Friedman loved it because it felt natural, not pushy. The lesson is counterintuitive - being honest about pricing actually increases conversions when you respect your users' intelligence.
Mantis VC, co-founded by The Chainsmokers, raised $100 million for its third fund, 25% more than its previous $80 million fund. Despite its celebrity status, Mantis focuses on B2B startups, with investments in companies like Chainguard and Rogo. It leverages its extensive network from private shows for Fortune 500 companies for valuable customer introductions.
Everyone thinks AI agents will kill systems of record - why log in Salesforce when agents do everything? David Peterson thinks that's backwards. Agents need constraints and permissions, making systems of record essential. However, the pricing power shifts. Companies will pay for work, not data storage. When incumbents throttle APIs to maintain revenue, customers revolt. The winners won't be today's systems desperately locking down data - they'll be whoever builds ground-up "systems of intelligence" without legacy baggage.
2025.07.30·
newsletter.angularventures.com·4 min read·TLDR Founders
Anthropic got backlash for adding usage limits to Claude Max, following similar issues with Cursor and Windsurf. It's always coding tools that face pricing revolts. The problem is simple - coding requires flow state, not paywalls interrupting a 5-hour feature build. Running out of credits in an SEO tool works because it's discrete tasks. However, limits mid-debug ruin the IDE experience. These companies aren't being deceptive, they're stuck. AI costs too much for unlimited subscriptions that developers want. Like expensive mobile data in 2003, this will look ridiculous once cheaper models make unlimited usage possible.
Coinbase's $65M observability bill showed how fast costs can grow. They're not alone: observability costs are spiraling out of control [5] across industries.
Trump scrapped Biden's AI safety rules to let markets "win the AI race" - but forgot to include a game plan. The 28-page blueprint backs open-source AI and cuts regulation while keeping technical standards. Good for researchers who need tool access. Bad for implementation: 103 vague actions with no deadlines, no funding, and a third don't even name who's responsible. They'll track job losses but ignore AI fraud and discrimination. Stanford HAI says without details, even good ideas like regulatory sandboxes will die in bureaucracy.
Palantir's stock surge has made it one of the 20 most valuable U.S. companies, with its market cap reaching $375 billion, surpassing Home Depot and Procter & Gamble. Investor confidence is driven by its robust AI segment and its strengthened ties with the U.S. government, reflected in a 45% revenue increase from its government business. Despite being valued at 273 times forward earnings, the company remains small in revenue compared to peers in the top 20.
Investors who poured $104.3 billion into AI startups in 2025 are now seeing lower returns upon exiting, with just $36 billion in cashouts reported. Despite this disparity, AI funding continues to surge, led by massive investments in companies like OpenAI and Scale AI. There are concerns of an "AI bubble" echo, with some major players like Microsoft pulling back, signaling caution around overvaluation risks.
Churn rate literally sets your maximum company size - divide new MRR by churn to see your ceiling. Buffer discovered this brutally: at 6% churn and $8,400 new MRR monthly, it was mathematically capped at $140k forever. Even tripling marketing couldn't help because exponential churn always catches linear growth. When COVID pushed churn to 7%, revenue crashed 35% overnight. Here's the magic: cutting churn from 6% to 3% doubles your ceiling instantly. Buffer finally grasped this, reduced churn, and growth exploded.
Reddit isn't just about Google rankings anymore - it's training the AI that your customers use to research products. One First Republic Bank user found something surprising. Reddit visitors converted just as well as Google visitors. However, paid Reddit ads completely failed. The difference is authenticity - Reddit users instantly reject anything that feels like marketing. The real value is using Reddit as a research lab. You can see exactly what customers ask about your competitors when no one's watching.
Managing product feedback is challenging due to its messy, disorganized nature, and often low-quality submissions. It requires significant effort to gather, organize, and prioritize feedback, differentiating between valuable signals and irrelevant noise. Despite these challenges, understanding customer feedback is vital for building products that users love and can significantly impact a company's success.