Venture Capital

Some investors backing OpenAI and Anthropic are also funding Neo Labs as a portfolio diversification strategy, betting on multiple approaches to AGI rather than concentrating in the current leaders.

Guy Siri and Alex describe venture as identical to A&R: maniacal focus on niche, blinders against herd noise, gut-driven conviction honed by pattern matching across cycles, and willingness to lead unpopular rounds (Anthropic SPVs were 'tough to fill') — a process they argue transcends market timing.

Diamandis/Blundin fund concentrating on singularity infrastructure: GPU asset class (Orin), AI longevity (Insilico), living products (NeoGenesis), autonomous coding (Blitzy) — all hitting exponential milestones.

All Raise data shows growth from 9% to 18.6% (2,500 women) in 7 years, but 'partner' title no longer guarantees deal-leading authority, carry, or IC voting power. Downturns disproportionately affect diverse hires ('first in, first out' risk). Female GPs face higher LP scrutiny on risk despite comparable track records. Data shows diverse teams generate alpha via unique networks/perspectives — opportunity for LPs allocating to emerging diverse managers.

Guy Siri translates music talent-spotting — identifying the 'chorus' in a founder's pitch, acting fast before competitors overpay, trusting gut over consensus — to venture, citing early OpenAI/Anthropic/Hugging Face bets that others passed on.

Early-stage funds focused on physical AI (compute, energy, defense, manufacturing) are generating strong realized returns as portfolio companies reach public markets, attracting LP capital despite broader venture slowdown.

Market downturns cause firms to revert to relationship-based decisions over performance data, leading to slower promotion and higher exit rates for women and non-binary VCs even with competitive track records, while LP fundraising bias persists against diverse fund managers.

Lior Susan reports strong LP appetite for new funds targeting the analog-to-digital transformation of physical industries, citing successful exits like Cerebras and potential SpaceX liquidity as proof points.

Some investors funding neolabs have also backed established labs like OpenAI and Anthropic, suggesting a diversification strategy to avoid over-concentration in any single AI lab. The $2.5 billion flowing into neolabs reflects VCs hedging across the frontier model landscape.

Guy Siri translates music talent scouting (signing Alanis Morissette, Muse on first song) to venture: maniacal focus on founder 'chorus', ignoring market noise via 'horse blinders' (David Geffen advice), and trusting gut over consensus — resulting in early OpenAI/Anthropic/Hugging Face positions now worth ~$1B combined.

All Raise data shows women and non-binary VCs doubled to 18.6% of decision-maker roles in seven years, yet title inflation masks true influence (lead deals, carry, IC voting). In the current VC contraction, diverse investors with strong track records face slower promotion and harder fundraising — LPs ask more risk-focused questions of female GPs versus male peers with weaker records. Data shows diverse teams generate alpha via unique networks and perspectives, but relationship-based decisions in downturns reverse progress.

Many traditional software VCs express interest in robotics but refuse to invest in hardware, forcing founders to either find specialist firms like Eclipse or structure as software plays; this expertise gap persists even as physical AI hype grows.

Fox Glove CEO Adrian McNell observes that many traditional software VCs are 'robotics curious' after Jensen Huang's physical AI comments, but few have hardware investing capability; this expertise gap favors specialized firms like Eclipse and creates fundraising challenges for robotics startups.

Investors who already backed established labs like OpenAI and Anthropic are funding neolabs partly to diversify their AI bets and avoid over-concentration in any single lab, even though the competitive viability of these startups remains unproven.

Traditional software VCs are exploring robotics/physical AI but often lack hardware investing capability. Companies structured as familiar SaaS infrastructure (margins, pricing) can access broader capital while still capturing the physical AI tailwind.

Guy Siri applies A&R intuition — acting fast on 'chorus' moments, ignoring herd noise with 'blinders,' and trusting gut over consensus — to identify generational AI platforms, acknowledging the bar rises after $1B outcomes in Anthropic/OpenAI.

Founders experiencing rapid wealth creation (seed to large rounds in months) need guidance to focus on long-term substance and principles rather than valuation; Eclipse advises ignoring valuation marks and building for decades, with opportunity cost of not starting companies now being very high.

Early conviction in foundational labs (near $1B deployed) created a high watermark; today's crowded landscape requires more work and pattern recognition to distinguish signal from noise, but the core instinct — back visionary founders early — remains the same.

NATO has stepped up venture capital efforts to work with private businesses and maintain technological edge, recognizing that most advanced technologies (AI, cyber, space) are now developed by private companies like SpaceX rather than government programs.

Founders increasingly object to early-stage competitive investments (Index/Town/Instinct), while late-stage funds use massive secondaries ($170M for Wonderful) and flexible terms to win deals — a trend that will escalate beyond current 'peak crazy' structures.

Jordi Hays argues Miro's 90% valuation drop from $17.5B to $1.79B signals a broader late-stage venture bubble where many current high-valuation companies will face similar discounts in three years.

Current barbell problem — strong early-stage activity but weak incorporation and exit infrastructure — forces European founders to incorporate in Delaware for IPO readiness; solving the '28th regime' corporate law and insolvency harmonization would keep value creation in Europe and deepen local exit markets.

General Catalyst's Customer Value Fund treats sales & marketing spend as an asset class, providing non-dilutive, non-recourse capital that lets founders redeploy equity cash to product and M&A.

The best venture firms apply a positive-impact filter (e.g., rejecting prediction markets as 'another casino for Americans') and seek cult-like founder-investor philosophical alignment, believing that discarding values for returns is the default (80-95% of investors) but leads to misallocation.

The AI capital cycle has shifted: mega-cap AI companies (Magnamopa) hold $20T+ liquidity and recycle it via investments/rev-share deals to lock in corporate/sovereign partners. Traditional VC is bypassed; AI entrepreneurs and employees become the largest new capital source as Anthropic/OpenAI/SpaceX go public.

Private company founders retain freedom to invest in multi-year innovation without daily share price pressure, while public listing introduces short-term evaluation metrics that conflict with the time horizons required for breakthrough R&D.

Klein argues Europe's startup deficit stems from regulatory fragmentation requiring separate legal entities per country — not funding gaps — pushing founders to incorporate in the US where unified markets enable faster scaling.

Andreessen argues large institutional investors face a structural shortage of high-growth deployable opportunities, and should concentrate capital and influence on the few sectors and founders capable of generating real economic growth, particularly in defense, energy, healthcare, and education.

Ek contends Europe has ample capital, talent, and knowledge (via YouTube/podcasts) but lacks enough experienced operators who have 'been there, done that' to mentor founders, and a cultural belief that outlier success applies to them — success begets success through role-model density.

Singular's strategy of backing founders before market consensus, evidenced by over 50% preemption rate, allows the firm to secure allocation in category-defining companies like Aikido before they become hot.

Ex-founders from Stripe, Wayflyer, and Tines are becoming angel investors and starting new companies in the UK and Ireland, creating a dense talent network that early-stage investors can tap through accelerators like Dogpatch Labs and Ormeau Labs.

8VC's new $1.5B fund (equivalent to $400M a decade ago) deploys 30% into company creation, with portfolio companies growing 5-10x per year, reflecting accelerated pace of value creation in defense and AI.

Late-stage growth rounds at $1-3B valuations on $100-200M ARR have consistently marked up 3-4x within quarters; 40% of 2025 unicorns marked up by mid-2026. Sequoia's tranche rounds signal capital rushing in, compressing future returns.

Chinese AI startup raises $3.5B at $35B valuation, immediately targeting $50B pre-money round ahead of Hong Kong IPO; non-traditional staggered fundraising reflects intense investor demand and closing IPO windows.

Startups should set quarterly hiring pace based on real-time leading indicators (product-market fit and go-to-market fit metrics) rather than annual capacity models; green lights accelerate, yellow holds, red triggers emergency fix before continuing.

Venture returns follow a extreme power law where virtually all value comes from a handful of $100B+ outcomes; investors rationally ignore downside protection and only fund founders pitching maximum upside, making ambitious visions easier to fund than modest ones.

Expedition's edge comes from intensive outbound networking in talent-dense communities, building relationships with leaders, and maintaining a deep bench of active servicemen and special forces veterans who combine military domain expertise with business acumen to help portfolio companies decode fragmented European government go-to-market.

Being an early backer of OpenAI or Anthropic has become the premier fundraising story in venture; Khosla Ventures is raising $5.5B across funds just a year after a $4B raise, deploying capital faster than ever with focus on AI applications, robotics, climate, and hard tech, signaling intense competition and expense in venture capital today.

Firms like Meta, Center Systems (SpaceX alumni), and Hrian are building 4-week training for high-paying skilled roles (wire harnesses, plumbing, electrical) because labor markets demand it — private-sector training scales faster than government programs and directly links to hiring needs.

Nvidia's investments in CoreWeave, Lumentum, Coherent and others are not vendor financing but strategic capacity-building; these stakes will appreciate as portfolio companies expand to meet AI infrastructure demand.

Equity investors are increasingly interested in capital-intensive business models — energy, infrastructure, deep tech, robotics, roll-ups — that were previously considered unattractive but are now seen as having a moat. This shift pulls debt financing alongside equity rounds earlier in the company journey, even at pre-revenue stage for companies with hard assets or contracts.

The multi-billion dollar returns from SpaceX and Anduril have demonstrated that capital-intensive, long-horizon hard tech companies can achieve venture-scale exits, prompting generalist firms like Thrive Capital and Valor Equity Partners to allocate to nuclear startups despite traditional 10-year fund timelines.

Series A ($1-3M ARR at 200-400x) offers worst risk/reward — minimal PMF signal but huge step-up from seed; Menlo responds with barbell: expanded seed program (3 partners, $8M checks on spot) and inflection/growth focus on clear winners (>$10M ARR, anointed leaders), accepting lower ownership for access to outliers.

Despite YC's growth, the batch/pod structure preserves the early intimate experience; the single best predictor of startup success remains shipping pace, unchanged by AI tooling.

Investors can rationally allocate to high-volatility funds like Situational Awareness if sized appropriately within a diversified portfolio, as losses in one fund may be offset by gains elsewhere.

Leverage forces instantaneous, correlated liquidations across positions, whereas venture portfolios without leverage can right-size companies over years, letting winners like SpaceX offset losers like Bird at different times.

Sindana Capital's $3B fund-of-funds wins by leading 80%+ of its seed funds, betting on managers with structural founder access (scout programs, university pipelines) rather than picking skill alone, and benefits from power-law returns where a single 300x fund (Founders Fund 2) dominates portfolio economics.

Every 15-20 years a platform shift enables new startup waves; AI requires a new application layer translating model capabilities into domain-specific workflows (legal, finance, marketing, HR), creating opportunities across the stack from agent-first apps to infrastructure.

Early-stage funding frenzy for world model startups with aggressive 1357 valuation targets contrasts with difficult IPO paths and scarce M&A exits in China, risking VC returns.

Regulatory changes across Europe are enabling pension funds to allocate to higher-risk assets like venture and growth, unlocking massive dormant capital pools that have historically been sidelined, similar to the 40% pension/endowment LP base in US venture.

LP conversations have shifted dramatically in 12 months: early focus on DPI/liquidity has given way to frustration that large multi-stage funds now target 2-3x net (essentially venture beta), while smaller boutique funds like Axiom still underwrite for 10x+ outliers — the only justification for venture's illiquidity and high failure rate.

Sequoia replaced Monday partner meetings with anytime decision rights, explicit co-steward titles, and a culture where influence flows to expertise not tenure — data shows contentious deals perform equally to consensus ones, only conviction matters.

Large VC firms operate as index funds with capital deployment quotas, incentivizing follow-on investments in capital-intensive neoclouds that later pivot and face markdowns when GPU supply-demand normalizes, harming founders who over-dilute.

Clean 1x non-participating preferences ensured all Airtable investors recovered capital despite 90% valuation drop; demonstrates downside protection value in frothy markets, but also shows late-stage investors accepting money-back outcomes rather than 10x venture returns.

Top-100 private names offer efficient secondary liquidity at small discounts or premiums; selling 20% to return 25% of new fund capital beats waiting 6+ years for IPO+lockup.

Coastal Ventures argues sustainable returns come from early, high-ownership bets (e.g., Rocket Lab at $5M for 33%), not aquihire-driven late-stage rounds where investors don't recover headline prices; failure rate must be 60-70% to justify risk premium.

The neo lab category (AI research lab spinouts raising $100M+ seed rounds at $1B+ valuations) presents severely skewed risk/reward for most deals due to absent business models and technical clarity; VCs should pass on the vast majority and only back generational founding teams pursuing massive markets, using Anthropic's trillion-dollar outcome as the benchmark for potential returns.

Benchmark's edge comes from passing on 'investment-grade' opportunities lacking founder chemistry, using the 'green button test' (would I take a 9pm Saturday call?) and 'life's work' filter (could I recruit a friend to join?) — optimizing for 1% better decisions compounded over decades.

SPC's new fund reflects a structural shift in founder ambition — from limited software ideas to nuclear-powered cargo ships, alternative space launch, and US infrastructure transformation — requiring larger checks and high-conviction partnership further down the capital stack, starting from a 'gestation' phase for idea refinement.

Foundation Capital sees dangerous trend of 'lab companies' raising hundreds of millions at multibillion valuations without commercial traction; discipline required as some will fail to raise follow-on capital.

Trillion-dollar strategics paying 1% market cap for strategic AI assets creates exit comps that mathematically support billion-dollar seed valuations — but consumer AI agents (Instinct) may be overpriced vs B2B (Cursor).

Slipstream paused deployments for five quarters in 2021-22, arguing GPs must resist 'play the game on the field' pressure; co-invest conviction bar should be 'bet my LPs' next fund commitment on this,' and fund-of-funds should smooth vintage exposure rather than time markets.

Resilient M&A appetite from trillion-dollar buyers paying 1% market cap for strategic assets validates high early-stage valuations, with Cursor at $60B, Hugging Face at $15B, and OpenRouter at $8B cited as comps.

Masa Son taught Arora that operator instinct to fix struggling investments is wrong; venture returns come from concentrating on compounding winners that can quadruple. This power-law mindset applies to both investing and internal resource allocation.

Jason Lemkin declares any non-founder-led investment 'a zero' — founder-led companies (Revolut, Rippling, Owner.com) acquire competitors and attract talent; professional CEOs cannot navigate AI transition.

Bootstrapping forces companies into small markets that cannot support ambitious growth; founders with extreme ambition should target VC-backable markets large enough to justify venture capital and enable category leadership.

Systematically testing 20+ ideas by interviewing founders about pain points, validating willingness to pay within 72-hour cycles, allows teams to efficiently discover venture-scale opportunities like CPQ.

New AI-native firms are adopting a barbell model — raising SPVs for billion-dollar frontier model/robotics bets while maintaining a traditional early-stage fund for capital-efficient founders — reflecting the bifurcation of AI startup capital needs.

Nvidia's depreciation insurance and reference designs enable banks to underwrite GPU-backed debt as asset-backed securities, allowing pension funds and insurers to fund AI infrastructure at real-estate-like cost of capital instead of venture equity returns.

Regular investor updates — even during struggles — create trust and optionality for bridge rounds, while silence signals trouble and reduces follow-on probability; the 'only update when winning' approach fails for non-hot companies.

IVP's 46-year discipline of 10 investments/year is becoming more critical as AI winner-take-most dynamics mean 1-2 companies per fund generate essentially all returns, making selection and concentration paramount.

GV is shifting more investment partner time to San Francisco due to unprecedented deal concentration, while noting private valuations are rising exponentially driven by massive TAM expectations and a coming wave of AI IPOs.

Early-stage AI ventures like Discovery Loop command massive rounds at lofty valuations based on founder pedigree and TAM potential, forcing investors to recalibrate return expectations as $1B unicorns become commonplace.

Modern IPOs require massive revolving credit facilities as balance sheet signals; banks compete aggressively to join syndicates (Anthropic facility could exceed 10 banks) to secure roles in landmark offerings, making credit commitments a leading indicator of IPO pipeline quality.

Most AI application companies (neolabs) will cease as independent businesses within 18 months. Survivors own durable, proprietary workflows (e.g., legal) that won't be obsoleted by frontier model improvements and have high barriers to new entrants. General computer-use agents (Excel/Jira automation) are undifferentiated and vulnerable.

Sequoia data across decades shows best investments always from sponsors with highest conviction, not consensus. SpaceX had a 1/10 vote but sponsor forced partner visits. Airbnb was controversial (air mattresses). Conviction must persist post-investment (keep investing). Partnership encourages fierce debate (devil's advocate, premortems) but sponsor decides — politics removed by transparency.

Mayfield's 55-year track record of inception-stage investing (70% of deals at idea stage) with extreme selectivity (8-10 deals/year) has produced 120 IPOs and 225 acquisitions; the firm believes AI amplifies returns 2-3x and is deploying $3B with conviction that A+ founders in B/C markets will pivot to iconic outcomes.

European venture capital funds remain slow and fragmented, limiting their capacity to write large checks and compete with US investors for top startups, creating a structural headwind for the region's VC ecosystem.

Callosum's record UK seed round, backed by Atomic, Plural, DCVC, and UK Sovereign AI, signals a financial and sociological shift making London a competitive base for ambitious deep-tech AI companies.

After a slow year, large tech companies are restarting dealmaking (Nvidia, SoftBank, Salesforce) to acquire AI talent and technology, driven by capex budgets shifting to M&A and fear of missing strategic assets.

With seed entry valuations at $600M+ effective (post-dilution), even $9B exits only yield 15x — insufficient for fund returns. Requires either lower entry prices or $50B+ outcomes; changes seed strategy fundamentally.

Private equity firms are forming joint ventures with AI labs (Anthropic's Ode, OpenAI's deployment co) to drive AI adoption across their portfolio companies, creating a captive customer flywheel and potential equity upside through operational improvements.

Obrecht credits a AUD 1M commercialization grant and R&D tax concessions for doubling early runway, enabling Canva to stay in Sydney where talent is "incredible" and distraction low—arguing that non-dilutive government capital and secondary hub focus can be a strategic advantage for deep-tech founders outside Silicon Valley.

With $20T in unicorn-value companies now funded weekly, early-stage valuations have detached from traditional benchmarks; Samir's fund bets Recursive Self-Improvement (RSI) for scientific discovery — led by Jeff Dean's team — could match frontier model economics ($100B revenue/$2T cap) with private structure enabling more risk-taking than public labs.

Revolving credit facilities have become critical IPO signaling tools; banks compete to provide liquidity backstops that demonstrate financial resilience to public market investors, with Anthropic's facility potentially exceeding $10B.

Sequoia's 5-decade data shows best fund returns always from deals where sponsor had highest conviction, even if controversial (SpaceX vote of 1); process optimizes for conviction not agreement.

The Section 8 Clayton Act investigation into a16z for Ben Horowitz (Databricks board) and Martin Casado (FiveTran board) is a baseless PR attack likely from a rival CEO; private company boards have no material confidential info, and independent directors would resolve any theoretical conflict instantly.

London has undergone a financial and sociological shift: investors now back super-ambitious hard-tech ideas at scale (Callosum's $100M seed), and founders are unapologetically ambitious, capturing selective elements of US startup culture while leveraging deep local talent from universities like Imperial and Cambridge.

Anthropic's retroactive super-voting shares, OpenAI's 0% founder ownership, and long-term benefit trusts (Bernanke et al.) are governance experiments to insulate ASI decisions from quarterly pressure. However, capital flows to highest ROI; 'Moloch' (market forces) ultimately dictates outcomes regardless of legal structures.

Stripe ($7B OpenRouter) and SpaceX ($60B Cursor) demonstrate that well-capitalized private companies can issue stock for transformative acquisitions — bypassing public market scrutiny and timing — a structural shift in how late-stage value creation occurs.

A decade after 'unicorn' was coined for $1B valuations, $20T+ of such companies exist and are funded weekly; venture firms must recalibrate return expectations for pre-revenue AI startups raising hundreds of millions at lofty valuations.

Sequoia's historical analysis shows best returns always come from highest-conviction sponsors, not from optimizing ownership percentages. Partnership encourages fierce debate and 'front-stepping' to test conviction; sponsor retains final decision authority even against unanimous opposition.

The DOJ's century-old Clayton Act investigation into a16z partners sitting on Databricks and FiveTran boards is a baseless attack likely triggered by a jealous portfolio CEO; VC board interlocks are standard practice with firewalls, and a16z's SpaceX/Cursor returns validate the model.

Companies now raise at 2x valuations within 4 weeks (e.g., unnamed company 10B to 21B in a month) on a single strong month's data, versus 3-4 months previously. Jane Street participation signals hedge fund crossover into venture. This speed reflects extreme capital availability for AI leaders but raises diligence concerns.

Across decades of Sequoia funds, the best returns consistently come from deals where the sponsoring partner had highest conviction; increasing ownership percentages mathematically never beats conviction-driven selection.

Mayfield's 55-year track record (120 IPOs, 225 acquisitions) shows that investing at the paper-and-pencil stage in exceptional founders — not hot markets — drives outsized returns, with AI amplifying this edge 2-3x while the firm stays small and disciplined.

Current $100M+ seed rounds reflect hype cycle; only few companies can achieve hyperscale growth; investors must distinguish true platform companies from also-rans using discernment on founder commitment and market structure.

a16z's SpaceX position ($157B valuation) and Cursor exit demonstrate top-quartile returns; DOJ Section 8 investigation over interlocking directorates (Databricks/FiveTran boards) is trivial — private company boards hold no material confidential info — and likely instigated by a jealous competitor.

Thomas observes that Europe's fragmented markets generate unique problems—like multi-language dubbing or cross-border currency exchange—that force founders to build globally scalable solutions from day one, turning a structural weakness into a competitive advantage.

NSF and NIH are launching meta-science experiments including golden tickets for unilateral reviewer picks and grant durations from 6 months to 5 years, mimicking venture capital's tolerance for failure to reverse Eroom's law decline in research productivity.