Anthropic Files to Go Public | Cognition Raises $1BN at $26BN Valuation | The 996 Work Ethic

Anthropic's IPO and Market Impact

• Anthropic filed to go public following a $65 billion raise, positioning itself as a potential fastest IPO to scale.
• The deal is projected to value the company at $1 trillion within five years, comparable to SpaceX's trajectory.
• This valuation reset raises the "bar" for VCs and employees, potentially making $400M–$2B exits feel like "wasted time" compared to trillion-dollar outcomes.
• Investors express a preference for "billion-dollar positions" in their portfolios, dismissing smaller opportunities as insufficient for career excitement.
• The public listing is viewed as a necessary step to remove market "mystery" and allow the ecosystem to move forward, though it may increase psychological pressure on non-participants.

Capital Race and Public Market Dynamics

• A "grab it now" mentality has driven major tech firms (Anthropic, OpenAI, SpaceX, Google) to rush public market filings and capital raises.
• Google announced an $80 billion equity raise, shifting from a cash-flow machine to a heavy Capex consumer to fund AI data centers.
• The total potential equity issuance across these four entities (SpaceX, Google, Anthropic, OpenAI) could reach $300–$400 billion, all AI-related.
• The shift to equity raises is seen as a strategic move to secure capital at premium valuations before debt markets tighten, rather than relying solely on debt.
• The narrative of "staying private is cool" has ended, with the market favoring immediate liquidity and capital access.

SaaS "Saspocalypse" and Earnings

• The SaaS sector is experiencing its best earnings week in two years, signaling the end of the "Saspocalypse" panic.
• Cloud stocks (WorldCloud ETF) rebounded 25–30% in a month, returning to flat year-to-date performance after a 30% drop.
• The rebound validates that the previous over-correction was irrational, but fundamental concerns regarding seat contraction remain.
• Bifurcation in Software Value:
  • Companies with "agentic" or AI-fueled products (Twilio, Okta, Datadog) saw 57–100% annual gains.
  • Traditional "human-per-seat" software is facing declining demand as companies cut licenses to fund AI token spend.
  • Salesforce split its business into "Gen Trigger" (growing 12–13%) and legacy software (single-digit growth).
• Multiples for SaaS companies remain elevated (Atlassian at 4x ARR, HubSpot at 3.8x), but growth re-acceleration is the primary driver of value, not multiple expansion.

Cognition and Autonomous Agents

• Cognition raised $1 billion at a $26 billion valuation, with its core product Devon reaching $492 million in ARR.
• The market vision is shifting from "empowering mediocre engineers" to "autonomous AI engineers" (Devin) that execute tasks without human intervention.
• Jason (VC) views the vision of autonomous agents as more compelling than traditional "code generation" tools, arguing for the removal of middle-ground engineers.
• The deal highlights the intense competition for talent, with "hottest" startups now out-competing traditional VC portfolio companies for founders and CTOs.

Token Spend, Budgeting, and the "Tokens over Humans" Trend

• Corporate CFOs are moving from "crank all you want" to strict budgeting after Q1/Q2 token spend exceeded estimates by 10x.
• Uber announced a $1,500/month token cap for employees to manage costs, signaling a broader industry trend toward cost containment.
• Projected Shift: By end-of-year 2027, a significant portion of engineering budgets will be reallocated from headcount to tokens.
  • Rory estimates a potential 33% split ($100 in tokens for every $200 in human salary).
  • Jason predicts a more conservative 10% split initially, though this may rise.
• Impact on Roles:
  • QA and Customer Success departments face the highest risk of replacement by tokens/agents.
  • Best engineers are unlikely to be replaced; they will be given "unlimited tokens" to increase productivity.
  • Marginal roles (e.g., entry-level QA, CSM) will be cut to fund the remaining high-performance human and token mix.
• Rebuttal: Rory argues token costs are currently "noise" in customer support budgets, but Jason contends that in engineering, the trade-off is binary and quantifiable.
• Market Validation: The fact that companies are unwilling to cut token spend entirely validates a $500B–$1T market category, even if adoption slows due to budget constraints.

Legal Industry Disruption

• Kirkland & Ellis announced a $100 million/year investment over five years to build proprietary AI (competing with Harvey and Legora).
• This move is interpreted as a defensive tactic to protect "secret sauce" and data, rather than a fundamental threat to third-party AI providers.
• The consensus is that "full-stack AI law firms" will not replace high-end human judgment for multi-billion dollar transactions.
• Consumer vs. Enterprise: AI will drive massive market expansion for low-cost legal services (wills, divorce, small business) but cannot replace the "human in the loop" required for mission-critical, high-stakes decisions.
• Jason notes that high-stakes legal work requires human accountability for 100% accuracy, a standard AI cannot yet meet to the satisfaction of partners.

Financial Services and Wealth Management

• Robinhood's introduction of AI investment agents is viewed as a potential disruption to the "commoditization" of trading advice.
• Differentiation: AI can effectively automate asset allocation and financial planning (education/autonomy) but has not yet proven it can generate alpha through active trading.
• The primary value of AI in wealth management is creating "true experts" in users, providing tailored insights rather than just executing trades.
• Hedge funds (e.g., Citadel) will likely use LLMs for information synthesis, but replacing human traders with AI for alpha generation remains unproven.

Private Equity and Capital Markets

• Apollo and Harvard note that PE software returns are becoming "disastrous" due to overpayment and slowing growth in the SaaS sector.
• Valuation Squeeze: PE firms bought SaaS at 10x revenue during high-growth periods; with growth slowing to 8–9% and multiples contracting to 5–6x, equity returns are severely compressed.
• Debt struggles (high leverage in private credit) imply that equity positions below them are in even worse shape.
• Exit Challenges: Overpaid firms may face 10-year "grind" scenarios or "miserable" 1.2x–1.3x returns unless they can execute bolt-on acquisitions.
• VC Fund Retention: Massive distributions (e.g., Menlo's potential $10B+ from SpaceX/Anthropic) may cause key talent to quit, as the "next fund" math (smaller carry relative to life-changing wealth) fails to incentivize continuing work.
• Solution: Founders with massive carry might shift to LP roles to recapture full economic upside, or simply retire if the next deal's ROI doesn't justify the effort.

Work Culture and "996" Intensity

• The "996" work culture (9 AM to 9 PM, 6 days a week) is not new but remains a strategic choice for high-intensity, high-reward startups (e.g., Cognition, Corgi).
• Quid Pro Quo: Founders promising 8-figure exits to early employees justify the intensity; failure to deliver returns makes such expectations "toxic."
• Sustainability: The "marathon" of startup work is unsustainable for 24/7 intensity over a career, but essential for the first 50 employees of a high-growth firm.
• Judgment Risk: Extreme work intensity can degrade judgment, requiring leaders to balance effort with psychological health to avoid "rage working."
• Hiring Contradiction: While the valley predicts mass unemployment via automation, companies are simultaneously struggling to hire top talent, suggesting AI adoption is accelerating human productivity rather than replacing it en masse.

Macro Economic Outlook

• Contrary to "AI doom" narratives regarding GDP, the long-term growth rate (historically ~2%) is expected to hold as productivity gains are offset by "weak links" in the value chain (productization, sales, packaging).
• The "SaaS apocalypse" fear was an over-correction; the sector is stabilizing with a focus on AI-integrated growth.
• The ecosystem is shifting from "human-centric" capital allocation to "token-centric" allocation, fundamentally altering how engineering and support budgets are structured.