AI EcosystemsSaaStr

20VC x SaaStr: Anthropic at $900B, Salesforce Spends $300M on Tokens, SpaceX Files for the Biggest IPO in History, and Why the Tech Lash Is Just Beginning

Read original
vendor-fundingai-policymarket-consolidationai-coding-tools

“Salesforce will spend $300M on Anthropic tokens this year, almost entirely on coding. That works out to $15,000-$20,000 per engineer per year. Most public company CEOs have no idea where they sit on that curve.”

Key takeaways

  • Anthropic at $900B/18x revenue is cheaper than typical Series A/B rounds at 20-50x ARR, making it potentially the best venture trade available despite the absolute valuation
  • Salesforce's $300M token spend ($15-20K per engineer) reveals enterprise AI economics are becoming material line items, but most CEOs lack visibility into their positioning on the cost curve
  • Massive tech layoffs (38,875 jobs across Cisco, LinkedIn, Meta, Intuit) coinciding with AI boom signals beginning of 'tech lash' - public sentiment shifting from ChatGPT applause to booing Eric Schmidt at graduations
  • Anthropic's fundraising philosophy (trade at discount, close fast, get full check) contrasts sharply with OpenAI's approach (maximize valuation, contingent capital, borrowed funds) - reflects different capital strategy philosophies
  • Compute infrastructure war is real: 5-6 gigawatts at $40-50B per gigawatt means persuading hyperscalers to spend hundreds of billions on your behalf, not direct capital deployment

Why this matters for operators: Enterprise leaders budgeting AI spend need benchmarks; VCs evaluating AI infrastructure plays need valuation context

I cover AI×GTM intelligence like this every Wednesday.

Get STEEPWORKS Weekly

More picks

Personal Productivity & AI-Augmented Workr/ClaudeAI

🚀 Skills for small businesses, officially released by Anthropic

  • Anthropic released 31 small-business AI skills that hit 382K downloads on day one, signaling massive demand for packaged AI workflows
  • New category emerging: 'AI business operating templates' - standardized, AI-readable skill files that replace manual automation stitching across multiple tools
  • Skills are vendor-agnostic at core (.md files) - can be adapted for Codex, Cursor, Gemini, or other AI agents, not locked to Claude
ai-business-templatesautomation-stacksai-agent-orchestration
Read original Full analysis →
Enterprise AIExponential View

👀 The AI backlash is the only thing growing faster than AI revenues

  • Anthropic achieving extraordinary growth ($10.9bn Q2 revenue, profitable 2 years early) while public backlash intensifies
  • Cultural resistance emerging at grassroots level - college students booing AI proponents at commencements, demanding 'human' solutions
  • Infrastructure impacts creating local opposition - data centers affecting water quality, environmental concerns becoming political flashpoints
ai-policyregulatory-impactmarket-consolidation
Read original Full analysis →
Enterprise AIAxiosVictor's pick

AI backlash becomes a real business risk

As we go all in on AI, the meta of whether its good, how its positives wlil be distributed, and just how far society is from optimizing for it is so real

  • AI sentiment has collapsed across all demographics: only 18% of Gen Z feels hopeful, 70%+ of Americans think it's moving too fast, and negative views doubled from 34% to 50% in three years
  • AI executives are dangerously disconnected from public sentiment - Superhuman CEO was literally unaware of negative polling, seeing AI adoption as 'inevitable as the internet'
  • Backlash is creating real business consequences: record data center cancellations in Q1 2026 due to community resistance, threatening the compute infrastructure AI companies depend on
ai-policyregulatory-impactmarket-consolidation
Read original Full analysis →

This analysis was produced using the STEEPWORKS system — the same agents, skills, and knowledge architecture available in the GrowthOS package.