Jun 26, 2025, 10:13 AM ETFireside Chat, Interview, Conference Presentation, Webinar

Navan Files to Go Public and Canva Pulls the Brakes: Why and What Happens?

Meta's acquisition of Scale AI (and potentially Nat Friedman/Daniel Gross teams) is driven by the fear that ChatGPT could become the primary internet interface, siphoning user attention and revenue minutes away from Meta's ecosystem.
The acquisition is not a direct revenue play against Anthropic but an existential insurance policy costing roughly $100 billion (8% of Meta's $1.8T market cap) to prevent deplatforming or irrelevance.
The "magic moment" for AI development is concentrated: only those present in the early OpenAI environment (e.g., OpenAI, Anthropic, xAI, Inflection) possess the proprietary knowledge to command billion-dollar valuations.
California's non-compete laws facilitate this "knowledge leeching," allowing talent to leave and found new competitors, whereas other jurisdictions might legally lock this talent in place.
Sam Altman confirmed that Meta has made multiple offers totaling hundreds of millions of dollars to OpenAI talent, creating a hyper-transactional market where loyalty is secondary to immediate financial gain.
The "magic" of AI is facing rapid commoditization; while the initial innovation is exclusive, the ability to replicate it is becoming cheaper, shifting value from pure invention to execution and market capture.

Harvey.ai raised $300 million at a $5 billion valuation despite product limitations, relying on early "mindshare" creation rather than immediate product superiority.
Harvey's strategy involved establishing itself as the "lawyer's choice" early, signing high-profile clients (e.g., Allen & Overy) to create a stampede effect, effectively "freezing" the market before the technology fully matured.
For a $5 billion valuation to be justified, the business model must evolve beyond selling software licenses to "eating the work" of lawyers, effectively replacing human labor budgets rather than just augmenting them.
The Total Addressable Market (TAM) for legal AI is significantly larger if the tool replaces human labor (charging lawyer rates) rather than acting as a research adjunct (charging software rates), as labor spend is 5x that of data/tools.
Competitors like Harvey, Legora, and Crosby are fragmenting the legal market, unbundling specific processes (patents, NDAs, corporate law) which may actually reduce the overall TAM compared to broad generalist predictions.
Legacy law firms face a structural threat: they risk losing business to agile, AI-native firms (like Crosby) that can offer near-instant, low-cost analysis of standard documents, forcing incumbents to either adapt or lose market share.

Navan's recent IPO filing signals a broader trend where 62.5% more IPOs are happening this year, driven by a "Pavlovian" response to recent massive gains like Circle's 5x IPO debut.
Circle's post-IPO surge to $231 (5x the IPO price) was driven by speculative trading momentum rather than fundamental changes, highlighting the current "bubble" behavior in public markets.
The decision to go public is increasingly dictated by price: if the public market offers a "stupid price" higher than the private market, companies will go public despite the operational friction.
Canva's delay in IPOing is rationalized by its strong cash flow ($3B+ ARR) and low capital needs; the founders may prefer to buy back shares and pay dividends rather than endure public market scrutiny.
Larry Ellison's strategy at Oracle demonstrates a shift from using cash flow for buybacks (increasing ownership) to heavy CapEx investment in AI ($30B+), prioritizing long-term positioning over short-term share price maintenance.
The "price lever" dominates economic decisions: if private valuations stagnate or drop relative to public liquidity, the pressure to IPO increases even for profitable, non-capital-hungry companies.

A prevailing trend in AI is "Leverage Beta": companies winning are not necessarily building superior products initially but are claiming market territory through aggressive marketing while the technology matures.
Companies like Harvey and Lovable are "arbitraging the obvious" by claiming market leadership now, knowing the underlying LLM tech will inevitably improve to meet the hype, forcing late movers to scramble.
The "implementation will be boring" reality: many successful AI wrappers (e.g., Lovable) are essentially basic integrations of existing models with "pretty buttons," yet they succeed by capturing user attention first.
Sales tools for AI lag behind developer tools; the market lacks consumer-grade, real-time AI assistance for sales reps to overcome their own knowledge gaps, creating an opportunity for tools like Clue.ly.
Marketers are increasingly using shock value and controversial tactics (e.g., social media stunts) to cut through noise, betting that the "first mover" perception outweighs reputational risk in the short term.

Major B2B players (Salesforce, Slack, LinkedIn) are locking down APIs and data access in response to the existential threat of Model Context Protocol (MCP), which allows external AI agents to access proprietary data.
The trend of "locking the wagons" is a sign of a decaying empire; monopolies are moving from open integration to fee-based access or strict restrictions to protect revenue streams.
HubSpot's early embrace of MCP integration is highlighted as a strategic win, viewing the threat as an opportunity to offer superior AI-native solutions rather than resisting them.
Slack's restriction of AI access to its data is seen as a dangerous signal to customers, potentially driving adoption of open alternatives or forcing a shift to priced API models for data access.
The "efficiency grind" will eliminate mediocre knowledge workers; AI can already handle complex tasks (like reviewing LP agreements or commercial contracts) with near-perfect accuracy, rendering slow, human-dependent processes obsolete.

OpenAI vs. Microsoft Antitrust: Bet taken on "Yes" (36% odds); Sam Altman has likely already made threats sufficient to count as accusations, making a formal lawsuit a logical next step to pressure Microsoft.
US Gov Takeover of AI: Bet taken on "No" (68% odds); the current administration and tech leadership (Sacks, Andreessen) favor rapid AI development over government seizure, with no active hints of such a move.
Vanguard Divestment Impact: Divesting assets prior to joining public service likely cost money (due to the crypto/late-stage market dip), representing a financial sacrifice rather than a strategic genius play.
Trump Smartphone Release: Bet taken on "No" before September; lack of supply chain leaks or production evidence suggests the project is either non-existent or a mere marketing sticker on existing hardware.