Thomas Lefant, partner at COTU managing $70 billion ($30 billion private side), discusses leading Anthropic's recent funding round and the firm's crossover investment strategy. COTU was founded in December 1999 and survived the 80% market decline in their first 2.5 years.

The conversation covers Claude's explosive growth, the AI adoption surge across portfolio companies, and the structural changes facing SaaS valuations. Lefant shares insights on the private-to-public market transition, engineering productivity debates, and COTU's 'big idea' investment philosophy that led to early Nvidia conviction.

Key topics include Anthropic's rapid scaling, the democratization of private market access, SaaS multiple compression, and whether AI will create or destroy software engineering jobs. Lefant also discusses the firm's risk management approach and his collaboration with brother Philippe.

Anthropic's Explosive Growth and Claude Code Adoption

Anthropic's revenue projections grew materially during the 2-3 month fundraising process, with the company publicly disclosing revenue in excess of $13-14 billion

Claude Code's inventor Boris previously worked at COTU for 2.5 years developing software, making the investment particularly meaningful for the firm

Board meetings across portfolio companies consistently show AI tool adoption reports, with one company planning to triple their AI spending next year

Companies want to ensure they're using the best tools and being 'AI forward' to avoid being outcompeted

"We're spending X on this tool and we think it's way too low. We want it to be much bigger. We expect the spend to at least triple" - Thomas, quoting a board meeting slide

The Next Generation of Public Market Leaders

The current Mag 7 has been essentially flat over the past year, with Microsoft losing almost a trillion dollars in value due to AI positioning concerns

The next Mag 7 will likely include names like SpaceX, OpenAI, Anthropic, Revolut, and DataBricks - companies that may go public in the next 12-24 months

"If you're going to want to outperform the index over a long period of time, you're going to need exposure to these companies" - Thomas

The innovation coming from late-stage private companies is "unbelievable" and represents a critical asset class for future returns

SaaS Valuation Crisis and AI Disruption Threat

SaaS companies face decelerated growth (Workday now growing 13% vs. historical mid-20s to low-30s) while still trading at expensive multiples of 28-30x earnings

Investors can now buy semiconductor companies like Broadcom growing almost 40% at cheaper multiples than slower-growing SaaS companies

Companies must either reaccelerate growth through AI adoption or see multiples rerate down to ~20x gap earnings to match market comparables

"In three or four years, if cloud code can rewrite their entire business, that's harder for companies to control" - Thomas on terminal value concerns

AI's Impact on Software Engineering Jobs

None of COTU's portfolio companies are saying they want to cut engineering staff in half due to AI efficiency gains

Companies hope engineers become significantly more productive to enable features that were never possible before, potentially shifting from selling software to selling work

Example: Rippling moving from HR software to actually providing HR services through AI automation

The bank teller analogy: ATMs introduced in the 1970s led to an explosion in bank teller jobs through 2000s, not the predicted 70% reduction

ATMs reduced branch costs, enabling more branches, which increased total employment despite fewer employees per branch

There are approximately 400,000 software engineers in the Bay Area, with the impact of AI on this number still uncertain

COTU's Big Idea Investment Philosophy

Thomas focuses on TAM growth over time and additional TAMs rather than initial market size, calling initial TAM size "irrelevant"

The firm uses "BFI" (Big F***ing Idea) internally because big ideas should be "jolting" and memorable

Uber example: The initial taxi TAM wasn't compelling, but it grew 5-10x through reduced friction and added grocery/food delivery TAMs

Apple/iPhone analysis in early 2000s wrongly predicted 5% annual price declines, when prices actually increased as the TAM expanded massively

Thomas was the Apple analyst from 2003 for almost 20 years, witnessing the iPhone's TAM expansion firsthand

Risk Management and Market Volatility Perspective

Thomas read the Century Paper but disagrees with comparisons to "screaming fire in a crowded room," viewing early AI risk discussions as healthy preparation

"If everyone thinks we're in a bubble, then we're not in a bubble" - Thomas on the importance of ongoing market questioning

Daily volatility and questioning is preferable to no volatility followed by a massive crash 3 years later

COTU's formative experience of 80% market decline in first 2.5 years created a "face tattoo" focus on risk management that has kept them in business for three decades