Venture Capital Investor Ethan Choi Makes Radical Pivot to AI

Ethan Choi doesn’t sugarcoat it. Entry-level jobs are dying. AI killed them.

The Khosla Ventures partner joined the firm in 2024 after a decade doing growth-stage deals at Accel and Spectrum Equity. Now he’s stage-agnostic. Backs founders at any round. Complete flip from his old playbook.

When I was at Greycroft, I saw plenty of investors stick to their lane. Series A guy stays Series A. Growth guy stays growth. Choi torched that script.

Why the pivot? AI changed everything.

“I’ve led investments in several AI-first companies,” Choi explained in a recent interview. “AI is a massive threat to entry-level jobs.”

He’s right. I’ve watched this play out across portfolio companies. Junior roles vanish first. Always do.

Choi built his reputation backing enterprise software and fintech infrastructure at scale. At Accel, he led growth investments in 1Password, Klaviyo, and Pismo—which Visa acquired. Also backed Nuvemshop and commercetools. Big checks. Late-stage stuff. Growth equity territory.

Before Accel, Choi worked at Spectrum Equity. Portfolio included Lynda.com, which LinkedIn bought. Also Headspace, Lucid Software, and PicMonkey (Shutterstock grabbed that one). Pattern recognition: back proven companies, write big checks, chase exits.

That was then.

Now Choi invests at any stage. Seed. Series A. Growth. Doesn’t matter. Focus shifted entirely to AI and founder-first conviction. That’s the new thesis.

What Makes This Venture Capital Investor Different

Most VCs won’t admit their entire strategy changed. Choi owns it.

He flipped his investing philosophy completely. Used to focus on growth-stage companies with proven metrics. Now? Stage-agnostic. Bets on founders building AI companies regardless of traction.

Follow the money. Incentives explain everything.

Khosla Ventures has been AI-focused for years. Vinod Khosla made that call early. Choi joined knowing the mandate. Growth equity doesn’t make sense when AI companies scale differently than traditional SaaS. You need to get in earlier. Smaller checks. Higher risk. Bigger potential multiples.

The math only works if you catch AI companies at formation. Not Series C.

Choi’s founder-first approach matters here. AI founders don’t follow typical playbooks. Revenue models break traditional SaaS metrics. Customer acquisition looks nothing like enterprise software. Conviction in founders becomes the only underwriting criteria that holds.

I’ve seen this movie before. When cloud happened, growth investors who stayed rigid missed the biggest winners. Same pattern now with AI.

The Entry-Level Job Crisis

Choi isn’t wrong about AI killing entry-level roles. The data backs him up.

VC-backed companies typically hired armies of junior employees. SDRs. Junior engineers. Analysts. Coordinators. All that dried up. AI tools replaced them.

When I was deploying capital at Bessemer, every Series A company planned to hire 50-100 people over 18 months. Now portfolio companies hire 10-15 senior people and use AI for everything else. Junior headcount vanished.

Choi calls this “a shifting social contract in the modern workforce.” He’s being polite. I’ll say it plainly: the career ladder just lost its bottom rungs.

Companies used to hire junior talent, train them, promote from within. That model dies when AI handles entry-level work better and cheaper than humans. No one’s training junior employees anymore because those jobs disappeared.

VCs won’t tell you this, but we created this problem. We pushed portfolio companies to “do more with less.” AI gave them the tool to cut headcount. Now we act surprised entry-level hiring collapsed.

What This Means for Founders

Choi’s pivot signals where smart capital is flowing. Away from traditional growth equity. Toward earlier-stage AI companies.

Founders should pay attention. This venture capital investor spent a decade writing $20M-$50M growth checks. Now he’s backing companies at formation. That shift matters.

Here’s what the term sheet doesn’t say: when a growth investor goes stage-agnostic, it means growth-stage economics broke. Returns dried up at later stages. All the value creation moved earlier.

I’ve sat in enough portfolio review meetings to know what happened. Growth funds deployed billions into 2021 vintage companies at 50x-100x revenue multiples. Those investments are underwater. DPI went to zero. LPs stopped getting distributions.

So growth investors like Choi pivoted earlier. Smaller checks. Lower entry multiples. Higher ownership. Better shot at returning the fund.

Valuation is vanity. Terms are sanity. Stage discipline is survival.

The Bigger Picture

Choi’s move from Accel to Khosla in 2024 wasn’t random. Timing matters.

Accel remains a phenomenal firm. Top-tier brand. Great returns historically. But their model centers on Series A and growth equity. Khosla goes earlier and swings harder on conviction bets. Different risk profile entirely.

For an investor pivoting to AI and founder-first deals, Khosla makes more sense than Accel. The fund structure supports it. LP base expects it. Partnership encourages it.

Fund economics drive everything. Here’s how: Accel raises $3B-$4B funds and needs to deploy $50M-$100M per company to move the needle. Can’t do that at seed stage. Khosla raises smaller, more flexible vehicles. Can write $5M seed checks or $50M growth checks from the same fund. Structure enables strategy.

Choi’s background gives him an edge few AI investors have. He’s seen scaled enterprise software and fintech companies up close. Knows what $100M ARR looks like. Understands unit economics at scale. That pattern recognition helps when underwriting AI companies that haven’t proven metrics yet.

Most seed investors never saw a company reach $100M ARR. Choi backed multiple. Different lens entirely.

Questions Ahead

The big question: does founder-first conviction work in AI, or does it lead to massive blow-ups?

AI companies burn capital faster than traditional software. Training models costs millions. Compute expenses never stop. Customer acquisition remains unproven for most AI-first products. Churn rates are unknown. Revenue durability is untested.

Betting on founders without proven metrics works great when the founder is right. Catastrophic when they’re wrong. AI hasn’t had enough time to separate signal from noise.

I’ve seen conviction-driven investing create 100x winners and total wipeouts. Nothing in between. AI will follow the same distribution.

Choi’s career performance—exits at Lynda.com, Pismo, PicMonkey—proves he can pick winners. Whether that translates from growth-stage software to early-stage AI remains unclear. Results take 7-10 years to prove out.

For now, one of venture capital’s top growth investors just went all-in on AI at every stage. That’s the signal. LPs are watching. So are founders. The pivot is complete.