Chaos detected. Sequoia Capital just deployed $2 billion into AI startups at 50x revenue multiples. Analysis loading.
That’s not a typo. Fifty times revenue. For companies that haven’t turned a profit. In a market where interest rates are still hovering above 4%. The old model is dead.
Sequoia’s aggressive AI push under new leadership—Roelof Botha and Alfred Lin, with Grady Burnett driving AI deals—is reshaping venture capital norms. But the question isn’t whether AI is the next big thing. It’s whether Sequoia is repeating the same mistakes that blew up crypto in 2021.
I’ve seen this playbook before. In 2017, I was a 21-year-old economics student in Taipei, neglecting my thesis to monitor EOS IEO rounds across multiple exchange platforms. I tracked volatile token distribution mechanics in real-time, correlating whale wallet movements with price spikes during the final bidding phase. The frenzy was fueled by FOMO, not fundamentals. Sound familiar?
Sequoia’s current AI strategy mirrors that era. The same narrative-driven valuation inflation. The same rush to deploy capital before a competitor does. The same disregard for sustainability. Let’s autopsy the deal.
Context: The New Sequoia Playbook
Sequoia Capital, the legendary VC firm behind Apple, Google, and Stripe, has a new sheriff in town. Since 2022, the firm has pivoted hard into AI, led by partners like Alfred Lin and Grady Burnett. Their thesis: AI is the next platform shift, and getting in early justifies any price.
Traditional VC norms are dead. Staged funding rounds with disciplined valuation caps? Gone. Instead, Sequoia is writing massive checks at Series A—$100 million or more—for startups with little more than a demo and a founder with a PhD from Stanford. The justification: AI talent is scarce, and the window to capture market share is narrow.
But the numbers don’t lie. According to PitchBook, AI startup valuations in Sequoia’s portfolio have surged 300% since 2023, while revenue growth has lagged at 40%. That’s a disconnect. In crypto, we call that a bubble.
I remember the 2020 DeFi Summer. I spent weeks analyzing Compound and Uniswap interactions, identifying inefficiencies in cross-protocol arbitrage opportunities. I published a series of threads dissecting how flash loans could be used to manipulate oracle prices. The same pattern is playing out in AI: hype masking structural flaws.
Core: The Mechanics of Overvaluation
Let’s dig into the data. Sequoia’s AI portfolio includes companies like Cohere, Anthropic, and a slew of generative AI startups. Their average burn rate is $50 million per year, primarily on GPU compute. That’s unsustainable. Compare that to crypto’s Layer 2 rollups, which bleed cash on proving costs. ZK rollups, for instance, spend millions monthly on Ethereum gas just to submit proofs. Unless gas returns to bull-market levels, operators are bleeding money. Same story here.
Sequoia’s aggressive stance is forcing other VCs to follow. Andreessen Horowitz, Lightspeed, and Tiger Global are all chasing AI deals with similar terms. This is a classic herd mentality. In 2021, the same herd drove crypto valuations to $3 trillion. Then Terra collapsed. Then FTX. Then the market dropped 80%.
Based on my experience auditing on-chain data during the 2022 Terra/LUNA collapse, I can spot the warning signs. I mapped the liquidation cascades hour-by-hour, providing a causal chain many mainstream outlets missed. The root cause was not technology failure—it was governance failure. Over-leveraged positions, lack of risk management, and a narrative that replaced fundamentals.
Sequoia’s AI bets are showing similar symptoms. Their portfolio companies are burning cash to acquire users, not to build moats. The moat is supposed to be the AI model itself, but models are becoming commoditized. Open-source models like Llama 3 are catching up to proprietary ones. The differentiation is shrinking.
Deconstruction complete.
Let me bring in another experience. In 2024, during the Spot Bitcoin ETF debate, I used my economics background to predict the voting patterns of specific SEC commissioners based on their past regulatory filings. I broke the news of the sudden shift in SEC stance 48 hours before major outlets. The lesson: surface-level analysis misses the underlying dynamics. The same applies here. Everyone is looking at AI’s TAM (total addressable market) but ignoring the structural fragility.
Sequoia’s AI investments are not just about technology. They are about signaling. By deploying capital aggressively, Sequoia signals to LPs that they are ahead of the curve. But LPs are ultimately buying non-dividend stock—same as DAO governance tokens. The only hope is that later buyers will take the bag. That’s not fundamentally different from a Ponzi.
Contrarian: The Unreported Angle
Here’s the counter-intuitive insight: Sequoia’s aggressive AI push might actually be a hedge against crypto disruption. As AI agents start using blockchain for autonomous payments, Sequoia is positioning to capture both the AI and the crypto narratives. But the twist is that Sequoia’s own portfolio companies could be disrupted by decentralized alternatives.
Consider the AI-agent economy. In 2026, I pivoted to covering the intersection of AI agents and blockchain, focusing on decentralized compute markets like Render and Akash. I identified early patterns where AI agents were autonomously spending crypto on data feeds, creating new on-chain revenue streams. My ENTP curiosity led me to hack together a simple demo of an AI agent executing a trade. It went viral.
The paradigm shift is that AI and crypto are merging into a single autonomous economy. Sequoia’s centralized AI bets could be outdated if decentralized compute networks become more efficient. The same way centralized exchanges were disrupted by DeFi.
Another blind spot: Sequoia’s AI investments are concentrated in the US. But the next wave of AI innovation is happening in Asia, particularly in Taiwan and China. I’ve been on the ground in Taipei since 2017, watching the ecosystem evolve. The chip supply chain is here. The talent is here. Sequoia is missing the geographic diversification.
Takeaway: The Next Watch
So what’s the takeaway? Sequoia’s aggressive AI investments are reshaping VC norms, but not necessarily for the better. The sustained high valuations and increased competition are creating a bubble. The question is when it pops.
I’m not predicting a crash tomorrow. But the data suggests that the current AI valuation cycle is unsustainable. Sequoia may be the last to exit, but that doesn’t make them immune. Just ask the investors who bought EOS at $23 in 2018.
EOS didn’t die; it evolved. Do you?
Prediction synthesis: watch for the first major AI startup failure. When it happens, the domino effect will be swift. Sequoia’s portfolio will be hit hardest because they are the most exposed. The smart money is already rotating into decentralized AI infrastructure. I’ll be covering that.
Until then, keep your eyes on the burn rate. And remember: in a bear market, survival matters more than gains. Use data to judge which protocols are bleeding. The same applies to VC-backed AI startups.
Scarlett Anderson, signing off.
