Venture capital has always been a power-law business. In the AI cycle of 2023–2026, that power law became visible at unprecedented scale: a handful of foundation-model and application-layer companies absorbed the majority of private-market value creation, and the firms that backed them early—with meaningful ownership—pulled away from everyone else.

Earthian tracks this shift across three live rankings: the AI Venture Capital Ranking (/realtime-venture-capital-leaderboard), the Earthian Alchemist List 2026 (/earthian-alchemist-list-2026), and the AI Startup Leaderboard (/realtime-ai-startup-leaderboard). Read together, they tell a consistent story: concentrated bets beat diversified spray, especially when the underlying companies become category-defining platforms.

The line that separates winners from the rest

If you plot firm-level AI ecosystem value against portfolio breadth, a clear pattern emerges.

• Thrive Capital ranks #1 in Earthian’s AI VC ranking with **$864.9B** in direct ecosystem value—anchored by **OpenAI**, **Cursor**, and **Physical Intelligence**—despite only **three** top AI ventures and **three** early lead investments.
• Lightspeed Venture Partners ranks #2 with **$408.2B** in direct ecosystem value, driven primarily by **Anthropic**, plus **Mistral AI**, **Glean**, and **Safe Superintelligence**—again with just **four** top AI ventures and **two** early leads.
• By contrast, a16z and Sequoia Capital—both excellent firms—show seven to eight top AI ventures and seven to eight early lead investments, yet sit at $427.9B and $298.7B in direct ecosystem value respectively.

The gap is not about activity. It is about where capital was concentrated and when conviction was expressed.

On the startup side, the same concentration shows up in valuations: Anthropic leads Earthian’s AI Startup Leaderboard at $965B; OpenAI follows at $852B ($201B total funding). The firms that owned meaningful stakes in those two companies before consensus—Thrive on OpenAI, Lightspeed and a cluster of partners on Anthropic—captured disproportionate markups relative to dollars deployed.

Analogy: poker, not roulette

Think of recent AI venture investing as a high-stakes poker table—not a roulette wheel.

• Roulette strategy (spray-and-pray): place many small bets across dozens of AI startups, hoping one hits. You get activity, logos, and conference panels—but diluted ownership and weak signal when the market reprices.
• Poker strategy (concentrated conviction): identify the few hands worth pushing chips into, **lead** when others hesitate, and **stay in** as the pot grows.

Thrive and Lightspeed played poker. Thrive pushed early into OpenAI when frontier-model economics were still debated, then extended the same pattern to Cursor—applied AI with distribution that compounds. Lightspeed did the same on Anthropic, pairing foundation-model exposure with enterprise-grade applied AI (Glean) and international model bets (Mistral).

The Earthian Alchemist List makes this human: Joshua Kushner (Thrive, #1) and Ravi Mhatre (Lightspeed, #4) sit at the top not because they backed the most companies, but because they backed the right companies early, with partner-level conviction.

Strategy 1: Concentrated bets on platform winners

The dominant strategy in Earthian’s rankings is concentration on platform-scale outcomes—not maximum deal count.

Thrive’s ranking profile is the clearest example: one bet above $500B in ecosystem value (OpenAI), three early leads, and an ecosystem value score of 1.0000—the benchmark in our framework. That is not luck; it is a repeatable posture: fewer investments, higher ownership, higher bar for “this could define a category.”

Lightspeed’s profile rhymes with Thrive’s but on a different axis: Anthropic as the anchor, with a portfolio that connects foundation models to enterprise deployment and global model competition. Lightspeed’s $44B AUM and $7.0B Lightspeed XIV fund give it scale—but the ranking edge comes from where that capital landed, not how many logos appear on the website.

Strategy 2: Early lead investor before consensus

Earthian weights early lead investments heavily because leading rounds before product-market fit is obvious is how firms earn ownership quality.

The Anthropic coalition on the Alchemist List illustrates this:

• Yasmin Razavi (Spark Capital, #3)—early lead in Anthropic’s Series C
• Ravi Mhatre (Lightspeed, #4)—early lead in Anthropic
• Matt Murphy (Menlo Ventures, #5)—key partner-level role in Menlo’s early Anthropic investment

Spark Capital ranks #6 in the AI VC ranking with $371.2B in direct ecosystem value from essentially one top AI venture—Anthropic. Menlo ranks #5 with $356.6B from two top ventures. These are not diversified AI franchises; they are precision strikes on a generational company.

That is the second strategy: be the firm that writes the term sheet when the category is still controversial, not when every growth fund is competing for allocation.

Strategy 3: Ownership quality through the repricing cycle

Concentrated early bets only dominate if firms retain ownership through subsequent mega-rounds. Thrive’s repeated early involvement in OpenAI—and its application-layer adjacency via Cursor—reflects a partnership model built for multi-year compounding, not quick markups on small stakes.

Lightspeed’s Anthropic position similarly benefits from staying relevant across rounds as Anthropic moved from research lab to $965B enterprise AI platform. The AI Startup Leaderboard’s funding history (Anthropic: $128.7B total funding; OpenAI: $201B) shows how quickly paper marks become fund-defining outcomes for the investors who held through.

Firms optimized for breadth—a16z with seven early leads and seven top AI ventures, Sequoia with eight and eight—still produce strong returns and strategic optionality. But in Earthian’s ecosystem value per dollar of conviction framing, they rank below Thrive and Lightspeed because value accretion in this cycle has been top-heavy: a few companies captured most of the private-market AI valuation stack.

What this means for founders and LPs

For founders, the lesson is not “only talk to two firms.” It is: the partners who have dominated recent years are the ones who lead early, understand category risk, and stay committed when rounds get large. Kushner, Mhatre, Razavi, and Murphy appear on the Alchemist List because they exhibit that behavior repeatedly—not because they have the largest funds.

For LPs, the rankings suggest a portfolio construction insight: in power-law vintages, manager selection matters more than manager count. A LP allocated to Thrive and Lightspeed captured asymmetric exposure to OpenAI and Anthropic; a LP spread equally across ten “active AI” GPs may have gotten deal flow without getting ownership in the outcomes that mattered.

How to use Earthian’s rankings

• AI Venture Capital Ranking — firm-level ecosystem value, early leads, and concentration metrics
• Earthian Alchemist List 2026 — partner-level conviction mapped to top companies (OpenAI, Anthropic, Cursor, SpaceX, Wiz)
• AI Startup Leaderboard — real-time valuation and funding context for the companies driving VC marks

Bottom line

The most successful VCs in recent years did not win by being everywhere in AI. They won by drawing a straight line from early conviction to platform ownership: Thrive on OpenAI and applied AI winners; Lightspeed and the Anthropic coalition on the other side of the frontier-model divide. In a market where two companies alone exceed $1.8T in combined valuation on Earthian’s leaderboard, concentrated bets are not a stylistic preference—they are the strategy that the data rewards.