RVI gives you OpenAI, Anthropic, and xAI at a 30%+ NAV premium and a ~2.5% expense ratio. ARKK gives you Tesla, Coinbase, and Roku at NAV and a 0.75% expense ratio. That is the short answer. The longer answer is more interesting.
Both products show up in retail brokerage accounts under the same "disruptive innovation" label, but the structural differences — closed-end fund vs ETF, private vs public, 18 vs 35 holdings, ~2.5% vs 0.75% in fees — change the answer completely depending on what you actually want exposure to. I have run this comparison for friends, founders, and LPs five separate times in the last 60 days, and the right pick is rarely what someone walks in assuming.
RVI vs ARKK Comparison: Which Tech Fund Wins in 2026?
The RVI vs ARKK comparison is not apples-to-apples — RVI is a closed-end fund holding 18 private AI and tech companies (OpenAI 12%, Anthropic 10%, xAI 9%, Stripe, Databricks), while ARKK is an open-end ETF holding 30–35 public high-growth stocks (Tesla 10%, Coinbase, Roku). RVI carries a 30%+ NAV premium and ~2.5% expense ratio. ARKK trades at NAV with 0.75% in fees and roughly $5–6B in AUM.
| Attribute | RVI (Robinhood Ventures I) | ARKK (ARK Innovation ETF) |
|---|---|---|
| Structure | Closed-end fund (CEF) | Open-end active ETF |
| Launch | November 2025 | October 2014 |
| AUM (mid-2026) | ~$300M NAV | ~$5.5B |
| Holdings count | 18 private companies | 30–35 public stocks |
| Exposure type | Private, illiquid | Public, daily liquid |
| Top holding | OpenAI ~12% | Tesla ~10% |
| Expense ratio | ~2.5% | 0.75% |
| NAV premium | +30% to +100% | ~0% (trades at NAV) |
| Tax structure | RIC, passes through distributions | 1940 Act ETF, low cap-gains distributions |
| 1-year return (mid-2026) | Trading premium, not NAV change | +22% from June 2025 trough |
| 5-year return | N/A (launched late 2025) | −45% cumulative |
| Daily trading volume | ~500K shares | ~10M+ shares |
The winner depends entirely on what you are trying to own. If the goal is direct private AI exposure — OpenAI, Anthropic, xAI specifically — RVI is the only retail-accessible listed wrapper that gets you there. If the goal is liquid, low-cost disruptive innovation in public markets, ARKK is cheaper, more transparent, and trades at fair value. Owning both makes sense for some investors. Owning RVI as a Cathie Wood ARKK replacement does not.
RVI Holdings: Private AI Bets You Cannot Get Anywhere Else
RVI's 18 disclosed positions as of June 2026 form one of the only listed pure-play portfolios of private AI infrastructure. Track the live breakdown on the Robinhood RVI Fund dashboard. The concentration matches where late-stage venture money has flowed since 2023:
| Rank | Holding | Approx. Weight | Sector |
|---|---|---|---|
| 1 | OpenAI | ~12% | Foundation model |
| 2 | Anthropic | ~10% | Foundation model |
| 3 | xAI | ~9% | Foundation model |
| 4 | Stripe | ~7% | Fintech infrastructure |
| 5 | Databricks | ~6% | Data infrastructure |
| 6 | Perplexity | ~5% | AI search |
| 7 | Mercor | ~4% | AI talent platform |
| 8 | Ramp | ~4% | Spend management |
The top three foundation-model names — OpenAI, Anthropic, xAI — account for over 30% of the fund. That is the entire RVI thesis: this is a venture-style concentrated bet on who wins the LLM race, wrapped in a listed share class retail can actually buy. SpaceX is conspicuously absent because SpaceX excludes retail-facing pooled vehicles from its tender offers. If you specifically want SpaceX exposure you need DXYZ, ARK Venture Fund (ARKVX), or secondary platforms like Forge and Hiive.
ARKK Holdings: Public High-Growth Tech, Heavy on Tesla
ARKK's 30–35 holdings are required by ETF rules to be liquid public securities. The portfolio is actively managed by Cathie Wood's ARK Invest team and reshuffled daily. Mid-2026 weights cluster around a handful of high-conviction names:
| Rank | Holding | Approx. Weight | Sector |
|---|---|---|---|
| 1 | Tesla (TSLA) | ~10% | EV / autonomy / AI |
| 2 | Coinbase (COIN) | ~9% | Crypto exchange |
| 3 | Roku (ROKU) | ~7% | Streaming platform |
| 4 | Roblox (RBLX) | ~5% | Gaming / metaverse |
| 5 | Block (SQ) | ~5% | Fintech |
| 6 | Palantir (PLTR) | ~5% | AI / defense data |
| 7 | Robinhood (HOOD) | ~4% | Retail brokerage |
| 8 | CRISPR Therapeutics | ~4% | Gene editing |
ARKK's only AI-pure-play exposure runs through Tesla (autonomy + Optimus + Dojo), Palantir (enterprise AI/data), and a few smaller positions in companies like Tempus AI and Recursion Pharmaceuticals. If you want the actual foundation-model layer, ARKK does not own it — there is no OpenAI ticker to buy. That is the structural gap RVI fills.
Expense Ratio, NAV Premium, and the Total Cost of RVI vs ARKK
Stated expense ratios do not capture the real cost difference. Closed-end fund mechanics mean RVI can trade at large premiums or discounts to its underlying net asset value, while ETF authorized-participant arbitrage keeps ARKK pinned to NAV.
| Cost Component | RVI | ARKK |
|---|---|---|
| Stated expense ratio (annual) | ~2.50% | 0.75% |
| NAV premium at entry (mid-2026) | +30% to +100% | ~0% |
| Embedded private-fund fees on underlying SPVs | Variable — included in NAV | None |
| Bid-ask spread (typical) | ~0.40% | ~0.02% |
| 5-year all-in cost on $10K | $1,250+ on fees alone (excluding premium loss) | ~$375 |
The 30%+ NAV premium is the single biggest hidden cost in RVI. If you buy $10,000 of RVI at a 35% premium and the premium contracts to historical CEF norms of 5–15% over 24 months, you lose 15–25 percentage points of return before the underlying portfolio moves at all. That is the math closed-end fund analysts have been pricing for decades — and it has burned every retail wave that piled into a hot CEF at a high premium (see GBTC's 2017 unwind, MFD's 2021 unwind, and arguably DXYZ's 2024–25 round trip).
Performance: ARKK's Round Trip and Why RVI Has No Track Record
Performance comparison is asymmetric because RVI launched in late 2025 and most of its 7-month return is premium expansion rather than underlying NAV change. ARKK has 11.5 years of public data — and it is sobering.
| Period | ARKK | S&P 500 (SPY) | Nasdaq 100 (QQQ) |
|---|---|---|---|
| Inception (Oct 2014) to mid-2026 | +165% | +285% | +440% |
| 5-year (Jun 2021 – Jun 2026) | −45% | +65% | +75% |
| 3-year (Jun 2023 – Jun 2026) | +40% | +50% | +72% |
| 1-year (Jun 2025 – Jun 2026) | +22% | +18% | +24% |
| Peak-to-trough drawdown (Feb 2021 – Dec 2022) | −81% | −25% | −35% |
ARKK's 5-year cumulative return is roughly 110 percentage points behind the S&P 500. The 81% peak-to-trough drawdown from February 2021 to December 2022 — the worst of any major themed ETF that survived — is what most retail investors remember when they think about "Cathie Wood." RVI does not have this kind of track record to evaluate yet. What we can model is that if the 30%+ premium collapses to a normal 5–10% premium and the underlying NAV grows 25–35% annually (rough late-stage venture base rate), 5-year RVI returns land somewhere around +60–90% net of fees — well above ARKK's 5-year trailing but not by enough to ignore the premium risk.
Tax Treatment: How RVI and ARKK Differ at the IRS
Both vehicles are Regulated Investment Companies (RICs) under Subchapter M of the IRC, but the structural plumbing differs in ways that matter for a taxable account.
RVI tax behavior
- • Distributes realized gains, dividends, and interest each year
- • Most early-year distributions will be ordinary income, not qualified
- • Premium does not affect your cost basis — only the share price you paid
- • Long-term capital gains apply on sale after 12 months held
- • No K-1 — issues a standard 1099-DIV
ARKK tax behavior
- • In-kind creation/redemption minimizes capital gains distributions
- • Most years see near-zero year-end capital gain distributions
- • Dividends from underlying stocks pass through as qualified or ordinary
- • Tax bill is concentrated at sale, not annually
- • Standard 1099-DIV and 1099-B reporting
For taxable accounts, ARKK's structure is meaningfully more efficient. For tax-deferred accounts (IRA, 401k, Roth), the difference disappears. If you are buying RVI in a taxable account, expect annual 1099 paperwork and some ordinary-income drag every year.
RVI vs ARKK Decision Framework: Which Should You Actually Buy?
Buy RVI if:
- ✓ You specifically want OpenAI/Anthropic/xAI exposure
- ✓ You are sizing 1–5% of your tech allocation as a venture-style bet
- ✓ You can hold 5+ years and tolerate the premium round-tripping
- ✓ You are buying in a tax-deferred account
- ✓ You are willing to wait for the premium to compress before adding
Buy ARKK if:
- ✓ You want liquid, daily-priced disruptive innovation exposure
- ✓ You are comfortable with a Tesla-heavy public-equity portfolio
- ✓ You need the option to sell quickly without a discount cost
- ✓ You are tax-sensitive and holding in a taxable account
- ✓ You believe Cathie Wood's stock picks will revert to her 2017–2020 form
Owning both in modest size — say 2% RVI and 3% ARKK as part of a 20–25% high-growth-tech sleeve — covers private and public exposure to disruptive technology in a way neither vehicle does alone. Compare the dashboards on AI Valuations and SaaS Valuations before sizing — multiples in both private and public AI are at extended levels relative to 5-year averages.
RVI is the only listed way for retail investors to own OpenAI and Anthropic. ARKK is the cheaper, more liquid, more tax-efficient way to own public high-growth tech.
If you want private AI specifically, RVI wins. For everything else, ARKK still wins on cost.
Track live RVI holdings, NAV premium, and ARKK exposure on the Robinhood RVI Fund dashboard at Value Add VC. Originally published in the Trace Cohen newsletter.