Top IPOs to Watch — Vol. 1: SpaceX, Anthropic, OpenAI, Databricks, Stripe, Cursor, Polymarket, Crusoe, Kraken — May 2026

This is Vol. 1 of our weekly pre-IPO tracker. The job is simple: tell you what changed this week across the names retail investors actually care about, give you the valuation and timeline, and tell you what to do about it.

Pre-IPO investing for retail used to be theoretically possible and practically locked. That has shifted in the last twelve months. Closed-end funds like DXYZ trade on the NYSE. ARK’s Venture Fund (ARKVX) takes $500 minimums. The SpaceX IPO will reserve up to 30% for retail. The walls are coming down — but the noise is overwhelming. This series is the filter.

Disclosure: I do not currently hold direct private positions in any company below. Aedilis has no affiliate relationships with the brokerages or funds named as of publication. This is analysis, not investment advice.

The Roster — Week of May 11, 2026

Company	Valuation	IPO Status	Heat
SpaceX / Starlink	$1.75T–$2T	Roadshow June 8; trade ~June 20–30	🔥🔥🔥
Anthropic	~$900B (closing)	Bank talks; possible Q4 2026 IPO	🔥🔥🔥
OpenAI	$852B	Filing H2 2026; listing 2027	🔥🔥🔥
Databricks	$134B	S-1 expected Q3 2026	🔥🔥
Stripe	$159B	No IPO planned; tender instead	🔥
Cursor (Anysphere)	~$50B	xAI buyout option; IPO ~late 2027	🔥🔥
Polymarket	$15B	Possible NYSE direct listing Q3 2027	🔥🔥
Crusoe	$13B	Pre-IPO round live; ex-MongoDB CFO hired	🔥🔥
Kraken	$13.3B (down from $20B)	S-1 confidential; paused	🔥

1. SpaceX / Starlink — The Headline IPO

What changed this week: The Motley Fool reported May 5 that the public S-1 is roughly two weeks out, putting the filing window at May 15–22. The roadshow targets the week of June 8 and shares likely begin trading on Nasdaq between June 18 and June 30. Up to 30% of the offering is being reserved for retail, with E*Trade leading and Robinhood, SoFi, and [Fidelity](https://www.fidelity.com/?ref=aedilis) expected to get smaller pools.

Valuation: $1.75T–$2T target, $75B raise, implied float of 3.75%–4.3%. Dual-class structure leaves Musk with ~79% voting control on ~42% equity.

Pros: Largest U.S. IPO in history with a real recurring-revenue story underneath — Starlink alone now produces the majority of SpaceX revenue at 10M+ subscribers. Retail allocation is unusually generous. Index inclusion and ETF demand will create structural buy pressure.

Cons: Pricing assumes near-perfect execution on Starship development and continued military contracting. Voting control concentration in one person is a real governance risk. Mega-IPOs typically retrace 20–40% within 6–9 months as lockups expire — Facebook, Uber, Snap all traded below their IPO price within a year.

Retail action: Open or fund an E*Trade account this week. Enable IPO Access on Robinhood and IPO investing on SoFi. Decide your indication-of-interest number before the price range is set. Full mechanics in today’s separate Starlink deep-dive.

2. Anthropic — The Accelerating Outlier

What changed this week: TechCrunch reported April 30 that Anthropic’s next funding round could close at a $900B+ valuation within two weeks — meaning it may have closed by the time you read this. The round size is ~$50B. More importantly, The Information reports executives have discussed an IPO as soon as Q4 2026 with bankers expecting a $60B+ raise.

Valuation: Series G closed February 2026 at $380B post-money. The new round prices the company at $900B pre-money. That’s a 2.4x mark-up in three months. Revenue run rate just crossed $30B (announced this month).

Pros: Of all the AI labs, Anthropic has the cleanest enterprise revenue story — Claude is winning paid seats in Fortune 500 deployments at a faster clip than competitors. The valuation is high but tied to real ARR growth. A constitutional-AI safety brand resonates with regulatory-cautious enterprises.

Cons: Burning cash to keep up with model compute costs. Heavily dependent on Anthropic-Google and Anthropic-Amazon strategic deals — those are not arms-length transactions and could draw scrutiny in an S-1. $900B prices in essentially no risk that any competitor catches up.

Retail action: No direct path. Indirect exposure: ARKVX holds Anthropic in its top positions. DXYZ also has exposure. Both trade with minimums accessible to retail. If Anthropic does file in Q4 2026, this becomes the second-largest IPO of the year after SpaceX.

3. OpenAI — The Slower Sibling

What changed this week: Reporting solidified that CFO Sarah Friar is targeting regulatory filing in H2 2026 with a potential listing in 2027. Goldman Sachs, JPMorgan, and Morgan Stanley are advising. Sam Altman is pushing for a faster timeline — Friar is the brake.

Valuation: $852B post-money on the latest $122B round. Some analysts have floated a $1T IPO valuation. Annualized revenue crossed $25B in February 2026 (up from $6B at end of 2024). Internal projections show $14B in losses for 2026 — profitability not expected until ~2030.

Pros: ChatGPT is the most recognizable AI brand in the world. Enterprise revenue accelerating. Microsoft partnership locks in distribution and infrastructure.

Cons: The complicated nonprofit-to-for-profit governance restructuring remains an open question for an S-1. Losses are enormous. Customer concentration risk through Microsoft. Anthropic IPO-ing first would steal the AI-lab narrative.

Retail action: Same indirect playbook as Anthropic — ARKVX, DXYZ. If you must pick one to buy through the funds, ARKVX has cleaner OpenAI exposure. Both Anthropic and OpenAI sit in many of the same funds, so you don’t have to choose.

4. Databricks — The Profitable Wildcard

What changed this week: Largely quiet on news flow, which is actually the bullish read. The Series L closed December 2025 at $134B with $4B equity + $1.8B JPMorgan-led debt facility in January 2026. Pre-IPO debt is the bankers’ tell. Jason Lemkin’s prediction stands: “unless the market crashes, I think they’re going to IPO in the back half of ’26.”

Valuation: $134B. Revenue run rate $5.4B (+65% YoY). AI product revenue surpassed $1.4B. Positive free cash flow.

Pros: Profitable, which separates Databricks from every other AI name in this article. Lakehouse architecture is becoming industry standard. Customers include over 60% of the Fortune 500.

Cons: Competes head-on with Snowflake (public, struggling), AWS, Google, and Microsoft — all of whom have moved fast in lakehouse territory. The $134B valuation is roughly 25x forward revenue, which is steep even for AI infra.

Retail action: DXYZ has 4.0% Databricks exposure. ARKVX holds Databricks as a top-10 position. If you want the cleanest pre-IPO bet on AI infrastructure that has actual margins, this is the one. Watch for the S-1 in August or September.

5. Stripe — The “Never IPO” Story

What changed this week: Nothing new on IPO timing — and that is the news. Co-founder John Collison repeated in January and again in private comments this spring that going public “isn’t one of our top five or ten or twenty priorities.” The February tender at $159B (up from $91.5B a year earlier) confirms the model: Stripe will create liquidity for employees and early backers through repeated secondaries rather than an IPO.

Valuation: $159B as of February 2026. $1.9T in total payment volume in 2025 (+34% YoY). Profitable in 2025. Revenue suite tracking to $1B run rate in 2026.

Pros: Best-run private fintech in the world. Profitable. Aggressively expanding into agentic commerce and stablecoins.

Cons: The IPO you cannot have. If you’re optimizing for retail-accessible pre-IPO exposure, Stripe is dead money in your tracker until at least 2028.

Retail action: Skip the IPO speculation. The only practical way to play Stripe is the public payment processors that compete with or complement it — Shopify, Block, PayPal. Or accredited-only access through EquityZen secondaries if you qualify.

6. Cursor (Anysphere) — The Buyout Optionality Trade

What changed this week: The story keeps getting more bizarre. November 2025: $2.3B Series D at $29.3B. April 21, 2026: xAI announced it has the right to acquire Cursor for $60B later this year, or to pay $10B for joint work. Current talks: a $2B round at $50B+ valuation. Revenue went from $100M ARR in January 2025 to $2B in February 2026 — the fastest B2B SaaS scaling on record.

Valuation: $50B private round in progress, $60B xAI call option, $2B+ ARR. Analyst expectation: IPO filing late 2027 if the xAI option isn’t exercised.

Pros: Genuinely category-defining product. Revenue growth that has no SaaS comp. Optionality from the xAI deal — either you get bought out at $60B (which would be a private liquidity event, not a public one), or Cursor IPOs independently into a friendly AI-coding market.

Cons: AI coding is a knife fight. GitHub Copilot, Replit, Codeium, Windsurf, and Claude Code all moved fast in the past year. Revenue growth at $2B ARR will be much harder than $100M→$2B was.

Retail action: No public-market route until the IPO. The closest proxy is buying public companies that compete or partner — Microsoft (GitHub Copilot), or holding $TSLA-adjacent xAI exposure if you believe the buyout fires.

7. Polymarket — The ICE-Backed Long Bet

What changed this week: A $400M round at a $15B valuation is closing. Intercontinental Exchange (NYSE’s parent company) previously invested $600M and is rumored to be involved again. Analysts speculate Polymarket could leverage that ICE relationship for an NYSE direct listing as early as Q3 2027.

Valuation: $15B private. $10.6B in monthly notional volume in March 2026 — on par with mid-tier traditional derivatives exchanges.

Pros: Prediction markets are a regulated category now — CFTC-approved, ICE-backed, legitimately positioned. The volume growth is real. Direct listing via NYSE removes the underwriter middlemen.

Cons: Regulatory risk if a new administration tightens crypto-adjacent rules. Direct listings typically have weaker first-day support than traditional IPOs. The user base skews crypto-native, which could limit mainstream growth.

Retail action: Wait. No public exposure available. If you have an NYSE-affiliated brokerage account when the direct listing happens in 2027, you should be able to participate without restriction (direct listings don’t have allocation games — everyone trades from day one).

8. Crusoe — The AI Data Center Pure-Play

What changed this week: A pre-IPO funding round is in market right now (Axios first reported March 16). An employee tender priced at $13B — a 30% premium to the $10B October 2024 Series E. COO/CFO Michael Gordon, who led MongoDB’s 2017 IPO, was hired in early 2026. That’s the strongest single signal in this entire roster that an S-1 is being prepared.

Valuation: $13B on the most recent tender. Projected revenue: $152M in 2023, $2B in 2026, $5.5B in 2028 — a 105% CAGR. Currently building 1.8 GW (Wyoming) and 1.2 GW (Abilene, TX) campuses.

Pros: Picks-and-shovels play on the AI compute boom. Real physical infrastructure with multi-year revenue contracts. Vertical integration into stranded natural gas for power (the original Crusoe model) gives it cost advantages over peers.

Cons: Capital intensity is brutal. Customer concentration in hyperscaler buildouts. If AI capex enters a digestion phase, Crusoe is exposed.

Retail action: Watch the S-1 calendar — Gordon’s hire suggests a 2026 filing. In the meantime, public AI data center comps to track: Equinix, Digital Realty, Vertiv, and CoreWeave (now public).

9. Kraken — The Paused Trade

What changed this week: Co-CEO Arjun Sethi confirmed at the Semafor World Economy conference in April that the confidential S-1 filed November 19, 2025 is still on file. But the IPO is paused — Kraken put plans on hold March 18 citing market conditions. The Deutsche Börse $200M investment in April marked the valuation at $13.3B, down from $20B in November.

Valuation: $13.3B (down 33% from peak). Crypto exchange revenue down vs. 2024 peak. Deutsche Börse strategic partnership for European expansion.

Pros: Confidential S-1 already on file — Kraken could un-pause within weeks of a crypto market rally. Coinbase comparison still holds (Coinbase IPO’d at $86B in 2021).

Cons: Valuation has dropped meaningfully. The “pause” might extend. Crypto regulation in the U.S. remains the wild card.

Retail action: Coinbase (public, $COIN) is the working proxy. If Kraken does un-pause, it will likely be on the heels of a $COIN rally, so watch the comp.

Honorable Mentions — The Watch List

Perplexity: $22.6B valuation (Jan 2026). CEO Aravind Srinivas said “no plans of IPOing before 2028.” Track for proxy exposure if Google’s AI Search defensiveness intensifies.

Rune Technologies: Defense/AI logistics startup. Raised $24M Series A in 2025 (Andreessen Horowitz, Point72, others). Way too early for IPO talk, but sits in a hot defense-AI category alongside Anduril and Shield AI.

Pre-IPO Fund Performance — May 2026 YTD

Fund	YTD Return	Access	Top Holdings
DXYZ (Destiny Tech100)	+30%	NYSE — any brokerage	SpaceX 16.2%, Shield AI 4.1%, Databricks 4.0%, xAI 3.5%, OpenEvidence 3.5%
ARKVX (ARK Venture)	+17%	$500 min via Titan/ARK	SpaceX 18.0%, Anthropic, OpenAI, Databricks
ARKX (ARK Space ETF)	+12% (est.)	Any brokerage	Public space + SpaceX proxy exposure

The simplest pre-IPO portfolio for someone who doesn’t want to chase individual IPO allocations: a 60/40 split between DXYZ (NYSE-traded, instant liquidity, higher fee at 2.5%) and ARKVX (lower fee, quarterly tender liquidity, broader coverage). That single pair covers SpaceX, Anthropic, OpenAI, Databricks, xAI, and ~30 other private names.

The fee math is real — DXYZ’s 2.5% annual fee is steep relative to public ETFs. The trade-off is liquidity. If you’d rather pay less and accept quarterly liquidity windows, ARKVX is the better long-hold vehicle.

Heads up on Forge Global: If you’ve been reading older pre-IPO content (including the SpaceX article I published this morning), Forge Global was named as a secondary-market access route. As of March 2, 2026, Forge was acquired by Charles Schwab and delisted at $45/share. Functionally that means Forge’s secondary marketplace is now part of Schwab — accredited investors can still access private shares, but the standalone FRGE ticker is gone. EquityZen and UpMarket remain the other two major secondary platforms.

FAQ

Which of these is most likely to IPO in 2026?

SpaceX (June), Anthropic (Q4), and Databricks (late Q4 or Q1 2027) are the top three candidates with active bank involvement. OpenAI’s listing is more likely a 2027 event.

I can only fund one new brokerage account — which?

E*Trade for the next 60 days, specifically for SpaceX retail allocation. After the SpaceX IPO settles, the same account positions you for whatever comes next, because the lead underwriters rotate but Morgan Stanley sits on most of the largest deals.

Should I buy DXYZ or ARKVX?

If you want one-click access through your existing brokerage, DXYZ. If you want lower fees and don’t need daily liquidity, ARKVX. Holding both removes the choice.

What happens if SpaceX prices below the $1.75T rumored target?

It would be a positive sign for first-day demand. Mega-IPOs that price at the low end of their range historically pop harder on day one. The risk is the other direction — pricing at $2T+ with weak demand and a flat or down first day.

Is the AI IPO wave a bubble?

Hard question. Revenue at Anthropic, OpenAI, and Cursor is real and growing fast. Valuations relative to revenue are not unreasonable by recent SaaS standards. What’s bubble-like is the speed of valuation increases (Cursor 3x in 6 months, Anthropic 2.4x in 3 months). The right framework: assume any single AI-lab name could draw down 50% in a sector rotation, but the basket should outperform on a 5-year hold.

What’s coming next on this tracker?

Vol. 2 will add Klarna, Discord, Plaid, and Canva to the watch list, drop any name that goes fully public, and provide week-over-week valuation deltas. New every Monday at 8 AM ET.

Marcus Webb — Aedilis. Disclosure restated: I do not currently hold direct private positions in any company discussed. Aedilis has no current affiliate relationships with the brokerages or funds named. This is analysis and education, not investment advice. Always consult a fiduciary before making investment decisions.

Favorite