☕ The biggest IPO wave in history is coming. If SpaceX doesn't crash it first.

Sunday, April 26, 2026 | Private Markets Intelligence

Good morning. An AI coding assistant raised $2B at $50B valuation—six months after its last round. A chip company filed for the largest semiconductor IPO in history. And SpaceX is targeting a $1.75 trillion public debut in June that could either unlock the biggest exit wave ever or crash the entire market.

This is what happens when AI eats 80% of all venture funding—and concentrates in just four companies.

Alberto, Founderscrowd

In today's rundown:

Founderscrowd: Cerebras Systems filed publicly for IPO on April 17, targeting May at a $35B+ valuation. The company makes massive AI inference chips that process queries faster than Nvidia. It's raising ~$2B—the largest US semiconductor IPO in history.

Why it matters:

Nvidia dominates AI training chips. But inference (running AI models for users) is a $500B+ market with no clear winner yet. Cerebras built chips 56x larger than Nvidia's with custom cooling systems to handle heat.

The bet: AI needs speed. Users won't wait 10 seconds for ChatGPT responses. Cerebras promises sub-second inference at lower cost than Nvidia.

The risk? Cerebras tried to IPO in 2024 but got delayed by government inquiry. If this IPO flops, expect the hardware IPO window to slam shut for months. If it works, expect Groq, SambaNova, and Graphcore to follow.

Bottom line: First major AI hardware IPO of 2026. This either opens the floodgates or closes them for chip companies.

Founderscrowd: X-Energy went public Friday at $23/share (above range), raising $1 billion at a $12B valuation. The company has zero reactors built, zero regulatory licenses, and zero customers. First reactor delivery: 2030 (maybe).

Why it matters:

AI's hitting a power wall. Data centers are maxing out grids and states are banning new builds. Maine banned data centers entirely. X-Energy's pitch: small modular reactors built next to data centers for dedicated power.

Amazon is the lead backer. Why? AI needs nuclear or growth stops. No nuclear = no new data centers = OpenAI can't scale = Microsoft loses.

Nuclear's the only tech that's carbon-free, always-on, and big enough for AI's scale. Google and Microsoft are making similar bets.

Bottom line: Investors paid $12B for a pre-revenue infrastructure bet because AI can't scale without solving power first. Speculative? Absolutely. But whoever solves AI's energy problem wins hundreds of billions.

Most of these opportunities disappear before they hit the news.

Founderscrowd Premium members get first look at exclusive private market deals, the kind institutional investors see before they go mainstream. We're talking early-stage access to companies raising now—AI infrastructure, fintech solving broken markets, real businesses preparing for exits.

This week, Premium members got access to:

✅ XYZ — Fintech platform solving $80B broken market (CEO has 3 exits, Jeremy Liew invested, 40% NJ tax rebate)
✅ XYZ— $20M revenue business, 43% CAGR, preparing 2030 exit to Helen of Troy or Lifetime Brands

Premium includes:

✅ Weekly investment memos on live deals
✅ Full pitch decks, financials, and founder contact
✅ Private markets intel before it's public
✅ Pre-IPO secondary opportunities (SpaceX, OpenAI, Databricks)

The difference? When billion-dollar deals hit the news, Premium members already had a first look at the investment months earlier.

Become a Premium member →
Lock in $40/month before it goes to $100+ (this is your last week)

Founderscrowd: Kailera Therapeutics (obesity drugs) raised $625M in an upsized IPO last week the largest biotech debut of 2026. Odyssey Therapeutics and Mobia Medical also filed. Nuclear and defense companies are going public. Biotech is hot.

Why it matters:

The IPO market is open—but only for biotech, energy, defense, and chips. Software? Frozen.

Figma is down 66% from its peak. Navan is down 50%. Liftoff withdrew its IPO this month. Zero SaaS unicorns have filed in 2026.

The reason: AI disruption fears. Anthropic's Claude Cowork wiped $2 trillion from software stocks in February when it showed AI agents could automate entire workflows. Investors think AI will replace SaaS products, so they're not buying software IPOs.

Bottom line: If you're in SaaS, expect to stay private or sell to strategic buyers. Public exits are off the table for 2026. But if you're in biotech, energy, or defense—IPO window is wide open.

Founderscrowd: Q1 2026 shattered every funding record: $300 billion invested globally (up 150% QoQ and YoY). But the distribution is insane.

The top 4:

Those four companies took $188B—65% of all global VC funding. AI overall captured 80% of total venture dollars.

Why it matters:

For everyone else? Deal count dropped 26%. If you're raising and you're not AI, you're competing for the remaining 20% while reading headlines about record venture funding.

The market is bifurcating: mega-rounds for AI infrastructure giants, discipline everywhere else. Seed and Series A rounds are up marginally, but late-stage non-AI rounds are down dramatically.

Bottom line: Venture is back, but only if you're building AI infrastructure or you have bulletproof unit economics. Everyone else is in a different market than the headlines suggest.

What this week reveals:

1. AI infrastructure is the only game that matters. Nuclear (X-Energy), chips (Cerebras), compute (Anthropic's $65B), data centers (PE's $200B)—infrastructure is where the money went. Apps are speculative. Infrastructure is a toll road.

2. The IPO market is bifurcating. Biotech, energy, defense, chips: yes. SaaS: no. The "software-mageddon" triggered by Claude Cowork wiped $2T from software stocks. Investors think AI will replace SaaS, so software IPOs are frozen.

3. Venture funding is concentrating. Q1 2026: $300B raised, but 65% went to 4 companies. If you're not AI infrastructure, you're fighting for scraps.

4. SpaceX June IPO is make-or-break. $1.75T valuation on $16B revenue (109x multiple) either works or crashes the market. Every other 2026 IPO is waiting to see what happens.

5. PE has no exit except secondaries. Firms are sitting on trillions in unsold assets. IPOs are slow. Strategic buyers aren't paying 2021 prices. So PE is selling to… other PE firms. Continuation funds are the new normal.

For investors: Infrastructure (nuclear, chips, data centers) is getting funded at pre-revenue valuations. AI coding tools (Cursor) are going parabolic. The biggest IPO wave in history is coming H2 2026—but only if SpaceX doesn't crash the market first.

SpaceX June IPO:

Cerebras May IPO:

Q2 funding concentration:

PE secondary market:

An AI coding tool went from $29B to $50B in six months. A chip company filed for the largest semiconductor IPO ever. SpaceX is targeting a $1.75 trillion public debut in June. And Q1 2026 broke all records—but 65% went to just 4 companies.

The pattern: AI infrastructure is everything. Apps commoditize. Infrastructure captures value. Whoever controls power, chips, and compute wins.

⭐⭐⭐⭐⭐ This is the signal
⭐⭐⭐ Average
⭐ Needs work

See you Sunday,

Jose ☕

Founderscrowd

P.S. Sunday we're breaking down why $300B in one quarter doesn't mean what you think it means—and what the concentration in 4 companies reveals about where venture is actually heading.