# 2026 Is The Biggest IPO Year In History. But Take Out SpaceX And Right Now It’s One of the Worst. > Source: > Published: 2026-08-03 12:58:53+00:00 ### A good year for IPOs. A bit less so for their performance. And not a year at all for SaaS. ** Per PitchBook there has been $1,786.8 billion in US VC-backed IPO value through July 28.** That’s not a record. That’s five prior years stacked on top of each other and then doubled. Then you do the subtraction. SpaceX went public on June 12 at $135 a share, at a valuation of roughly **$1.77 trillion**. It sold 555.56 million shares and raised $75 billion, $85.7 billion once underwriters took their overallotment. The single largest IPO ever executed, by a wide margin. So of that $1,786.8B: about $1,770B is one rocket + coding/AI + tweet company. **The other 43 VC-backed IPOs this year combined are worth roughly $17 billion.** That’s not a boom. That’s 2023 with a trillion-dollar asterisk stapled to it. ## 1. The count says “normal.” The value says “one guy.” 44 VC-backed IPOs through July 28 versus 50 for all of 2025. Annualize it and you land somewhere around 75, so the *count* is up meaningfully. But the count per se was never the problem even in lean years. 2023 and 2024 both had roughly 40-45 IPOs too. The problem has always been that the good companies weren’t going out. Looking at the 2026 class, that hasn’t changed as much as the chart suggests. Databricks’ CEO said 2026 isn’t a good time to go public. That says it all for B2B. Also note **PitchBook’s “deal value” is post-money valuation at listing, not money raised.** Nobody handed $1.8 trillion in cash to founders and LPs this year. The actual proceeds number is a rounding error against that bar. ## 2. The aftermath has been a bit ugly This is the part everyone skips because the headline valuations are so loud. **SpaceX: down ~30% from its debut** as of Thursday’s close**Cerebras: down ~34.7%**, and that’s*after*a 19.9% one-day gain The excuse being offered is macro. The war in Iran, Chinese chipmakers coming for US AI compute. But look at what these companies have in common instead: **Enormous capex. Long roads to revenue. Businesses whose entire story is “the AI market keeps going up.”** SpaceX spent $10.1 billion on capex in the three months ending March, versus $4.1 billion the year before — most of it on AI. It lost nearly $5 billion in 2025. CFRA initiated coverage at *sell*, citing capital intensity and elevated valuation expectations. Public investors are doing the thing they always eventually do. They’re at least in part pricing the cash flows, not the narrative. ## 3. The private-public gap didn’t close. It moved. For four years the story was “the IPO window is shut.” It wasn’t, exactly. What was shut was the willingness of public investors to pay 2021 private-round prices. 2026 tested that. A handful of the most-hyped private companies in the world went out at or near their private marks, and the market marked them down 30%+ within weeks. **The gap between the last private round and what a public investor will pay is still there.** It just got tested at a much larger scale, with a much better company, and the answer came back the same. ## 4. Not one of the 2026 IPOs is classic B2B software Go down the 2026 list. SpaceX. Cerebras. Quantinuum. X-Energy. HawkEye 360. Space, AI silicon, quantum, nuclear, geospatial intelligence. **Zero application software companies. Zero B2B software.** And it’s not a timing thing. Crunchbase looked at new IPO filings and found no venture-backed B2B unicorn had filed at all in 2026. Not one submitted paperwork. Enterprise software has gone quiet. Databricks was supposed to be the one. $5.4 billion in revenue growing 65%, a $134 billion valuation, exactly the profile that used to get you out. Ali Ghodsi went on Bloomberg in June and said: *“We will be a public company. I just think this is a terrible year to go public.”* When a company at $5.4B growing 65% says the market isn’t there, the market isn’t there. ## 5. And the last SaaS class is why This is the real reason the pipeline is empty, and it has nothing to do with Iran. **Figma** went out in July 2025 at $33, hit $115 on day one, ran to $143. It’s now around $20-21. Down roughly 83% from the high, and below its IPO price. The proximate cause wasn’t rates or macro — it was Anthropic announcing Claude Design in April 2026, a direct competitor to Figma’s core product. The stock fell 28% in March alone on general AI-disruption fear before that even landed.**Navan and Chime** are both trading below their opening prices.**Asana is down ~50% in 2026. Adobe is down 29%. ServiceNow is off more than 30%. Salesforce fell 35% over six months.** That’s the mechanism. Public markets aren’t just declining to pay 2021 prices for SaaS. They’re actively pricing in the possibility that a foundation model company ships your product as a feature. That’s a different and much harder problem than a multiple reset. ## 6. The market isn’t repricing B2B. It’s sorting it. Run the other list. **Datadog is up more than 80% in 2026,** trading near a 52-week high.**Cloudflare is up ~35%.****Snowflake** was 50% below its year-ago price in April and is now positive on the year, after its best single day ever.**MongoDB** was down more than 20% early and has recovered.**JFrog** jumped 43% on earnings. Bank of America’s “Fab Five” — Snowflake, Datadog, JFrog, MongoDB, Twilio — were collectively up 30% year to date against a software ETF down 12%. In May, software stocks had **their best month since 2001.** The “SaaSpocalypse” talk faded almost entirely. So the dividing line isn’t quality, and it isn’t growth. Salesforce is a great business. Adobe prints cash. Both got hammered. Datadog and Snowflake got re-rated to highs. **The line is whether AI makes you more necessary or less.** Datadog’s argument is that AI makes software systems more complex, and complexity means more to monitor. Q1 revenue crossed $1 billion for the first time with growth *accelerating* — 25% to 29% to 32%. OpenAI is reportedly its largest customer. Cloudflare’s Matthew Prince told investors AI is re-platforming the internet and is the biggest tailwind in the company’s history. Snowflake signed a $6 billion cloud and chip deal with Amazon and raised guidance. Figma’s problem is that Anthropic shipped a design tool. The pattern is hard to miss: **consumption-priced infrastructure that AI consumes more of is winning. Seat-priced application software that AI might do instead is losing.** That’s a much more useful read than “multiples compressed.” Multiples didn’t compress. They separated. ## 7. The bright spots in the IPO class itself Two things cut against the gloom, and both deserve to be said: **Biotech worked.** A Bloomberg biotech IPO index was up 55% through July 21 — outperforming the 2026 AI IPO class outright. When the story is a binary clinical outcome instead of a compute narrative, public investors showed up. **And the broader IPO cohort has held.** The Renaissance IPO Index — the largest and most liquid US IPOs of the last three years — was up 12.7% on the year through Wednesday, against the S&P 500’s 8.3%. Recent IPOs as a class are *beating* the market. So the aftermath isn’t uniformly bad. It’s specifically bad for capital-intensive AI infrastructure bets that priced at private marks. A record year for liquidity has not been a record year for the stocks that produced it. And for B2B SaaS specifically, it hasn’t been a year at all. *Source: PitchBook, US VC-backed IPO activity as of July 28, 2026.*