September Highlights: FAPMI outpaced FPMI as private market performance remained selective
September produced positive headline returns in the private market, but the composition of those gains differed. The Forge Private Market Index (FPMI) rose 1.39%, while the Forge Accuidity Private Market Index (FAPMI), which includes public-price exposure, gained 3.38%. FAPMI's stronger return reflected concentrated support among its larger constituents, while the equal-weighted private company benchmark registered a more uneven advance.
FAPMI's lead over FPMI reflected the effects of cap-weighting and public-price exposure during a selective month. SPY declined 0.33% while QQQ gained 3.31%, providing a public market backdrop of AI enthusiasm alongside concern about yields and inflation.1
The Forge Artificial Intelligence (AI) thematic basket led September, rising 7.74%. The Forge Aerospace & Defense thematic basket gained 4.96%, and the Forge Consumer thematic basket edged up 0.62%. The Forge Chips thematic basket declined 2.07% and the Forge Fintech thematic basket fell 3.26%, leaving the positive index returns supported by a selective set of companies and exposures.
FAPMI reflected the public-growth split
FAPMI's stronger return reflected a different mix from the equal-weighted FPMI. OpenAI (+32.26%) was its largest identified contributor as the company is reportedly raising additional capital ahead of a potential IPO in 2027.2
Among FAPMI's public-price components, SpaceX (+4.99%) advanced while Cerebras (-3.58%) declined, producing an offsetting effect; Anduril (+7.29%) also supported the cap-weighted benchmark.
The result is important because FAPMI does not simply replicate the broader private company cohort. Its larger weights and public-price exposure made it more responsive to the month's growth-stock divide, while the private-company benchmark registered a smaller, more distributed gain.
Private-company returns were positive, but uneven
The positive FPMI return included several substantial advances. Fin, fka Intercom (+78.73%), was the largest named gain, followed by Lambda (+40.84%), Cresta (+29.22%) and Lyten (+26.09%). Fin rose sharply as Salesforce acquired the company for $3.6 billion on September 10.3,4 Those moves provided support, but did not describe the full private company universe.
The downside was equally visible. Airtable (-56.12%) and Postman (-33.67%) recorded the steepest named declines, while Kraken (-19.09%) and SambaNova Systems (-12.43%) also weighed on the index. Airtable's downward move was due to an acquisition by Bending Spoons with a lower than anticipated price of $1.28 billion.5 The coexistence of those losses with September's strongest gains explains why a positive FPMI result should not be read as a broad-based rise across the private market.
September therefore reinforced the distinction between benchmark direction and underlying breadth. The FPMI gain was constructive, but it was assembled from sharply different company-level outcomes rather than a uniform repricing of private company marks.
Thematic results supplied support, not a single narrative
AI supplied the clearest upside. OpenAI added 5.78% to the AI basket, while Lambda and Anduril also moved higher. Aerospace & Defense gained 4.96%, with Anduril and Saronic (+6.16%) contributing to the advance.
The Forge Chips thematic basket moved in the opposite direction, declining 2.07%, as Ayar Labs (-2.68%) and PsiQuantum (-3.57%) weighed on the result. The Forge Fintech thematic basket fell -3.26%, pressured by Kraken and Ripple (-11.04%).
Consumer was closer to flat, rising 0.62%. WHOOP (+13.53%) provided support, while The Farmer's Dog (-8.87%) detracted. The mixed thematic picture complements the benchmark story: the month's gains were real, but their sources differed materially by company and category.
IPO attention centered on frontier-AI economics
IPO speculation in September was concentrated in AI. Anthropic (+0.00%) was reported to be preparing for a potential listing.6 Its prospectus showed nearly $4.6 billion of 2025 revenue, but also more than $8 billion in operating losses and almost $13 billion in operating expenses as computing costs rose; the company outlined $518 billion of future cloud, computing and infrastructure spending.7
A potential IPO would therefore test whether public investors are prepared to fund frontier-AI growth at that cost structure. OpenAI's management publicly signaled caution on its own timing, saying the company would not go public in 2026.8 TechCrunch reported that OpenAI was in talks to raise at least $30 billion in a pre-IPO round at roughly a $1.4 trillion valuation, following its $122 billion March raise at an $852 billion valuation; the proposed financing was described as a bridge to an IPO.9 Beyond AI, Kalshi (+0.01%) was reported to be considering an IPO as soon as 2027,10 while Revolut was reported to be evaluating a potential dual listing in London and New York.11
September was constructive, but selective
Taken together, September was positive for both Forge benchmarks, but the composition of those gains mattered. FAPMI's 3.38% return outpaced FPMI's 1.39% gain as AI-linked and public-price exposure found support. The same data showed weaker Chips and Fintech results, sharp declines among selected private companies, and an IPO backdrop defined more by readiness and timing than a broad rush to list.
