IPO Markets Rebound: $1.2 Trillion in Listings Resume Despite Firmus Setback

After September's silence, October has exploded with initial public offerings. Twelve IPO listings, announcements, and filings in just five days signal a dramatic reversal in the public markets—yet investor selectivity has become ruthless. When Firmus withdrew its $30 billion AI infrastructure IPO, a clear message emerged: not every high-valuation company deserves to go public, and founders are learning to read market signals faster than ever.

The past week shows a market finding its footing again, with winners and losers emerging fast. TRexBio priced its biotech IPO. RISC-V semiconductor maker ESWIN Computing listed in Hong Kong. MNT-Halan executed a blockbuster float. Infomaniak, an IT service provider, listed in Switzerland. Yet Firmus's collapse reminds us that size alone no longer guarantees success—profitability, unit economics, and demonstrated market traction are now table stakes for the public markets.

This shift represents a reset from the 2021 IPO mania, when "growth at all costs" carried companies worth tens of billions to public markets. Today's investors are asking harder questions about defensibility, geography, and cash generation. October's IPO activity proves selective capital isn't scarce—it's just far more discerning about where it flows.

Hong Kong's Ascent as the AI IPO Hub

Hong Kong is consolidating its position as the venue for AI and semiconductor IPOs, a geographic shift with profound implications for how capital flows across borders. ESWIN Computing, a RISC-V chip designer, listed this week. More significantly, Kuaishou's Kling AI—a generative video model company—has selected banks for a Hong Kong IPO targeting at least $1 billion. This follows the global pattern: as US and European IPO windows narrow due to regulatory tightness and macro volatility, Asia's capital markets are stepping in to absorb high-growth tech companies.

The geographic shift reflects two converging dynamics. First, Hong Kong investors—particularly sovereign wealth funds and Asian PE firms—have demonstrated greater tolerance for pre-profitability tech companies. These investors take a longer view on returns and understand infrastructure differently than US public market investors. Second, regulatory paths are clearer in Hong Kong than in the US, where SEC scrutiny of AI companies has intensified. The US Securities and Exchange Commission's heightened focus on AI model validation, data provenance, and labor practices has created friction for would-be IPO filers. Kling's $1 billion IPO target is conservative relative to Silicon Valley venture valuations, but it signals confidence that the company can achieve solid unit economics and defensible moats in generative video.

The shift matters because it redistributes capital allocation. AI companies that might have targeted $5-10 billion US IPOs three years ago are now comfortable with $500M-$1B Hong Kong listings. This bifurcation—with select US companies able to access American public capital, while others navigate Asian exchanges—is reshaping where innovation gets funded and where it gets scaled.

Australia's Infrastructure Play: Firmus's Cautionary Tale

Australia's equity market experienced two opposite forces this week. MNT-Halan, a mobile money platform, executed a "blockbuster" IPO, while Firmus withdrew its $30 billion listing on the Australian Securities Exchange. The contrast is instructive for understanding how the public market is sorting winners from losers.

The Firmus retreat is the week's most important data point. At $30 billion valuation, Firmus positioned itself as an AI infrastructure provider aiming to raise $4.4 billion for servers and deployment. The company faced structural headwinds: competitive pressure from established cloud giants (AWS, Google Cloud, Microsoft Azure) with better unit economics and customer relationships, questions about capital efficiency in infrastructure buildout cycles, and macroeconomic skepticism about the sustainability of the AI capex boom. Industry observers noted that raising $4.4 billion for servers is only meaningful if you can deploy them faster and cheaper than competitors—a difficult argument when giants control pricing.

Withdrawing the IPO rather than pricing at a 50% markdown suggests management correctly calculated it better to wait. Two years ago, such a retreat would have been seen as a disaster, a failure to capitalize on market momentum. Today, it's pragmatic. Founders are learning that hitting public markets at the wrong price is worse than waiting for conditions to improve. The fizzled listing demonstrated that even "hot" AI infrastructure narratives must clear a profitability hurdle.

MNT-Halan's success, by contrast, reflects investor appetite for fintech with proven traction. Financial services companies that have demonstrated path-to-profitability, geographic expansion potential, and unit economics improvements are seeing green lights from public markets. Mobile money platforms operating in emerging markets with regulatory tailwinds and low legacy competition have a clearer runway to public markets than generalist infrastructure plays.

Sectors in Motion: Biotech, Semiconductors, Services, and AI Specialization

The week's IPO mix reveals exactly where investor conviction lies in 2026:

Biotech: TRexBio announced pricing for its initial public offering. The company operates in a sector that has faced a multi-year funding drought post-2021, when IPO windows slammed shut for most biotech startups. A successful public market entry signals selective recovery—likely concentrated in companies with approved or near-approval assets, validated clinical data, and clear regulatory pathways. Speculative biotech platforms are still radioactive in public markets.

Semiconductors: ESWIN Computing's Hong Kong listing is part of a broader trend: chip design and advanced semiconductor companies are attracting capital even as hardware cycles remain uneven. RISC-V architecture adoption, the geopolitical pivot to sovereign chip capacity, and US export restrictions on advanced semiconductors create genuine long-term tailwinds for alternative chip designers.

Services and Established SaaS: Infomaniak, a Swiss IT service provider, listed successfully this week. The company exemplifies the type of IPO thriving in 2026: cash-generative, profitable, with clear unit economics and defensible customer relationships. It's the antithesis of the cash-burn model that fueled 2021 IPO frenzies when growth-at-all-costs was the mantra.

Specialized AI: The absence of large foundation model companies in this week's IPO listings is notable. Kling AI's $1 billion target is telling—it positions the company as a specialized AI tool (generative video) rather than a general-purpose LLM competitor. That focus is what makes it IPO-ready. General-purpose AI platforms are still too capital-intensive and margin-uncertain for public markets.

What Investors Want Now: The New IPO Criteria

The October IPO window reveals three non-negotiable criteria for public market success in 2026:

1. Unit Economics That Make Sense at Scale. Firmus's $4.4 billion deployment for infrastructure sounds impressive until you calculate payback periods and compare to incumbent cloud providers' capital efficiency. Successful IPO candidates either operate defensible niches (chips, specialized biotech assets, regional fintech) or have achieved profitability with predictable growth. Investors are done funding "hockey stick" projections from unproven teams.

2. Geographic Diversification and Sovereign Appeal. ESWIN's RISC-V focus benefits from the US-China semiconductor decoupling and emerging nations seeking alternatives to ARM. Kling's AI video model appeals to Asian platforms constrained by US export controls. Even MNT-Halan's fintech model resonates across markets underserved by traditional banking infrastructure. The most IPO-ready companies have multiple geographic revenue streams and defensibility against one regulator's decisions.

3. Realistic Timing and Valuation Reset. The fact that TRexBio priced its IPO (rather than withdrawn or postponed) suggests the company negotiated expectations with underwriters—likely a more modest valuation than a 2021 filing would have claimed. The pricing represents a reset to fundamentals: what's the company worth based on demonstrated cash flow and growth rate, not venture hype?

What This Means for Q4 and Beyond

October's IPO surge—12 filings and listings in five days—doesn't signal a return to 2021-level exuberance. Rather, it indicates a bifurcated market with increasingly clear dividing lines. Some categories (profitable SaaS, established semiconductors, fintech with proven traction, biotech with approved assets) have a clear path to public capital. Others (cash-burning infrastructure plays, pre-revenue AI platforms, speculative biotech) face years of private capital rounds or acquisition routes to liquidity.

The week's activity is the market's way of saying: we're selective, we're serious about returns, and we're moving on from last cycle's excess. For founders and executives preparing to go public, the message is unambiguous: traction beats hype, profitability beats growth-at-all-costs, and geographic optionality beats single-market narratives. Firmus's withdrawal and Kling's conservative $1 billion target are not failures—they're wisdom. The companies that read these signals will have a clearer path to the public markets than those still chasing 2021's playbook.