INDEPENDENT VERDICT
Fab2, Inc. is a verifiable Delaware semiconductor company formerly known as Atomic Semi, and its September 8, 2026 Form D documents one of the larger private semiconductor financings of the year. The filing reports a total offering of $587,997,030, with $556,021,222 sold to 89 investors and $31,975,808 remaining after a first sale on August 21. The securities consist of Equity together with securities to be acquired upon exercise of an option, warrant or another right to acquire securities; the company relies on Rule 506(b), is not a pooled investment fund and reports no sales commissions or finder fees. Sam Zeloof appears as President, CEO, executive officer and director, Jim Keller as a director, and Corey Nobile as a director. Most importantly, the filing contains an unusually useful clarification: $88,000,000 of the reported amount sold represents conversion of securities.
That $88 million sentence changes how the entire financing should be read. Fab2 publicly announced a $500 million Series A at a $3.7 billion valuation, while the Form D shows $556.02 million already sold. Those two figures look inconsistent until the conversion disclosure is incorporated. The SEC amount sold is not necessarily synonymous with new cash deposited in the Series A closing; part of it represents securities that converted into the securities being reported in this offering. FilingDossier therefore does not describe Fab2 as having raised $556 million of fresh Series A cash. The better interpretation is that the Form D captures a broader capitalization event around the Series A, including $88 million of converted securities, while the company's public announcement describes the new financing round as $500 million.
THE $88M CONVERSION IS THE KEY TO UNDERSTANDING THE ROUND
This is one of the rare Form D cases where Item 13 itself explains why media financing numbers and SEC numbers diverge. Fab2 reports $587.997 million as the total offering and $556.021 million sold, but then expressly states that $88 million of the amount sold includes conversion of securities. Conversion commonly occurs when earlier SAFEs, convertible notes or similar instruments turn into equity during a priced financing, although Fab2's Form D does not specify which instrument or instruments account for the $88 million. It would therefore be inappropriate to label the entire conversion amount as SAFE financing, debt conversion or a specific prior round without the underlying capitalization documents.
The economics are materially different depending on what converted. If an earlier security carried a valuation cap or discount, its holders may have acquired Series A-equivalent shares at economics different from new investors paying the current round price. If the converted securities were notes, accrued interest may also have affected the number of shares issued. If warrants or another right were involved, the capitalization impact could differ again. Investors evaluating Fab2's $3.7 billion valuation therefore need the fully diluted cap table, conversion schedules and preferred-stock terms rather than assuming the public $500 million divided by the headline valuation tells the entire ownership story.
The Form D itself gives another clue by selecting not only Equity but also "Security to be Acquired Upon Exercise of Option, Warrant or Other Right to Acquire Security." That selection is consistent with a financing involving more than one simple class of newly issued shares, but it does not identify each instrument. The remaining $31.98 million also means the SEC offering had not technically reached its stated ceiling when filed, even though Fab2 publicly characterized the transaction as a completed $500 million Series A. Public financing announcements and Form D ceilings measure different things and should not be forced into a single number.
FROM ATOMIC SEMI TO FAB2: THE COMPANY CHANGED ITS NAME BECAUSE ITS PRODUCT CHANGED SCALE
The SEC filing directly preserves `Atomic Semi, Inc.` as Fab2's previous name, creating a clean legal bridge between the startup first reported in 2023 and today's much larger Texas semiconductor company. Atomic Semi emerged around an unusual partnership between Sam Zeloof, known for building semiconductor fabrication capability at garage scale, and veteran chip architect Jim Keller. Keller's career spans major processor programs at companies including AMD, Apple, Tesla and Intel, while Zeloof approached semiconductor fabrication from the opposite direction: instead of designing increasingly complex processors for established mega-fabs, he demonstrated that simplified semiconductor manufacturing could be recreated with much smaller equipment and dramatically lower infrastructure requirements.
Early reporting around Atomic Semi described a roughly $15 million seed financing led by the OpenAI Startup Fund at an approximately $100 million valuation, with investors associated with the company including Naval Ravikant, Nat Friedman and Fred Ehrsam. By September 2026, Fab2 itself publicly stated that its Series A was $500 million at a $3.7 billion valuation. Even allowing for differences between reported seed terms and the later formal Series A, that represents an extraordinary increase in capital requirements and implied enterprise value over roughly three years. The reason is that Fab2 is no longer framing itself as merely a faster prototyping service or small chip startup; it is trying to industrialize the machinery and factories required to manufacture semiconductors.
The rebrand from Atomic Semi to Fab2 therefore carries more strategic meaning than a typical startup name change. The new name refers to what the founders call a "fab fab": a factory whose output is semiconductor fabrication factories. Fab2 says it designs the hardware and software needed to manufacture chips, produces the components and manufacturing tools itself, integrates those tools into complete production systems and ultimately wants to replicate the fabs themselves at industrial scale. The company's ambition is therefore recursive: manufacture the components, manufacture the machines, manufacture the fab, and then use those fabs to manufacture chips.
FAB2 IS TRYING TO ATTACK THE FAB EQUIPMENT STACK, NOT JUST THE FOUNDRY MARKET
Traditional semiconductor manufacturing depends on a highly specialized global supply chain. A large advanced fab can rely on lithography systems, vacuum equipment, deposition equipment, etch systems, metrology, gas handling, robotics, pumps and process-control technology sourced from numerous specialist suppliers. Fab2's proposed model challenges that architecture by vertically integrating far more of the tool stack. Public descriptions of the company say it builds equipment including pumps, valves, gas lines, vacuum chambers and lithography-related systems internally, then combines hardware with software-driven process automation.
That matters because Fab2 should not simply be compared with TSMC, Samsung or Intel Foundry. Those businesses optimize enormous fabs for high-volume wafer throughput, advanced process control and economies of scale. Fab2's concept instead emphasizes smaller, replicable, software-defined production units, fast process iteration and the ability to prototype or manufacture specialized chips without routing every product through a conventional mega-fab. Its lithography approach has included direct-write electron-beam techniques, which can eliminate expensive masks and accelerate iteration but carry an obvious throughput disadvantage compared with high-volume optical or EUV lithography.
That tradeoff defines the commercial thesis. If Fab2 succeeds, its strongest initial markets may not be leading-edge mass-market smartphone processors or gigantic AI accelerator volumes where wafer throughput dominates economics. The model may instead be better suited to prototyping, research, custom silicon, defense applications, specialized analog or sensor devices, unconventional process development and customers that value cycle time or manufacturing sovereignty more than maximum wafer volume. Over time Fab2 could improve throughput, parallelize equipment or expand process capabilities, but the public evidence does not yet establish that a small Fab2 production system can economically replace a conventional high-volume commercial foundry.
THE TEXAS MOVE IS PART OF THE MANUFACTURING STRATEGY
Fab2's geographic shift is also more than a headquarters relocation. The company moved its center of gravity from California to Texas during 2026 and now lists 6301 Stassney Lane in Austin as its principal business address in the SEC filing. Public reporting describes three operating locations: an Austin headquarters and R&D/production site, a Lockhart-area facility intended to house the "fab fab" manufacturing operation, and the original San Francisco facility associated with the company's earlier garage-fab work.
Some public descriptions have referred to approximately 120,000 square feet in Austin, roughly 30,000 square feet in Lockhart and approximately 25,000 square feet in San Francisco. Those numbers should be treated carefully. Local permit reporting around the Austin address has identified individual renovation phases covering substantially smaller areas than the widely cited 120,000-square-foot total, so a permit for one project phase does not independently prove or disprove the company's entire occupied footprint. The more defensible conclusion is that Fab2 has materially expanded in Texas and is building both R&D and production capability there, while exact currently commissioned manufacturing square footage should be verified from leases, permits and operational disclosures rather than copied from a single headline.
The location itself is strategically relevant. Central Texas already has a large semiconductor ecosystem including fabs, equipment suppliers, engineering talent and advanced manufacturing infrastructure. Fab2's model requires mechanical engineers, precision systems engineers, vacuum specialists, semiconductor process engineers, controls developers and software engineers in the same organization. Job postings also indicate that some positions involve technology subject to U.S. export controls, reinforcing that portions of Fab2's semiconductor manufacturing work may sit within strategically sensitive technology categories even though the company has not publicly positioned itself as solely a defense contractor.
STUDIO CREATES A SOFTWARE LAYER AROUND THE HARDWARE BUSINESS
Fab2's vertical integration does not stop at fabrication equipment. The company also operates Studio, formerly Atomic Studio, an in-browser collaborative EDA environment for schematic capture, chip layout and simulation. That software layer is important because the company's broader thesis depends on compressing the cycle between chip design, process definition and physical fabrication. Conventional semiconductor development can involve long handoffs among design tools, mask preparation, foundry process rules, production scheduling and external fabrication partners; Fab2 wants more of that loop inside a unified hardware-software system.
If that integration works, the defensibility may come from process data and manufacturing automation as much as from any single piece of equipment. A fleet of standardized small fabs could theoretically generate repeatable process data, allowing recipes and manufacturing improvements to propagate between machines. That would make the business resemble a mixture of semiconductor equipment company, foundry, EDA software developer and industrial automation platform rather than fitting neatly into one established semiconductor category.
But vertical integration is also expensive. Building custom pumps, chambers, motion systems, lithography equipment, process-control software and fabs simultaneously creates an enormous engineering surface area. Established semiconductor equipment suppliers spend decades refining reliability, contamination control, uptime, process repeatability and field service. Fab2's $500 million public Series A therefore looks less surprising when viewed against the number of industrial systems it is attempting to internalize. Investors are not financing one product; they are financing an attempt to rebuild a meaningful portion of the semiconductor manufacturing stack.
THE INVESTOR MIX ALSO SAYS SOMETHING ABOUT THE STRATEGY
Fab2's publicly identified Series A investors include Fundomo, Corner Capital, Paradigm, Duquesne Family Office, Maverick Silicon, StepStone Group, Protagonist, UDC Ventures and Naval Ravikant, with other investors also associated with the financing. Universal Display's venture ecosystem is particularly notable because Universal Display is an operating public-company supplier of OLED materials and technologies rather than a generalist software venture investor. Participation from investors with semiconductor or industrial exposure can provide commercial or technical network value beyond capital, although the public materials do not establish strategic supply agreements between every investor and Fab2.
The Series A's reported $3.7 billion valuation should also be understood in context. It represents the company's publicly stated financing valuation, not a value independently verified by the SEC. The Form D does not disclose preferred-stock liquidation preference, anti-dilution terms, participating rights, option pool expansion or the percentage of the company sold. Consequently, a simple calculation using $500 million and $3.7 billion cannot reconstruct the exact fully diluted ownership purchased by new investors, particularly because $88 million of converted securities is already embedded inside the SEC financing record.
The leap from a reported approximately $100 million seed-era valuation to $3.7 billion also means expectations have increased dramatically. At this valuation, investors are effectively underwriting much more than an interesting prototyping platform. They are underwriting Fab2's ability to manufacture semiconductor equipment reliably, reproduce fabs economically, develop useful process nodes, attract repeat customers, create a service and maintenance organization, and eventually support production economics capable of justifying a multibillion-dollar enterprise.
A "FAB FAB" CREATES A DIFFERENT RISK MAP THAN A NORMAL CHIP STARTUP
Fab2 faces several layers of technical and commercial risk simultaneously. Electron-beam direct writing offers maskless flexibility but faces throughput limitations; semiconductor tools must maintain extreme mechanical, thermal and contamination tolerances; process recipes must achieve consistent yield; and customers must trust a young company with sensitive chip designs and manufacturing dependencies. Even if individual machines work, reproducing an entire fab economically requires standardized installation, qualification, calibration, service and supply chains.
There is also a strategic question about the process-node target. Many valuable semiconductor products do not require the most advanced nodes, and specialized analog, power, sensor, RF, research and defense devices can remain commercially relevant on mature processes. That creates a potentially large addressable market for flexible smaller fabs. But if Fab2 tries to compete aggressively at leading-edge logic, capital requirements, lithography complexity and process integration difficulty could increase dramatically. Public materials reviewed do not provide enough information to assign Fab2 a verified production node roadmap or commercial wafer-volume target.
Customer disclosure is similarly limited. Fab2 has attracted major investor interest and is hiring aggressively, but public materials reviewed do not provide a detailed customer list, contracted fab orders, recognized revenue, gross margins or unit economics for a completed commercial fab. The SEC filing declines to disclose revenue range. That absence does not mean the company lacks revenue; it means investors cannot use Form D to measure revenue against a $3.7 billion valuation. The most important next evidence would be repeat fab deployments, disclosed production customers, uptime metrics, wafer or device throughput and economics showing that distributed small fabs can compete where the company claims they can.
FINAL ASSESSMENT
Fab2's September 2026 Form D confirms that Atomic Semi has evolved into a very different-scale semiconductor company. The legal continuity is clear: Fab2, Inc. lists Atomic Semi, Inc. as its previous name, remains a Delaware corporation and is now headquartered in Austin under Sam Zeloof's leadership with Jim Keller and Corey Nobile on the board. The filing reports $587.997 million offered, $556.021 million sold to 89 investors and $31.976 million remaining under Rule 506(b).
The most important independent finding is the $88 million conversion disclosure. Fab2 publicly announced a $500 million Series A at a $3.7 billion valuation, while the SEC amount sold is higher. Rather than treating one source as wrong, the Form D itself provides the reconciliation: $88 million of reported amount sold includes conversion of securities. That means the SEC total describes the broader securities transaction and should not be represented as $556 million of entirely new Series A cash.
The second independent finding is that Fab2's capital intensity follows directly from the business model. This is not merely a fabless chip company and not simply another foundry. Fab2 is attempting to design and manufacture semiconductor tools, integrate them into compact software-defined fabs, use its own EDA and automation stack, and ultimately manufacture the fabs themselves in volume. The financing therefore represents a bet on industrial replication: whether a semiconductor fab can become something closer to a standardized manufactured product instead of a one-off multibillion-dollar megaproject.
That thesis has potentially significant implications for U.S. semiconductor resilience because smaller distributed fabs could reduce dependence on a small number of gigantic manufacturing campuses for certain classes of chips. But it also concentrates extraordinary execution risk inside Fab2. Investors should demand evidence on process capability, yields, customer orders, fab installation cost, throughput, uptime, equipment lifetime, field-service requirements, gross margins and the amount of additional capital required to reach large-scale deployment. At a publicly announced $3.7 billion valuation, proving that a small fab can work is no longer enough; Fab2 must prove that it can repeatedly manufacture fabs that themselves manufacture commercially useful chips.
Form D is an exempt-offering notice. It is not SEC approval of Fab2, Atomic Semi, the $3.7 billion valuation, the Series A terms, the semiconductor technology or any projected commercial performance.
SEC SNAPSHOT
ISSUER: Fab2, Inc. | PREVIOUS NAME: Atomic Semi, Inc. | CIK: 0002151752 | SEC FILE NO.: 021-596715 | FILM NO.: 261365600 | ACCESSION NO.: 0002151752-26-000002 | FILED / EFFECTIVE: September 8, 2026
ENTITY: Delaware Corporation | FORMATION YEAR: 2022 | PRINCIPAL ADDRESS: 6301 Stassney Lane, Building 8, Austin, TX 78744 | PHONE: 510-250-7224
INDUSTRY: Other Technology | EXEMPTION: Regulation D Rule 506(b) | POOLED INVESTMENT FUND: No | INVESTMENT COMPANY ACT EXCLUSION: None claimed
SECURITIES: Equity | Security to Be Acquired Upon Exercise of Option, Warrant or Other Right to Acquire Security | BUSINESS COMBINATION: No
FIRST SALE: August 21, 2026 | OFFERING DURATION: One year or less
TOTAL OFFERING: $587,997,030 | AMOUNT SOLD: $556,021,222 | REMAINING: $31,975,808 | INVESTORS: 89 | MINIMUM INVESTMENT FIELD: $0
CRITICAL SEC CLARIFICATION: $88,000,000 of Total Amount Sold includes conversion of securities.
SALES COMMISSIONS: $0 | FINDER FEES: $0 | ITEM 16 RELATED-PERSON PAYMENTS: $0 | REVENUE RANGE: Declined to disclose
RELATED PERSONS: Samuel Zeloof — Executive Officer / Director | James Keller — Director | Corey Nobile — Director
FORM D SIGNATORY: Samuel Zeloof | TITLE: President and Chief Executive Officer
PUBLIC SERIES A: Fab2 publicly announced a $500M Series A at a $3.7B valuation in September 2026.
IMPORTANT FINANCING DISTINCTION: The SEC reports $556.021M sold, but specifically says $88M of that amount represents conversion of securities. The Form D amount sold therefore should not be described as $556M of entirely new cash financing.
EARLIER COMPANY IDENTITY: Atomic Semi | FOUNDERS: Sam Zeloof and Jim Keller | EARLY REPORTED FINANCING: approximately $15M seed financing in 2023, widely reported as led by OpenAI Startup Fund at an approximately $100M valuation.
CURRENT BUSINESS MODEL: Fab2 designs semiconductor manufacturing equipment and software, assembles the equipment into small software-defined fabs, and aims to mass-produce the fabs themselves.
EQUIPMENT / SYSTEMS PUBLICLY DESCRIBED: pumps | valves | gas delivery | vacuum chambers | lithography systems | process equipment | automation | semiconductor fab infrastructure.
SOFTWARE: Studio, formerly Atomic Studio | browser-based collaborative EDA for schematic, layout and simulation workflows.
OPERATING FOOTPRINT: Austin, Texas | Lockhart-area, Texas | legacy San Francisco, California operation. Public square-footage claims should be distinguished from currently visible individual construction-permit phases.
KEY TECHNOLOGY TRADEOFF: Fab2's maskless / electron-beam-oriented approach can improve flexibility and prototype turnaround but has materially different throughput economics from high-volume optical and EUV lithography.
PUBLICLY IDENTIFIED SERIES A INVESTORS INCLUDE: Fundomo | Corner Capital | Paradigm | Duquesne Family Office | Maverick Silicon | StepStone Group | Protagonist | UDC Ventures | Naval Ravikant | other investors.
PUBLIC SERIES A VALUATION: $3.7B. THIS IS A COMPANY-ANNOUNCED FINANCING VALUATION AND NOT AN SEC DETERMINATION OF FAIR VALUE.
CORE INDEPENDENT FINDING: Fab2's 2026 Form D tells a deeper story than the public $500M Series A headline. SEC records show $556.021M sold because the capitalization event includes $88M of converted securities, while the company's rebrand from Atomic Semi to Fab2 reflects an equally important strategic transition: it is no longer merely trying to manufacture chips faster, but to industrialize and replicate the semiconductor fabs and equipment that manufacture those chips. The central investment question is therefore whether the company can turn an inventive small-fab concept into repeatable semiconductor manufacturing infrastructure at economically viable scale.
Form D is an exempt-offering notice and is not an SEC-issued certificate, approval or endorsement.