Quantum Upside for $0.24/Share

🌾 Welcome to StableBread's Newsletter!

Most of the internet relies on public-key encryption to secure bank logins, private messages, and software updates. But a large enough quantum computer breaks that entire security layer.

The most-cited expert survey, the Global Risk Institute's Quantum Threat Timeline, puts the odds of such a machine arriving within a decade at 28-49%, and IBM has committed to delivering its first large-scale fault-tolerant quantum computer by 2029.

One company addressing that threat while also running an established chip business is SEALSQ (Nasdaq: LAES). The Swiss company designs, produces, and sells semiconductors, PKI services (systems that issue and manage the digital IDs devices use to prove they're genuine), and post-quantum technology products.

In FY2025, the company lost $34.2M on $18.3M of revenue, down 39% from its FY2023 peak. But the company holds a massive cash pile worth near $500M against only ~$8M of debt. That works out to net cash of $2.19/share, with the stock trading just 11% above it at $2.43, though the pile came at the cost of diluting existing shareholders 14-fold.

Regardless, you're paying $0.24/share ($2.43 minus $2.19 of net cash) for the entire business, which includes the QS7001 post-quantum chip now in certification and a $200M quantum investment portfolio.

The question is whether SEALSQ keeps diluting shareholders to fund an affiliated quantum empire, or whether the quantum bet and the post-quantum chip ramp pay off instead. Below, I cover the business, the quantum story, and whether it's a bet I'm taking.

SEALSQ is a fabless chip company, meaning it designs chips but owns no factory. It hands the manufacturing to outside foundries, mainly TSMC.

SEALSQ was carved out of WISeKey, a Swiss cybersecurity group that's still its controlling shareholder, and listed on Nasdaq in 2023.

The core product is a chip that stores a device's secret keys and identity, built so they cannot be copied or extracted. That lets a smart meter, a drone, or a payment reader prove it's genuine and encrypt what it sends.

SEALSQ sells these chips under the VaultIC name, alongside a line of secure microcontrollers—small processors with that same key protection built in. Cisco, the US networking-equipment maker, has bought SEALSQ's chips since 2014.

Two services are offered with the chips:

In August 2025 SEALSQ bought IC'Alps, a French chip-design house, for roughly EUR 12.5M. IC'Alps builds ASICs, custom chips made to one customer's exact specification, for regulated markets like medical, automotive, space, and defense.

So the business today is three parts: (1) secure chips, (2) PKI and provisioning, and (3) custom ASIC design. The entire business made $18.3M in FY2025.

A fourth part came online this year. Quantix Edge Security, a chip test-and-personalization joint venture in Murcia, Spain backed by €19.6M from the Spanish government, generated its first revenue in H1 2026, though the company hasn't disclosed how much.

SEALSQ's quantum story, including its quantum-computer stakes, satellite venture, and machine-payments token, is everything else the company attaches its name to. But despite all of the media and investor hype, none of that generates revenue yet.

One reason I'm particularly interested in an $18M chip company is a shift that regulators are now mandating.

Today's internet encryption (RSA and elliptic-curve, the cryptography behind the padlock in your browser) can in theory be broken by a large enough quantum computer. In practice that would mean whoever owns the machine could read intercepted bank traffic, government files and private messages, and forge the digital signatures that tell your device a software update is real.

Post-quantum cryptography, or PQC, is a new set of algorithms built to resist that. In August 2024 the U.S. standards body NIST finalized the first three, named ML-KEM, ML-DSA and SLH-DSA. Chipmakers will not build a chip around an algorithm that's still being revised, so final standards are what opened the door to post-quantum chips.

The near-term driver is not that a quantum computer exists yet. It's "harvest now, decrypt later," the risk that an adversary records encrypted data today and decrypts it once a capable machine arrives.

CISA, the NSA and NIST have all published joint guidance telling organizations to start migrating now, and they name that risk as the reason. That’s the practical reason a smart-meter or defense customer starts moving years before any quantum computer exists.

SEALSQ's answer is the Quantum Shield QS7001, a security chip that runs the NIST algorithms in dedicated circuits instead of in software on a general-purpose processor.

The same hardware also ships as the QVault TPM, SEALSQ's post-quantum take on the TPM (trusted platform module), the standardized security chip most PCs and servers already use to hold their keys.

The new algorithms take more computing work than the ones they replace, and a smart meter or sensor running off a coin cell (a watch battery) cannot spare the extra time or power.

Peer-reviewed work on dedicated Kyber and Dilithium circuits, the algorithms behind ML-KEM and ML-DSA, reports speedups between 3x and 9x over software. The specific "10x faster" figure is SEALSQ's own, from its September 2025 launch release, and I couldn't find it verified anywhere outside the company.

That's not the only claim worth a second look. Two others come up repeatedly in SEALSQ's press releases and investor decks, and both are not entirely true.

The first is the company's line that it's "one of the only 6 semiconductors companies in the world" able to develop certified secure microcontrollers. The certification behind that claim, Common Criteria EAL5+, is an international rating where an accredited lab reviews a chip's design and tests how well it resists physical tampering, and it's the normal bar for this product class rather than an unusual one.

But when I went through the Common Criteria certified-products list, the public registry where every one of these certificates is published, I counted 354 live EAL5+ certificates in the smart-card and secure-IC category alone. Most belong to card and passport makers, but at least a dozen of the holders design the chips themselves.

Even counting only makers of certified secure microcontrollers, the registry shows at least seven in the past five years: NXP, Infineon, STMicroelectronics, Samsung, China's Tongxin and CEC Huada, and SEALSQ itself.

Clearing EAL5+ still takes years and an accredited lab, so the club is small, just not as exclusive as the company says. It's also possible the six is simply an outdated claim on SEALSQ's website that predates the newer certificates.

The second is a "2027 U.S. government deadline." No such deadline exists. The 2027 date comes from the NSA's CNSA 2.0 schedule, which covers national-security systems only, and it marks when those systems should start preferring the post-quantum algorithms over today's RSA and elliptic-curve ones in new purchases. It's not a switch-off date for anyone else.

The real deadlines are further out, but they are now binding. The government-wide target was 2035, set by national security memorandum NSM-10, and in June 2026 an executive order pulled it closer. Federal agencies must move their most critical systems to post-quantum encryption by the end of 2030 and post-quantum signatures by the end of 2031, with a matching rule for federal contractors now being written.

The regulatory push behind post-quantum is real and funded. However, the urgency is often overstated, so it's worth pricing the stock on the mandate, not on a countdown.

To be clear, a code-breaking quantum computer is still not imminent. The 28-49% odds from the introduction are the highest reading in that survey's history, but even the top of that range is under 50%, so a machine arriving this decade is still less likely than not.

Here's the hardware evidence, for and against quantum:

For SEALSQ, what matters is that the odds are high enough that a customer protecting 20-year secrets has to act now, and the mandates above force buyers to move regardless of which side of the hardware debate proves right.

SEALSQ talks constantly about its pipeline. As of mid-2026 it expects more than $225M in opportunities through 2029, with more than $60M tied to the QS7001 and QVault chips. On the year-end call three months earlier the same figure was $200M through 2029.

None of that's backlog. A pipeline is the company's own estimate of what it might win. Backlog is signed orders a customer has committed to buy, and it's obviously the more meaningful number.

CFO John O'Hara laid out how the pipeline is built on the September 2025 call. The company identifies opportunities, applies "a relatively low success %," weights them, and looks three years out. On the year-end call he cautioned that "this is a management estimate and is subject to conversion risk, customer validation, timelines, and the certification process."

Here’s the part the pipeline talk buries. The QS7001 earns almost nothing today.

The 20-F, SEALSQ's annual report, says "revenue generation is not expected until Q4 2026," with first production revenue in H2 2026, gated on certifications that only finish late in the year. The chip the entire quantum thesis rests on is still pre-revenue.

It carries ordinary hardware risk too. The filing warns that a "remask," a chip respin, could cost a six-to-nine-month delay, and regulated buyers wait for certification before ordering volume.

What you should track here is the certification schedule, and so far it's holding.

SERMA, the accredited French lab SEALSQ hired to run the evaluation, confirmed in April 2026 that the QS7001 passed fault-injection and side-channel testing, the hardest physical checks in the process. Since SEALSQ commissions and pays for the work, read that as a certification milestone the company reported, not an outside endorsement. Even so, it's how Common Criteria operates industry-wide, with the vendor engaging an accredited lab and a government body reviewing the result.

The chip has since cleared a NIST entropy validation, and full EAL5+ and FIPS 140-3, the U.S. standard for validating a cryptographic chip, are targeted through the end of 2026.

So value the chip business on what it sells today. Secure chips, ASIC and PKI form the working business, guided to $27-37M of revenue this year and clawing back toward the old $30.1M peak, and the post-quantum layer adds nothing to that until late 2026 at the earliest.

It's worth mentioning here that the ten largest customers are 63% of revenue, and per the 20-F, "it is not industry practice to work with long-term contracts." Customers sign a framework and buy on purchase orders, with no volume commitment.

The way to evaluate that risk is customer stickiness, whether the same names come back year after year, and the evidence leans positive. Cisco has bought SEALSQ chips for 11 years, Asia Pacific revenue nearly doubled in FY2025, and SEALSQ logged eight design wins in Matter devices (the smart-home standard backed by Apple, Google and Amazon), meaning products that chose its chip.

If one name keeps climbing as a percentage of sales over the next few quarters, consider that an increasing risk, as there's no contract that locks in churned customers if something were to go wrong.

The management pitch is a "vertical quantum stack," from secure silicon up through quantum computers and satellites, that SEALSQ says nobody else has.

Through an internal "Quantum Fund" now sized at $200M, with more than $65M deployed, SEALSQ has taken stakes in a set of early-stage names. None of the minority stakes pays SEALSQ anything today, and the two companies it took control of in June 2026 only start counting toward its financials in H2 2026:

On SEALSQ's last audited balance sheet, from December 2025, the stakes it then held carried for ~$16M.

Do the five make strategic sense? They cover four different ways of building a quantum computer plus the software layer, and each deal wires SEALSQ's security chips into the partner's platform, so there's a logic beyond collecting logos.

But spending shareholder cash on minority stakes in pre-revenue startups is questionable venture investing, not chip building. I'd judge the program by whether more agreements like Quobly's $5M follow.

Two of the headline names management and investors like to discuss, WISeSat and SEALCOIN, are not SEALSQ companies at all. Both belong to the parent WISeKey.

WISeSat is the satellite arm, with 21 small satellites launched as of March 2026 as rideshare payloads on SpaceX rockets with SEALSQ's post-quantum chips aboard, 14 of them active today, and a 100-satellite constellation planned by 2027. SEALSQ paid $10M for an 8% stake in June 2025, so it would only capture a thin slice of any WISeSat success.

SEALCOIN is a cryptocurrency token built to let machines pay each other, a delivery drone paying a charging station on its own, with no human in the loop. It belongs to SEALCOIN AG, a WISeKey subsidiary, and SEALSQ discloses no stake in it.

The newest venture is a blank-check merger, and it makes Quantisimo SEALSQ's second quantum vehicle after the fund. In June 2026 SEALSQ and WISeKey formed Quantisimo Corp and signed a non-binding letter of intent with GigCapital8, a listed shell company holding $253M in trust that raises money from investors first and buys a business later.

The plan is to move "selected assets" from the quantum portfolio into Quantisimo at a $575M pre-money value, with talk of reaching $2B through five more acquisitions and a closing penciled in for Q1 2027.

GigCapital8's founder says investors have few pure quantum vehicles to buy. A dedicated listing could put a market price on the stakes that today are buried inside SEALSQ, and outside investors would fund further deals instead of SEALSQ's own treasury. None of it's binding, and what SEALSQ shareholders would actually receive has not been disclosed.

I would treat the minority stakes, the acquired companies, and the blank-check vehicle as something you should not pay up for. There's no quantum revenue in the stack yet and no signed volume contract beyond the $5M Quobly agreement, only investments that could pay off one day.

SEALSQ generated $30.1M of revenue in FY2023. Then shrunk to $11.0M in FY2024, a 63% drop.

Management's explanation is that customers held back orders while they waited for the post-quantum chips and worked through stock they had over-ordered during the 2021-22 shortage. The industry data backs part of that. FY2024 was brutal for microcontroller makers even while AI chips boomed, with STMicroelectronics' microcontroller segment down 39% and Microchip down 42% in its fiscal 2025.

The 20-F shows the rest, actual customer churn. A contract manufacturer that was 15% of FY2023 revenue fell to 4% in FY2024, and a digital-security company that was 12% went to zero and never came back. Those two were $8.1M of FY2023 revenue (15% plus 12% of $30.1M) and almost none of FY2024's, a big part of why SEALSQ fell harder than its industry.

Revenue recovered to $18.3M in FY2025, up 66%. But $3.6M of it was IC'Alps revenue for the five months after the deal closed, so part of the "growth" was bought, not grown.

Even at $18.3M the business is still 39% below its FY2023 level ($18.3M against $30.1M).

Gross margin held up better than revenue: 46.7% in FY2023, 34.0% in FY2024 as volumes fell, back to 47.3% in FY2025. IC'Alps (20% of FY2025 revenue) helps here too, because design work carries a near-88% gross margin against ~37% on the chips, so the blended number gets a lift from the mix.

Management guides FY2026 to grow 50-100% off the FY2025 base of $18.3M, which works out to $27-37M.

H1 2026 came in at ~$11M, up 120% y/y, but still 26% below the $14.8M SEALSQ did in H1 2023, when the old chip business ran at full strength.

Regardless, revenue is recovering, but the losses are worth a closer look.

FY2025 revenue was $18.3M at a 47.3% gross margin, so gross profit was $8.6M. Operating expenses were $48.4M. The operating loss was $39.8M and the net loss $34.2M.

The loss looks worse than the cash reality. The single largest line was an $11.3M non-cash charge for stock-based compensation to employees, board members and advisors.

Against $18.3M of revenue, $11.3M in SBC is quite significant (62%). However, FY2024's charge was only $0.1M, and most of the FY2025 grant was a one-time award booked in the first half, so I expect SBC to fall back toward zero rather than repeating in the future.

Interest on the cash pile brought in $6.1M, which is why the net loss of $34.2M is smaller than the $39.8M operating loss.

Back out the non-cash stock charge and the operating loss is $28.5M ($39.8M less $11.3M). That's what it costs to run SEALSQ for a year.

The number that matters is operating cash flow, which was negative $31.3M in FY2025. That's the rate at which the business consumed cash before any investing, and it's what the giant cash balance funds.

At the end of 2025 the company held $427.7M in cash and short-term investments. By June 2026 that was ~$495M. Debt is only ~$8M including lease obligations, with the borrowings being small French government loans inside IC'Alps.

So net cash is ~$487M ($495M less $8M). Against 222.8M shares that's $2.19 a share. The stock is $2.43. About 90% of the market value is the cash, and you pay $0.24 a share for everything else ($2.43 less $2.19), which is where the asymmetry lies.

At a ~$31M burn, ~$495M funds the business for many years. What you should evaluate instead is where the cash goes, because it fell ~$30M in Q2 2026 on acquisitions and minority bets, not on operations.

Dilution is the primary reason the stock has been driven down to $2.43, so you should understand it before evaluating SEALSQ any further.