Opportuna Newsletter #11 | Sep-25

After a packed summer in the Bay Area, we are back in Zurich, focused on sourcing new transactions. Direct secondaries remain active, even as IPOs return. Our portfolio is developing well, with mark-ups and early signs of distribution, validating our approach.

Our last edition on stablecoins struck a chord. Since then, OpenAI and Stripe have released the Agentic Commerce Protocol, underscoring momentum in programmatic money. With Mesh Connect in our portfolio, we are positioned to benefit. If you want to know more about other opportunities we are working on, contact us here.

This edition remains focused on the convergence of private and public tech investing.

The Chart of the Month maps out the developer tooling for the new era of Software building.

In Current Topics, we dive into US firms’ earnings transcripts to understand what they are really saying about AI. Our analysis suggests: 1) AI remains mostly a technology-sector theme; 2) Outside of utilities and tech, it shows up as a cost-saving tool more than a revenue driver; 3) Workforce reduction is surfacing even in technology. For investors, this points to AI as a force that may expand both margins and revenue in tech—a potent mix for valuation.

Our Long-Term Theme dives into Quantum Computing. The domain is inflecting and drawing in retail investors. Prices and placements prove it. But the bridge from physics to profit runs through logical qubits, scarce capacity, credible roadmaps, and real-world commercialisation. The market is shifting from public spending reliance towards tapping companies IT budgets. We argue that winners will be those who start with hybrid approaches; Pasqal’s neutral-atom deployments into EU/MENA HPC centers are examples of this hybrid path.

You can see past editions here. Anybody can sign up following this link.

🚨 Highlights of the Month

 OpenAI and Stripe have released the Agentic Commerce Protocol, underscoring momentum in programmatic money. It is clear that internet purchasing modalities are going to change a lot over the next three years! We are excited for the team at Mesh Connect, one of our portfolio company.

IPO window: back open, but with higher bars. Post–Labor Day is historically a favorable listing window, and 2025 is no exception. PitchBook’s analyst note says the US IPO backdrop is stronger than earlier in the year but still unlikely to produce a surge; standards have tightened vs. 2021-era deals. Through June 30 there were just 18 US VC-backed IPOs, reinforcing a “trickle, not flood” setup. For issuers: prioritize scale, profitability progress, and governance. For buyers: expect pricing powerIPO window: back open, but with higher bars.

Venture debt is being securitized by GPUs. Lenders are writing asset-backed facilities against Nvidia-class hardware as AI infra demand outstrips equity appetite. PitchBook reports growing skepticism among some lenders even as volume rises, and separate data show AI/ML startups taking ~38% of US/EU venture-debt dollars YTD.

Capital concentration is extreme—and AI-skewed. PitchBook tallies 41% of all 2025 VC dollars going to just 10 startups (eight are AI), roughly $81.3B of $197.2B deployed YTD.

📈 Chart of the Month: Developer Tooling for Software 3.0 (BVP)

🌐 Current Topics: What Are Companies Saying About AI?

AI has dominated headlines this year, but questions remain about its economic payoff. An MIT Survey claimed “95% of organizations are getting zero return”. In contrast, Google reported that three-quarters of IT leaders see measurable ROI from generative AI. Both surveys come with biases, so we dug into something harder to spin: company earnings calls.

We analyzed 345 US-listed firms reporting between July 14 and September 18. Using keyword analysis, we measured “AI airtime” in transcripts. Technology unsurprisingly led, devoting 35% of its discussion to AI. Other sectors gave it less than 10%. The takeaway: despite the noise, investors outside tech don’t yet treat AI as material.

Next, we classified AI mentions as either revenue opportunities or cost-saving initiatives. Growth plays excite investors more than efficiency gains, which are finite. Yet only technology and utilities leaned toward revenue. Everywhere else, AI was cast as a cost-cutting tool. The worry that AI adoption will shrink jobs is grounded in this: most industries view it as a labor arbitrage lever, not a growth engine.

Finally, we looked at headcount. We measured “hard reduction” mentions within workforce commentary. Surprisingly, technology ranked second highest after basic materials. For a sector long associated with hiring sprees, this is unusual. Other industries clustered around 20%. Taken together, the data suggests a broad cooling in job creation—and a pivot to efficiency—even in growth sectors.

Our analysis leaves three conclusions. First, AI remains mostly a technology-sector theme. Second, outside of utilities and tech, it shows up more as a cost-saving tool than a revenue driver. Third, workforce reduction is surfacing even in technology. For investors, this points to AI as a force that may expand both margins and revenue in tech—a potent mix for valuation.

🧭LT: Hype, Hope and Hard Physics in Quantum Computing

Since summer, quantum is back in fashion. Retail chatter and sharp moves in IonQ, Rigetti, and D-Wave signal an inflection. Capital has followed: IonQ priced a $1 billion raise; Rigetti and D-Wave completed $350 million and $400 million ATMs. Excitement is real. Proof is thinner.

A qubit’s appeal is combinatorics. A classical bit is a coin lying flat—heads or tails. A qubit is a coin spinning in the air—both states at once until measured. Two qubits encode four possibilities at once; three, eight; fifty, more than a quadrillion. That parallelism is why quantum could search vast spaces faster than classical computers. It is also why errors kill value so quickly.

Not all qubits are built alike—just as “bits” once lived in vacuum tubes before the transistor won. Today’s contenders: superconducting circuits (IBM, Google), trapped ions (Quantinuum, IonQ), neutral atoms (Pasqal, QuEra), photonic qubits (PsiQuantum, Xanadu), and spin qubits (Intel). Each trades speed, stability, and scalability differently. No architecture has “won” yet.

Physics is cruel. Qubits decohere with noise, heat, and time. Fault tolerance solves this by stitching many imperfect, physical qubits into one reliable logical qubit. Logical qubits give enterprises what they buy: repeatability, auditability, and unit economics, plus a stable software target. Watch three milestones: (1) logical-qubit counts rising from single digits to hundreds; (2) falling physical-to-logical overhead; (3) credible roadmaps to fault-tolerant subsystems.

Money, for now, is concentrated in the public sector. Quantum-computing revenue totaled $650–750 million in 2024, expected to surpass $1 billion in 2025. McKinsey pegs the 2024 “market size” (revenues + investments + internal big-tech funding) at ~$4 billion, rising to $16–37 billion by 2030. Pasqal’s neutral-atom systems illustrate the near-term path: integrate quantum as an accelerator inside classical/HPC workflows. Deployments span EU and Middle Eastern supercomputing centers; pilots include Crédit Agricole (credit-risk modeling) and EDF (EV-charging optimization).

The domain is inflecting and drawing in retail. Prices and placements prove it. But the bridge from physics to profit runs through logical qubits, scarce capacity, and credible roadmaps. Treat 100–300 logical qubits as the first ROI band—and ~1,000 logical as the policy and market watershed. For investors, the question is simple but brutal: are you buying the future of computing—or subsidizing physics research?

📌 Conclusion

AI, quantum, and secondaries all point to the same theme: convergence is accelerating. As capital shifts and computing evolves, opportunities open up where public and private markets meet.

At Opportuna, we are tracking this closely—and backing companies at the forefront. If you’re in Zurich or beyond and want to exchange perspectives, reach out.

Warmest regards,
The Opportuna Team