We’re deep in the belly of earnings season—Cisco Systems (CSCO: NASDAQ) and Applied Materials (AMAT: NASDAQ) close out a stretch that’s offered more spreadsheets than sleep. Cisco’s inventory bloat remains a focus. At AMAT, export controls and China dependencies loom over any upside. Both companies now face a bigger question: what happens to hardware demand when Trump’s 10% blanket tariff plan starts distorting supply chains again? The next set of clues lands this week.

As always, I’m focused on three things:
1) Technology-driven change;
2) the latest in innovation and startup trends, and;
3) stock fraud.

Companies Discussed

In This Note:

• Companies Discussed

• Switching Gears

• The Aussie Anti-Bank

• Tweet O’ The Week

• Epistrophy In The News

• 📆 of Epistrophy Events

• Availability This Week

Switching Gears

Cisco Systems is set to report its fiscal Q3 2025 earnings this week, following a second quarter that showcased both progress and persistent challenges. Revenue rose 9% year-over-year to $14.0 billion, largely driven by the integration of Splunk. That acquisition contributed observability and security revenue to offset weakness in Cisco’s traditional hardware lines. But strip Splunk out, and the company saw a modest decline—down 1% compared to Q2 last year. Cisco is growing, but selectively.

One area of strength has been AI networking gear. Cisco booked roughly $700 million in AI infrastructure orders in the first half of the fiscal year, with strong demand for high-throughput switches like the Nexus 9800, optical interconnects in the Silicon One family, and 800G-capable systems. On the software side, Cisco’s Duo Security and Umbrella platforms continued to expand, especially in zero-trust and DNS-layer protection. And Splunk, now integrated, anchors Cisco’s push into observability, offering real-time monitoring across applications and infrastructure. Together, these three categories now form the strategic spine of Cisco’s transformation: faster networks, safer networks, and visible networks.

Less visible, but no less important, is inventory. Cisco had been dogged by questions about rising hardware stockpiles, particularly as lead times improved and customers slowed deployments. But Q2 data shows a shift. Inventory dropped to $2.93 billion, the lowest in over a year. More importantly, Days of Inventory fell to 55—down from 61 in Q1 and 64 at the beginning of 2024. Cisco hasn’t just stabilized the supply chain; it’s shipping more product, faster, even as macro demand remains uneven. That’s a meaningful reversal from the overbuilds of 2022 and early 2023.

CEO Chuck Robbins deserves credit for steering Cisco away from hardware dependency without swinging too far into hype. His team has focused on long-cycle investment in AI fabrics, security integration, and cloud software, resisting the temptation to simply chase hyperscaler volume. The Splunk deal remains controversial—$28 billion is a high price—but the early signals suggest strategic fit. Cisco is now in rooms it didn’t used to be in: enterprise observability reviews, cloud security planning, multi-cloud compliance audits. It’s not leading those rooms yet. But it’s present, and that’s new.

The upcoming quarter won’t settle the bigger questions: Can Cisco permanently replace routing and switching revenue with software and AI infrastructure? Can Splunk scale inside Cisco without cultural erosion or sales overlap? And can the company grow through a tariff environment that may penalize Chinese-sourced components just as networking demand rebounds? But it can show whether Robbins’ long-game is beginning to pay dividends. Cisco doesn’t need a blowout. It just needs traction.

The Aussie Anti-Bank

The FT reported a few weeks back that $5.5 B  payments startup Airwallex is seeking banking licenses in the United Kingdom and the United States. This big step for the Australian-founded fintech company (currently headquartered in Singapore) could put the company on center stage and show how it’s building a bank that is better than a bank. 

In a world where you can stream 8K video in a jungle but can’t pay a supplier in Colombia without invoking five middlemen, Airwallex is doing something audacious: bypassing the 1970s. That’s the effective birth date of SWIFT, the system we still rely on for international payments—something closer to a telex than a transaction.

Airwallex, founded in Melbourne by Jack Zhang, Max Li, Lucy Liu, Xijing Dai, and Ki-lok Wong in 2015, built a workaround. A practical, quietly revolutionary architecture that lets businesses move money across borders, open foreign bank accounts, and transact in 20+ currencies—without ever touching the institutional sludge of correspondent banking. We know Ripple intimately (scar tissue!) and cover the sector by tracking Block (XYZ: NYSE) and PayPal (PYPL: NASDAQ) – Airwallex is well overdue.

The Infrastructure Gap

The geopolitical era of dollar dominance is shifting. More companies are building global customer bases, remote workforces, and supply chains spanning five time zones. Yet the infrastructure supporting business payments still relies on SWIFT.

The COVID pandemic only accelerated this realization. Digital-first businesses needed to spin up subsidiaries in Vietnam, contractors in Argentina, and pay platforms in Kenya—all without physically stepping foot in those countries. The tools didn't exist. Airwallex stepped into that void with a bold premise: global financial infrastructure should work like cloud software.

And that idea is resonating. Airwallex recently acquired MexPago in Mexico and entered Brazil through its acquisition of payments company UniPay. These aren’t vanity expansions. They reflect a deeper strategic bet: that Latin America’s inefficiencies are a fintech gold mine.

What Airwallex Avoids

To understand why Airwallex matters, you have to understand how bad the status quo is. SWIFT isn’t a payments network. It’s a messaging protocol—a way for banks to send instructions to each other to move money. But those messages are only part of the story. A typical international transfer might go through six financial institutions. A business in Arequipa, Peru tells its local bank to send money to a supplier in Perth, Australia. That bank doesn’t have a direct relationship with the Australian bank, so it routes the payment through a correspondent in Lima to another in Mexico City, to another in New York to another in Melbourne before finally getting to the recipient in Peru. Each hop adds time, fees, and a fresh layer of compliance friction. That’s not theoretical—those delays are measured in days, not minutes.

Now compare that to what Airwallex built. Instead of sending instructions to a bunch of third-party banks, Airwallex built direct links. It holds accounts in each country it operates, uses local payment rails, and nets out transactions across its own ledger.  The result is near-instant settlement. It doesn’t move money, it relabels money. 

The Airwallex stack includes:

• Local payment rails in each country to avoid SWIFT altogether.
  

• Virtual multi-currency wallets that let businesses hold, receive, and convert funds across 20+ currencies at interbank rates.
  

• Banking infrastructure-as-a-service APIs, allowing clients to embed international payments, issue corporate cards, and manage expenses in their own apps.

The analogy I keep returning to: It’s like Stripe and Revolut had a baby—with Stripe’s developer DNA and Revolut’s regulatory ambition—but raised by an operations manager who cares about latency and cost-per-transaction.

What Airwallex really replaces is the need for banks to trust one another at a distance. That’s what SWIFT was designed to enable in the 1970s: trustless, paper-based instructions between institutions that didn’t share infrastructure. But we live in a world where distributed ledger systems, treasury APIs, and real-time payment systems exist in nearly every major economy, to the advantage of Aiwallex.

Americans Can’t Get It

American companies don’t understand international payments.

It’s not their fault. The U.S. has had functioning ACH since the 1970s, instant domestic transfers via Zelle, payments through Venmo and low-friction credit networks like Mastercard (MA: NYSE) and Visa (V: NYSE) that have been around for decades. Most American business owners never think twice about how long it takes to move money across state lines.

But send money to Indonesia? Or pay a vendor in Nigeria? You’ll encounter fees, delays, bank errors, and inexplicable compliance hurdles. Try issuing refunds or managing treasury balances in three currencies without losing 6% to conversion spreads? Good luck.

It’s the real pain point for tens of millions of entrepreneurs across Asia, Latin America, and Africa. And Airwallex is building for them—because they’re building for everywhere, not just the U.S.

Competitive Landscape

The private Airwallex last funding round in October 2022  reportedly  valued the company $5.5 billion. Airwallex was said to have processed over $50 billion in annualized transactions in 2022, up from $20 billion in 2021. It counts Tencent, Square Peg, and Sequoia among its backers.

Looking Ahead

When the IPO window someday reopens, we’d expect Airwallex to be at the front of the line of fintechs looking to cash in with the US markets. And with trailing revenues likely $1 B, that could be a deal valued at $7-10B. But the tsunami of uncertainty from the White House means we won’t see calm enough waters for new offerings anytime soon. 

That said, we predict three things in the next 12–24 months:

1. Deeper embedded finance integrations, especially with global SaaS platforms and B2B marketplaces.
   

2. Licensing pushes in Europe and Southeast Asia, where regulators are warming to nonbank infrastructure.
   

3. A potential IPO, likely following expansion in U.S. enterprise sales or a major strategic partnership.
   

The biggest risk? Friction cuts both ways. As Airwallex expands into more regions, its ability to meet Know Your Customer (KYC) and anti-money laundering (AML) compliance requirements will be tested. Regulators in the U.S., EU, and Asia are increasingly scrutinizing cross-border financial infrastructure for vulnerabilities that could be exploited by bad actors. Unlike traditional banks with decades of embedded compliance teams, fintechs like Airwallex must constantly prove the integrity of their risk models and onboarding systems. Any lapse—real or perceived—could result in fines, restrictions, or reputational harm, particularly as geopolitical tensions make regulators more hawkish.

Getting too ambitious is also a concern. If Airwallex tries to be all things—card issuer, FX provider, expense manager, lending platform—it risks diluting the very edge that makes it compelling: speed, simplicity, and focus.

But for now, it’s one of the most technically sound, strategically disciplined, and globally aware fintechs in the market. And it’s quietly replacing infrastructure that everyone hates but no one notices—until it breaks.

Tweet O’ The Week

Epistrophy In The News

On Schwab Network, I broke down why Onsemi (ON: NASDAQ)’s apparent beat disguised a miss—auto revenue’s in freefall, and the so-called “transition to silicon carbide” sounds more like a stall. Over on Bloomberg BNN, I offered live commentary on AMD (AMD: NASDAQ)’s earnings call, where the real story wasn’t margins or MI300 units, but how Lisa Su sidestepped demand questions with the finesse of someone who’s played this game before.

📆 of Epistrophy Events

Availability This Week

I’ll be in Seattle on Monday delivering a lecture at the University of Washington’s Foster Business School titled “Fakes, Fads and Frauds: How To Detect Stock Fraud.” If you’re in town, come by—unless you're the kind of person who gets defensive about SPACs.

The rest of the week I’m in San Francisco. I hope these notes are useful; I’d love to talk more. Comments, corrections, and conspiracies are always welcome. And if you have a friend—or a frenemy—who might enjoy these, send them my way. Written reports are available to clients, with video summaries on YouTube, and of course our popular summaries of the summaries on Instagram, TikTok, and YouTube Shorts.