Last week, Microsoft (MSFT: NASDAQ) and Meta (META: NASDAQ) reminded us that the AI boom is, at this stage, a capital expenditure story. This week, we’ll find out who else can afford to play. Palantir (PLTR: NYSE)’s promises, Arista (ANET: NYSE)’s battle with Cisco (CSCO:NASDAQ), AMD (AMD: NASDAQ)’s margins, Supermicro (SMCI: NASDAQ)’s backlog and GlobalFoundries (GFS: NASDAQ)’s relevance all go under the microscope. In an industry defined by big talk and tight lead times, the earnings calendar has become a stress test—for both the companies and their investors.

And please check out our website https://epistrophy.beehiiv.com: it contains a searchable library of these notes and much of our research.

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

• GFS: The Turnaround Foundry?

• Tweet O’ The Week

• Epistrophy In The News

• 📆 of Epistrophy Events

• Availability This Week

GFS: The Turnaround Foundry?

GlobalFoundries (GFS: NASDAQ) reports second quarter earnings on August 5, and the story is less about what it has already delivered than what it still needs to prove. The foundry has positioned itself as the beneficiary of industrial policy, U.S. manufacturing ambitions and the domestic demand for “essential chips.” But a turnaround only works if it becomes self-sustaining—if margin, mix and momentum start reinforcing each other rather than requiring constant support.

This quarter will test whether that flywheel is beginning to spin. GlobalFoundries guided to $1.675 billion in Q2 revenue, up 6 percent sequentially. Margins are expected to improve modestly: 24.1% gross (25 percent non-IFRS), 10.8 percent operating (14 percent non-IFRS), and diluted earnings per share of $0.27 to $0.36 (yes for all the US boosterism, with headquarters in upstate New York and operations in the U.S., Europe and Asia -the company is officially incorporated in the Cayman Islands and therefore reports its financials under IFRS, not GAAP.)

That’s not breakneck growth, but it represents stabilization. After four quarters of uneven handset demand, utilization headwinds and wafer ASP pressure, the business appears to be narrowing its focus and tightening execution. 

If GlobalFoundries is going to complete a credible turnaround, it needs to get three things right:

1. Transitioning from underutilized mobile exposure to higher-value, long-cycle platforms.

Smartphones are shrinking as a percentage of GlobalFoundries’ revenue. That’s by necessity. The segment accounted for 44 percent of revenue a year ago; it was down to 37 percent last quarter, and falling. Revenue from smartphones dropped 21 percent sequentially and 14 percent year over year in Q1.

But that weakness is giving GlobalFoundries the room to reweight toward automotive, industrial IoT and communications infrastructure—sectors with longer design cycles, deeper customer relationships and less pricing volatility. The company says nearly 90 percent of design wins over the past four quarters were sole-sourced. Many are built on its differentiated 22FDX, 45SPCLO and 130BCD platforms, which serve radar, LiDAR, industrial vision, secure authentication and optical networking.

These aren’t speculative chips. Bosch is using GlobalFoundries silicon for its next-generation single-chip radar. Ayar Labs is using GF’s silicon photonics platform for chiplet-scale optical interconnects. And in April, GlobalFoundries announced new design wins across 77 GHz and 120 GHz radar SoCs, SatCom terminals, micro-LED displays and low-power medical DSPs.

Volume in these segments is still ramping. But the architecture is in place. To sustain the turnaround, GlobalFoundries will need to convert design win share into revenue share without the help of underutilization payments that previously propped up the model.

2. Controlling costs while scaling U.S. manufacturing—and proving CHIPS Act relevance without relying on it.

The company is midway through a $16 billion plan to expand domestic production. Over $7 billion has already been invested in New York, Vermont and Singapore. The rest is tied to CHIPS Act awards. The narrative—strategic reshoring, secure supply chains, non-China-aligned fab capacity—is working in GlobalFoundries’ favor, especially with U.S. and European defense contractors.

But for that to translate into financial leverage, the company must control capex burn, keep debt low and prove that customers will fund their own capacity. So far, it’s doing that. The balance sheet is clean: $3.7 billion in cash and marketable securities against just $1.1 billion in long-term debt. GlobalFoundries repaid $664 million last quarter. Free cash flow, while down year over year, remains positive.

Unlike rivals chasing hyperscale AI, GlobalFoundries has kept its investments aligned with process nodes where customers can’t afford risk or delay. That puts it in a stronger position to weather macro volatility—but it also means utilization and operating leverage must remain tight. Margin expansion in Q2 will be watched closely for signs that the model is scaling without subsidy.

3. Creating technical and commercial differentiation at the platform level, not just the process node.

GlobalFoundries won’t compete with Taiwan Semiconductor Manufacturing Co. (TSM: NYSE) on raw density. But it doesn’t need to. Its real edge is integration: RF plus logic, photonics plus packaging, embedded memory plus signal processing. These aren’t bleeding-edge, but they are hard to replicate.

The July announcement of a deal to acquire MIPS, the RISC-V IP vendor, signals that GF wants to move further up the stack. MIPS brings not only compute IP but also software tooling, including the Atlas Explorer design environment that enables “shift-left” optimization—power, performance, area—before physical design begins. If GlobalFoundries can integrate these tools into its silicon platforms, it may offer customers something that even larger foundries can’t: a process-validated, design-aware platform tailored to their application.

That has particular value in automotive, industrial and medical use cases where form factor, latency, and power efficiency matter more than raw speed. The company’s recent design wins reflect that positioning—targeting haptics, radar, imaging, audio, and secure edge inference rather than high-end GPU accelerators.

It’s too soon to call that a moat. But it’s a viable differentiator—one that becomes more defensible as customers commit to multi-year roadmaps and co-develop IP around GlobalFoundries’ mature but customizable technologies.

The way ahead

This quarter won’t resolve every question. Smartphone demand is still weak. ASPs are under pressure. The pace of CHIPS Act disbursements remains uncertain. And while design wins are encouraging, the revenue they promise has yet to fully show up in financials.

But for the first time in over a year, GlobalFoundries has a credible setup for forward momentum. The cost base is in control. The margin profile is stabilizing. And the customer mix is tilting toward durability, not dependence.

That doesn’t make GlobalFoundries a growth story. But it might make it a turnaround story worth watching next week.

Tweet O’ The Week

Epistrophy In The News

On Yahoo Finance, I joined Josh Lipton to break down Qualcomm (QCOM: NASDAQ)’s quarter and explain why the real story was in the modem roadmap and smartphone recovery. We also covered ARM (ARM: NASDAQ), and why its licensing model may insulate it from the more brutal swings of AI hardware demand. On NewsNation, Connell McShane and I unpacked Microsoft and Meta’s massive infrastructure bets, discussed whether the amazing billion-dollar-a-day spending by the five biggest names in AI.

📆 of Epistrophy Events

Availability This Week

I’m in San Francisco all week and making time for client check-ins, background calls, and coffee with people who have something to say and something to prove. If you’ve got a question about a company on this week’s docket—or just want to compare notes—I’m around.

Written reports are available to clients, with video summaries on YouTube, and of course our popular summaries of the summaries on Instagram, TikTok, and YouTubeShorts.

Are these notes are helpful to you? Suggestions? I’d love to discuss them further and, as always, comments, questions and ideas are appreciated. If you have a friend or even a frenemy whom you think might benefit from this note, have them reach out and I’ll put them on the list.