CAPEX MAXIMUS, EARNINGS MINIMUS

(INTC), (AMD), (AAPL), (NVDA), (TSM), (QCOM)

Wall Street is suddenly in love with Intel (INTC) again. The stock’s up, the demand narrative is back, and Pat Gelsinger is everywhere from Capitol Hill to CES. 

But squint at the balance sheet and you’ll see something else: shrinking margins, rising capex, and a business model that’s acting more like a utility than a tech titan.

Sure, I can see the growth, but it’s what I’m not seeing that worries me: the operating leverage. Call me cynical, but I like my comebacks with gross margins north of 50%.

Intel’s Q4 results were, in many ways, exactly what you’d expect from a company trying to straddle two worlds: the legacy of dominant x86 CPUs and the capital intensity of a foundry business that doesn’t yet scale. 

Revenue came in at $15.4 billion, up 10% year-over-year. AI and server-side demand gave the Data Center and AI segment a much-needed lift, rising 10%, while the Network and Edge group dropped 24%.

Margins, though, are still the Achilles’ heel. Gross margin was just 38.3% GAAP, 44.5% non-GAAP. That’s up a hair from last year, but still a long way off the mid-50s range investors remember fondly. 

Foundry ambitions are expensive. Intel spent over $27 billion in capex last year, and it’s not done yet. The company is counting on CHIPS Act subsidies, co-investments, and economies of scale to eventually bring those margins back. 

But there’s a difference between vision and execution. For now, we’re still firmly in the burn phase.

Client Computing did better than expected, up 33% year-over-year, as the PC market shows signs of stabilization. But it’s worth noting that a good chunk of that gain is cycling off a dismal 2022 base. 

And even here, pricing pressure remains an issue. ASPs are down, and Intel is having to fight harder to defend share against AMD (AMD) and Apple (AAPL) with its custom M-series silicon.

The company guided to $12.7 billion to $13.7 billion for Q1, a sequential decline, and gross margins are expected to compress further. 

EPS estimates are soft. And while investors are excited about the long-term AI positioning and internal foundry narrative, those are stories playing out over five to seven years, not five to seven quarters.

Intel Foundry Services (IFS) is being positioned as the next great catalyst, but it’s still generating minimal revenue – just $291 million last quarter. 

And the Mobileye (MBLY) spin-off, once thought to be a crown jewel, hasn’t contributed much to excitement or valuation. 

The IoT and programmable units are steady, but they’re not needle-movers. Meanwhile, R&D is up, SG&A is up, and free cash flow is still negative.

Yes, Intel has $24 billion in cash and short-term investments. But it also has $50 billion in long-term debt and a capital plan that would make a sovereign wealth fund sweat. 

The dividend was slashed last year, and while it’s been maintained since, the company is clearly in preservation mode.

That said, there are bright spots. Intel 3 is on track. Meteor Lake shipped on schedule. Gaudi 3, its AI accelerator, is reportedly gaining interest as enterprises look for alternatives to NVIDIA (NVDA). 

And government support – both financial and strategic – is about as strong as it gets. Intel isn’t going anywhere.

Still, competitors like Taiwan Semiconductor Manufacturing Company (TSM) continue to dominate high-performance foundry share, while AMD pushes aggressively in both server and AI. 

Even Qualcomm (QCOM) is taking swings at data center and edge AI. Intel may be getting back in the race, but the rest of the field didn’t stop moving.

But a durable re-rating requires more than geopolitical relevance. It requires earnings. And that’s still a few years out, assuming everything breaks in Intel’s favor.

So yes, the demand story is real. The comeback narrative is compelling. But until those margins get out of the ICU, I’ll be watching this rally with one eyebrow raised.

Tech titan? Maybe someday.

Utility? For now, it’s looking awfully familiar.