INTC Q3 2026 Earnings Analysis
Intel beat Q3 expectations with record $7.3B DCAI revenue (+80% YoY) from AI infrastructure demand, but faces PC weakness from memory constraints and Israel fab disruptions affecting Q4 guidance.
Key Metrics
要点总结
- DCAI revenue hit record $7.3B (+80% YoY), driven by AI infrastructure demand and Granite Rapids ramp.
- PC market declining due to memory supply constraints; Q4 guidance lowered, client revenue down 9%.
- Planning substantial 2027 CapEx increases despite near-term challenges; FCF recovery expected 2028.
收听平台
可用语言
Transcript
// Full episode scriptWelcome to Beta Finch, your AI-powered earnings breakdown! I'm Alex.
And I'm Jordan. We've got a big one today — Intel's Q3 2026 report.
Before we dive in, quick disclaimer: This podcast is AI-generated content for educational and entertainment purposes only. Nothing we discuss should be considered investment advice. Always do your own research and consult a qualified financial advisor before making any investment decisions.
Alright, with that out of the way — Alex, this was a tale of two businesses.
Totally. Let's start with the headline numbers. Intel posted Q3 revenue of $16.5 billion, beating guidance, with non-GAAP gross margin at 44.1% — about 200 basis points better than expected. EPS came in at $0.52 versus guidance of $0.38. That's the eighth straight quarter they've beaten their own expectations.
Eight in a row — that's a real streak for a company that's been in turnaround mode. And the engine behind it is clear: data center.
Right, DCAI — Data Center and AI — revenue hit $7.3 billion, up 17% sequentially and nearly 80% year-over-year. CEO Lip-Bu Tan called it the strongest growth for that segment in 15 years.
Server CPU demand tied to AI infrastructure is just on fire. And Intel's framing here is interesting — they're positioning CPUs as the unsung hero of the "agentic AI" era. The argument is that while GPUs are great at reasoning, CPUs are far more energy-efficient at the "finding" tasks — orchestration, data retrieval, that kind of thing.
And the TAM numbers they threw out were eye-popping. Management thinks the server CPU market could go from $40 billion today to over $200 billion in five years, mostly driven by agentic workloads.
That's a 5x expansion. Bold claim, but if even a fraction of that materializes, it's a massive tailwind for Intel's core franchise.
Now here's the flip side — client computing, the CCPG segment, which covers PCs. Revenue there was $8.1 billion, down 9% sequentially and below expectations.
Yeah, this is the part that's dragging things down. High memory prices and tight IT budgets are squeezing the PC market hard. Management expects full-year 2026 PC unit TAM to decline about 12%, with the second half down more than 20%.
And it's not getting better quickly — they're guiding PC TAM down another 10% in 2027.
The silver lining: AI PC revenue actually grew 35% year-over-year even as the broader category shrank. And their new Wildcat Lake chip, built on the Intel 18A node, saw unit growth of about 70% sequentially — it's actually sold out.
Let's talk guidance, because the Q4 outlook was a bit of a reality check. Intel's guiding revenue to $15-16 billion, with gross margin expected to drop to 38%, and EPS of just $0.25.
That's a notable step down from Q3, and there were a couple of specific culprits. First, CCPG revenue is expected to decline more than 20% sequentially as customers rapidly burn through PC inventory. Second — and this one's a gut-punch — there was a tragic accident at Intel's Israel fab, Fab 28, that caused an extended shutdown.
That's going to ding Q4 revenue by roughly $300 million and gross margin by about 200 basis points, mostly hitting their 10-nanometer server shipments. CFO Dave Zinsner said they expect a similar impact in Q1 before it fades.
It's a sobering reminder that even with strong demand, manufacturing is still fragile and complex. On top of that, there's an industry-wide shortage of advanced substrates needed for packaging — another supply bottleneck limiting how fast Intel can actually convert demand into revenue.
So demand is outstripping supply across the board right now. Management actually said they expect the industry to remain supply-constrained through 2027 and into 2028.
Which, counterintuitively, is actually bullish if you're Intel — as long as they can keep expanding capacity. They raised $23 billion in an equity offering back in August specifically to fund this buildout, and CapEx for 2027 is expected to be "up strongly" from this year's $20 billion.
Let's hit a few Q&A highlights, because there was some great color from analysts. Stacy Rasgon from Bernstein pushed hard on gross margins, and Zinsner's answer was basically: expect 40-60% fall-through on revenue growth next year, but it won't be smooth — underutilization on 18A and the lower-margin purpose-built silicon ramp will create some noise.
And on market share, there was a pointed question about AMD's aggressive growth guidance. Zinsner's response was refreshingly blunt — he said market share next year will largely be "a function of supply." Basically, whoever can actually ship more chips wins share, regardless of who has the better roadmap on paper.
There was also a nice nugget on the foundry side — Intel's working toward its 14A process node, they released an early "0.9 PDK" to select customers in October, and Lip-Bu Tan hinted at announcing a couple of external customers in the next two quarters.
And don't skip over the quantum computing announcement — Intel said they intend to acquire Classiq, an Israeli quantum software company, which would make Intel a "full-stack" quantum computing business. Early days, but it shows the company is still planting flags for the next decade, even amid near-term turbulence.
So, bringing it all together — what does this mean for investors?
The story right now is really a bet on timing. The data center AI business is genuinely firing on all cylinders, and the long-term TAM story is compelling. But the PC business is in a real slump driven by memory shortages that management doesn't expect to resolve until later in 2027 at the earliest. And then you've got this one-time wildcard with the Israel fab incident weighing on near-term numbers.
Zinsner was also candid that 2027 will likely NOT be a positive free cash flow year — this is an investment year, with CapEx ramping hard to build out 18A, Intel 3, and prep for 14A.
So patience seems to be the operative word here. Before we wrap — a reminder: everything discussed is AI-generated analysis for educational purposes. Past performance doesn't guarantee future results. Please do your own due diligence.
Well said. Intel's turnaround story continues to be a mix of genuine progress and real operational growing pains. The data center momentum is the bright spot to watch, while the PC softness and supply constraints are the hurdles to track into 2027.
We'll be watching closely for updates at CES in January, especially around Nova Lake and that external foundry customer news.
That's it for this episode of Beta Finch. Thanks for listening, and we'll catch you next time!
See you then!