Intel posted its fastest revenue growth in 15 years, with Q2 sales jumping 25% year-over-year to $16.1 billion — a surge driven almost entirely by AI data centre demand. The number caught analysts off guard and sent the stock higher in after-hours trading.
🔍 THE BOTTOM LINE
Intel has been the AI boom’s forgotten player for two years — Nvidia ate the GPU market, AMD positioned itself as the alternative, and Intel’s foundry ambitions looked like a money pit. A 25% revenue jump does not erase that narrative, but it does complicate it. The company raised Q3 guidance to $15.8–16.8 billion, suggesting the growth is not a one-quarter fluke. The real question is whether Intel can convert revenue growth into the margin recovery that justifies its $100+ billion foundry investment.
What is Intel’s situation?
Intel is the largest semiconductor company by revenue that still designs and manufactures its own chips (a model called ” IDM” — integrated device manufacturer). After years of losing market share to Nvidia in AI accelerators and to AMD in server CPUs, Intel underwent a radical restructuring under CEO Lip-Bu Tan, who took over in early 2026. The strategy: use the AI data centre boom to backfill revenue while the foundry business — which manufactures chips for other companies — ramps toward profitability. This quarter is the first evidence that strategy is working on the revenue side.
The Numbers That Matter
According to Financial Times and Yahoo Finance reporting:
- Q2 revenue: $16.1 billion, up 25% year-over-year — Intel’s fastest growth in over a decade
- Q3 guidance: $15.8–16.8 billion, raised from previous estimates, suggesting sustained demand
- Gross margin: 35.4%, improved from recent quarters but still well below the 50%+ margins Intel historically posted
- Stock reaction: shares surged in after-hours trading, with INTC trading around $100
The revenue figure is the headline, but the margin number is the subtext. A 25% revenue increase at 35% gross margin generates less profit than a 10% increase at 55% margin. Intel is growing again — but it is growing while carrying the enormous cost of its foundry buildout, which is not yet profitable.
Why AI Data Centres Are Suddenly Buying Intel Again
The FT reports that the growth was “fueled by surging demand from artificial intelligence data centres.” This deserves unpacking, because Intel is not selling the GPUs that power AI training — that is still Nvidia’s market.
What Intel is selling is everything around the GPUs: the server CPUs that manage data centre operations, the networking infrastructure that connects GPU clusters, and the Xeon processors that handle inference workloads where Nvidia’s accelerators are overkill. AI data centres are not just racks of H100s — they are full-stack infrastructure, and Intel’s data centre and AI segment (DCAI) covers the parts Nvidia does not.
The Infosys collaboration mentioned in Intel’s corporate profile — a “multi-layer AI fabric that unifies infrastructure, models, data, applications, and workflows into a composable and agent-ready ecosystem” — points to the strategy: Intel positions itself as the infrastructure layer for agentic AI, not the accelerator layer for model training. That is a smaller market than Nvidia’s, but it is growing fast enough to deliver 25% revenue jumps.
The Foundry Question
Intel’s foundry business — which manufactures chips for other companies, not just Intel’s own designs — remains the company’s biggest bet and biggest risk. The division is losing money as it builds out fabrication capacity in Arizona, Ohio, and Ireland. Lip-Bu Tan’s confidence in demand, cited in Yahoo Finance’s analysis, is partly about foundry: if AI chip designers need manufacturing capacity outside TSMC’s Taiwan-dominated supply chain, Intel Foundry is the only Western alternative at scale.
This connects to the broader sovereign AI infrastructure story we have tracked at The 18-Day AI Export Control Standoff Is Over. Commerce Lifted the Anthropic Ban. and Trump Just Opened the AI Chip Floodgates to the UAE — and Nobody — the geopolitical push to diversify chip manufacturing away from Taiwan creates a structural tailwind for Intel Foundry, regardless of whether Intel’s own chip designs win in the market.
But foundry profitability is years away. The current quarter’s 35.4% gross margin reflects the drag. Investors buying Intel at $100 are betting that foundry losses will narrow as external customer volume ramps — a bet that requires both execution and patience.
NZ Angle: What Intel’s Recovery Means Downstream
New Zealand’s data centre market is small in global terms but growing. Local providers — Datacom, Catalyst, and the hyperscaler regions planned by Microsoft and AWS in Aotearoa — buy the same Intel infrastructure that drove this quarter’s growth. If Intel’s AI data centre momentum continues, Kiwi cloud and AI infrastructure costs could benefit from increased competition in the server CPU market, where Intel’s pricing power has been eroded by AMD’s EPYC line.
The more interesting angle is sovereign AI compute. If New Zealand eventually builds domestic AI infrastructure — as discussed in When Data Centers Become Targets — What Iran — Intel Foundry is one of the few Western manufacturers that could supply custom silicon without depending on TSMC. That is a long-term consideration, not a current one, but it matters for policy planners thinking about 2030-era supply chains.
The Competitive Landscape
Intel’s 25% growth is impressive, but context matters. Nvidia posted revenue growth of over 100% year-over-year in its most recent quarters. AMD’s data centre revenue has been growing 50%+. Intel is growing from a lower base, recovering lost ground rather than capturing new markets.
The difference: Intel is the only one of the three that is also building a foundry business. If the foundry strategy works, Intel’s revenue ceiling is higher than any pure chip designer — because it can manufacture competitors’ chips too. If it does not, Intel is a legacy CPU company growing at 25% in a market where its competitors are growing faster.
Lip-Bu Tan’s Q3 guidance — $15.8–16.8 billion — signals confidence that the next quarter sustains the trajectory. The market will be watching margins, not revenue, for evidence the foundry drag is easing.
❓ FAQ
Is Intel back to competing with Nvidia in AI? Not directly. Intel’s growth is in server CPUs and data centre infrastructure — the layers around the GPU accelerators that Nvidia dominates. Intel’s Gaudi AI accelerators exist but hold minimal market share. The 25% growth is Intel winning the “everything except the GPU” market in AI data centres.
Why are gross margins only 35.4%? Intel’s foundry business is in heavy investment mode, building fabrication plants that cost tens of billions each. The foundry division is losing money on every wafer it produces for external customers because volume is still low. Until foundry utilisation rates climb, the division drags down Intel’s blended margin.
Could Intel’s foundry business actually challenge TSMC? Not in the next 2–3 years. TSMC has a multi-decade head start in process technology and customer relationships. Intel Foundry’s value proposition is geopolitical: it is the only Western fab at leading-edge scale. If the US government continues to prioritise domestic manufacturing through CHIPS Act funding and export controls, Intel Foundry gets structural support regardless of pure competitiveness.
What does this mean for the Intel stock price? The market reacted positively — INTC surged in after-hours trading. But Intel’s valuation still reflects deep uncertainty about foundry profitability. A 25% revenue jump that does not translate into margin expansion is a growth story without a profit story. The Q3 guidance raise helps; sustained margin improvement is what would change the thesis.
🔍 THE BOTTOM LINE
Intel’s best quarter in 15 years is a genuine milestone — the company that looked like it was losing the AI era has found a lane. Server CPUs, data centre networking, and the infrastructure layer around AI training are growing fast enough to deliver 25% revenue jumps. But the foundry business that represents Intel’s biggest bet is still bleeding, and 35% gross margins are a reminder that growth and profitability are different stories. Lip-Bu Tan’s raised guidance says the revenue trajectory holds. The margin trajectory — that is the one that determines whether Intel’s comeback is real or just a good quarter.
