Going into Thursday’s release, analyst opinion on Intel (INTC) was about as split as it gets for a stock this size. Morgan Stanley held a $75 target and a Hold rating as of 21 July, while HSBC sat at $200 with a Buy rating from earlier in the month.
Susquehanna, Stifel and KeyBanc filled the middle ground at $115, $120 and $155 respectively, all still rated Hold or Buy under a consensus that read Outperform overall.
Since every one of those targets was set before Intel actually reported, the real question now is whether Thursday’s numbers gave either camp reason to change its mind. Or whether the print simply handed each side more material to argue the case it already held.
Strong Quarter Gives Bulls Fresh Evidence
On the surface, the quarter reads like a win for the bulls. Revenue came in at $16.1bn, up 25.4% year on year and $1.8bn above the April guidance midpoint.
Non-GAAP gross margin reached 41.8%, some 2.8 percentage points ahead of forecast, while non-GAAP EPS of $0.42 beat the guided figure by $0.22.
Data Center and AI did the heavy lifting here. Segment revenue rose to $6.3bn from $3.9bn a year earlier, as segment operating margin climbed to 39.5% from 16.1%.
Purpose built silicon revenue nearly tripled year on year, and management pointed to a fresh batch of long-term customer agreements covering both pricing and volume. This is exactly the kind of evidence that supports HSBC and KeyBanc’s higher targets.
Foundry Progress Still Holds the Key
Foundry told a similarly encouraging story. Segment losses narrowed to $2.1bn from $3.2bn a year earlier, taking the operating margin to -36.2% from -71.7%.
18A yields have reportedly climbed to around 85% from roughly 65% the previous quarter, a level generally regarded as the point where a process node becomes commercially viable rather than merely functional.
Yet this is precisely where the bears retain their strongest argument. Reports linking Nvidia, AMD, Apple, Microsoft and OpenAI to design work on Intel’s advanced nodes remain unconfirmed as binding, high volume contracts.
CFO David Zinsner has said signals from external customers should become more concrete through the second half of the year and into early 2027. That sounds less like confirmation and more like an admission that the cornerstone customer Morgan Stanley and Stifel are waiting for hasn’t shown up yet.
Profit Improvement Comes With Spending Pressure
Guidance for Q3 tempers the picture rather than extending it. Revenue is expected between $15.8bn and $16.8bn, non-GAAP gross margin around 42% and non-GAAP EPS of $0.38, all representing continued growth without matching the scale of upside surprise seen in Q2.
The headline P/E ratio of -148.41 needs unpacking before anyone reads too much into it. The GAAP loss per share of $2.16 was driven largely by a $2.45 non-cash mark-to-market charge tied to escrowed shares linked to the CHIPS Act Secure Enclave agreement with the US government, rather than by weak trading.
Once that one-off is stripped out, the non-GAAP figure of $0.42 paints a considerably healthier picture of underlying profitability. Adjusted free cash flow still stayed negative at $8.4bn for the quarter though, a reminder that the capital intensity behind this turnaround hasn’t eased just because foundry losses are narrowing.
The Next Move Depends on Customer Commitments
None of this cleanly favours either side of the analyst split. The revenue and margin beats support a case for upgrades toward the higher end of the range.
The absence of a named, binding foundry customer leaves the lower targets just as defensible until that changes. Thursday’s release looks more like confirmation that both camps had a reasonable read on the business than the kind of print that forces a rapid re-rating in either direction.
The shares had already run hard through the first half of 2026 before July’s sharp pullback and partial recovery, so much of this quarter’s good news looks priced in rather than newly discovered.
I would want to see actual analyst revisions land in the days following this report, alongside a confirmed foundry customer commitment, before treating either end of the $75 to $200 range as the more likely outcome.
