JPMorgan upgraded Intel to Neutral from Underweight in an Oct. 18 report, setting a December 2027 price target of $85. Using the report’s listed Oct. 16 closing price of $89.85, the target still implies about 5% downside.
That gap helps explain the change. JPMorgan’s view is that after a roughly 36% drop from Intel’s June high, the stock now reflects more of the risks that had supported its cautious stance. The bank said execution is improving, but the company is still some distance from completing a turnaround that may take years to play out.
Valuation has reset, but earnings still need to catch up
JPMorgan said Intel’s shares have fallen about 36% from their June peak, a steeper decline than the roughly 20% to 25% drop seen across peers over the same period. In the bank’s view, some of the factors behind its earlier caution are now being priced in.
Competitive pressure over the past few years remains clear. Since JPMorgan cut the stock in November 2022, Intel’s revenue share in server x86 CPUs has dropped from 77% to 53%, while its PC CPU share has fallen from 81% to 68%. The bank expects share losses could continue over the next few quarters, especially in servers.
At the same time, the recent selloff has narrowed Intel’s valuation premium versus peers. Based on JPMorgan’s estimate of $3.60 in adjusted earnings per share for 2028, Intel is trading at about 25 times earnings, close to the 24 times peer multiple used by the bank.
The $85 target is derived from that 2028 earnings estimate multiplied by roughly 24 times earnings. In other words, even after the pullback, the current valuation still assumes a continued recovery in profit.
Share dilution had also been a major overhang. The report said Intel’s share count has risen from about 4.4 billion a year ago to about 5.1 billion. JPMorgan said the dilution from the last round of financing has largely been absorbed into valuation and earnings assumptions, and the market is increasingly pricing the risk of further large-scale dilution.
But the bank added that pricing the risk is not the same as removing the funding need.
Execution is improving, with data center driving profit recovery
More important than the valuation reset, JPMorgan said, is its growing confidence in execution under Lip-Bu Tan.
Within Intel’s core compute business, the Data Center and AI segment, or DCAI, has become the company’s largest profit contributor. According to the report, the segment posted $2.5 billion in operating profit in the second quarter of 2026, the first time it surpassed the client business.
JPMorgan said supply of Granite Rapids server processors is tight, and long-term agreements with customers including Google extend demand visibility into 2027. That gives the company support on both revenue and profit.
The bank also tied part of its thesis to changes in AI demand. It said agentic AI workloads carry relatively high CPU requirements, which means AI infrastructure spending could benefit Intel through its server processor business.
Even with recent market concerns about the durability of AI infrastructure spending, JPMorgan expects growth in agentic AI demand and tight supply of advanced chips to support upward revisions to Intel’s earnings outlook in the coming earnings season and through 2027. Its 2028 EPS forecast is about 15% above market consensus.
The client business presents a different mix of forces. The report said AI PCs now account for about two-thirds of Intel’s client product mix, and favorable pricing and mix have supported results. But rising memory prices are lifting bill-of-materials costs for complete systems, which could weigh on PC unit demand in the second half of 2026 and into 2027.
Whether mix improvement can offset softer end demand remains a key point to watch.
Manufacturing progress has improved confidence, but foundry still lacks scale
Manufacturing is one of the areas where JPMorgan sees the clearest improvement.
The report said 18A-P has entered risk production, while 14A is planned for volume production in 2028, with risk production targeted for the second half of 2027. JPMorgan also said yield, process and design-rule compatibility, and advanced packaging capability have all improved.
Backlog for EMIB-T advanced packaging is also building a base for capacity ramp-up in 2027. For JPMorgan, those changes strengthen confidence that Intel’s manufacturing technology foundation is becoming more stable.
Still, the bank said there is a meaningful gap between technical progress and large-scale external revenue.
According to the report, Intel’s external foundry business generated annualized revenue of only close to $1 billion in the first half of 2026. JPMorgan placed that figure alongside Taiwan Semiconductor Manufacturing Co.’s roughly $150 billion in annualized total revenue over the same period to show the scale difference. The report noted the business scopes are not fully comparable, but said Intel’s external foundry revenue remains small, which is one reason the bank is still cautious.
The next test is whether Intel can win major outside customers on advanced process nodes and convert technical progress into recurring orders. JPMorgan said named anchor customers would help validate the economics of the foundry business and the return on continued capital investment.
ASIC services offer another path for growth
Custom silicon is another area JPMorgan highlighted.
The report said Intel’s ASIC services business is approaching $2 billion in annualized revenue, about three times the level of a year earlier, and could move toward $4 billion over time.
ASIC stands for application-specific integrated circuit, chips designed for particular tasks. As large cloud companies increase spending on custom silicon, JPMorgan said Intel could use its chip intellectual property, system-on-chip design capability, and manufacturing capacity to win more business.
That revenue opportunity still depends on design wins and customer programs moving into production. The bank listed new large cloud or AI lab customers as an important upside catalyst to watch.
Profit may recover before cash flow does
JPMorgan’s financial model points to continued earnings growth. It expects Intel’s revenue to rise from about $63.3 billion in 2026 to $77.7 billion in 2027 and $94.3 billion in 2028. Adjusted EPS is projected at $1.58, $2.44, and $3.60, respectively.
But there is still a lag between profit recovery and positive free cash flow.
Under the bank’s estimates, Intel will generate about $19 billion in operating cash flow in 2026 against roughly $19.2 billion in capital spending, leaving free cash flow slightly negative. In 2027, operating cash flow is expected to increase to about $22.5 billion, but capital expenditures are projected to rise further to $26 billion, widening the free cash flow gap to about $3.5 billion.
JPMorgan does not expect free cash flow to turn positive until 2028, when it forecasts about $8.2 billion.
That helps explain why the bank stopped at Neutral rather than turning more constructive. Building out manufacturing capacity continues to consume cash, and the report cited management as saying free cash flow in 2027 will remain 「more challenging」. The company may need to return to capital markets for funding.
If that financing includes equity issuance, existing shareholders could face further dilution.
Why JPMorgan is still waiting before turning bullish
JPMorgan said it still wants to see several points validated: whether Intel can keep delivering on its data center CPU roadmap, whether it can stabilize share against AMD, whether AI chips and inference products can win more orders, and whether external foundry customers can generate meaningful revenue.
The bank is also watching for stronger data center networking capabilities.
Intel has provided more evidence of better execution, and the stock’s pullback has eased some of the earlier valuation pressure. For JPMorgan, though, a move beyond Neutral will require proof from customers, market share, and cash flow that the company’s multi-year transformation can produce sustainable returns.