Intel has set January as the month for the start of sales of its Panther Lake chips. The company is touting impressive statistics for the new CPU line, saying users can expect more than 50% better CPU and GPU performance — as well as more than 40% lower power usage, at a similar single-threaded performance — compared to Lunar Lake or Arrow Lake-H.

It is common for hardware companies to cherry-pick benchmarks and claim significant performance improvement. With that in mind, Intel’s performance uplift claims should be taken with a grain of salt until independent reviewers have the opportunity to examine the chips and conduct their own benchmarks.

A significant part of the claimed performance and power efficiency gains is attributed to Intel’s use of its 18A manufacturing process to produce the chips.

This is why these statistics have a dual purpose. The first is advertising new chips to consumers.

But Intel is also advertising its fabs, sending the message: “Hey, come to us, make chips at our fabs, and your chips will get an efficiency boost!”

Intel’s fabs vs. Intel’s salvation

Intel’s stock gained 47% in value over the last month. That is a huge jump for a short period of time. The main driver behind this growth was the company’s deal with Nvidia, as well as a couple of rumored deals.

Rumors suggested that Intel is attempting to strike a deal with Apple and AMD to use its fabs. There was also a quite absurd and refuted rumor about Intel trying to make a deal with Taiwan Semiconductor Manufacturing Company (TSMC) to invest in it.

Related: Bank of America updates Nvidia stock outlook on AI bubble fears

The reason these supposed talks were effective in boosting the stock is that if Intel could convince AMD or Apple to use its fabs, it might save the company.

Intel’s fabs are losing money, with more than $13 billion in losses in the last four quarters alone. Intel reported a net loss of $2.9 billion in Q2 2025, compared to a loss of $1.6 billion in Q2 2024.

More Tech Stocks:

• Musk’s Netflix boycott could actually hurt the streamer
• OpenAI’s deal with AMD proves AI race has just begun
• Goldman Sachs tweaks Nvidia’s stock price target with a twist
• The stock market laughed, then Palantir redefined the fight

The explanation for why Intel is having a big problem finding customers for its fabs, despite touting the performance and power efficiency gains its manufacturing process can provide, is very simple.

It is what the company isn’t saying.

Intel recently invited a small group of analysts and reporters for a series of private briefings and factory tours. It revealed a lot about Panther Lake, but didn’t disclose manufacturing yields.

As we’ve seen with the performance benchmarks, if they are good, the company will have no problem publicizing them. If the company remains silent about a factor that is most important to its potential customers, it is probably because it’s not great. 

Bank of America says Intel foundries unlikely to be profitable by 2027

Bank of America analysts Vivek Arya and his team provided an opinion on Intel stock following the $80 billion jump in Intel’s market cap.

The team noted several pain points for Intel: 

• The company struggles to gain market share against AMD and ARM. 
• It has no notable products in AI accelerators and scale-up/scale-out networking architecture/protocol. 
• Its foundry is unsustainable alone; 18A has no external support.

Related: Expert explains everything you need to know about this AI law

“Even if Intel Foundry maintains its internal customer AND potentially lands substantial external customer revenue [in the range of $5 billion to $10 billion], we estimate the foundry to remain unprofitable on both gross-profit and operating-profit bases by [calendar year 2027 estimates],” analysts said.

They added that the company would need approximately $15 billion to $20 billion in external foundry revenue to be profitable on a gross-profit basis, and even then, it could still be unprofitable on an operating-profit basis.

Analysts noted downside risks for Intel:

• Lower than yield/ramp at Intel Foundry, particularly for its new 18A and upcoming 14A nodes
• Lack of material external foundry customer in wafer processing
• Weaker-than-expected trends in a mature PC market
• Accelerated share loss to major CPU competitors

Intel’s identified upside risks:

• Key external foundry packaging/wafer deals that could significantly boost sales/utilization
• Greater-than-expected yields/ramps at 18A and upcoming 14A nodes, resulting in greater gross margin/utilization profile
• Stronger-than-expected PC market from Windows 10 refresh or AI uplift
• Geopolitical tensions boosting sentiment for domestic manufacturing assets

Arya downgraded the stock from neutral to underperform, with a target price of $34, based on 3.3 multiple of his enterprise value to sales ratio estimate for 2026, in line with a historical range of 1.7 to 4.

Related: Qualcomm shakes up AI on the edge with a huge surprise