Intel foundry eyes faster path to profit, vows more aggressive push on 1.4nm

Intel's foundry business is on track to turn profitable sooner than expected, the company said, as it pushes to overtake Samsung Electronics for the No. 2 spot in the global contract chipmaking market.

Once battered by massive losses that sparked layoffs and rumors of a business sale, Intel Foundry has been drawing in AI chip orders that even TSMC has struggled to absorb — a turnaround now gathering momentum.

Intel signaled it will make an even more aggressive bet on its 14A (1.4-nanometer) process than on its current 18A (1.8-nanometer) node, aiming to lock in major customers such as Apple and Nvidia by moving quickly to the leading-edge technology.

With AI-driven demand overwhelming TSMC's capacity and chipmakers scrambling for alternatives, the race between Samsung Electronics and Intel to capture those customers is set to intensify.

Intel Chief Financial Officer Dave Zinsner laid out the foundry division's current status and outlook at the Bank of America Global Technology Conference on Tuesday, according to industry sources.

Zinsner drew attention when he said Intel had originally targeted break-even by late 2027, but that the milestone would likely arrive <style ref="s0">at least a quarter earlier — possibly a bit more</style>.

He said the challenge at the start of last year was improving the performance and yield — the share of defect-free chips — of the 18A process. "We first stabilized performance, then worked every month to push yields higher," he said. "Yields moved up and down in the first few months, but after that we saw improvement in line with industry standards."

He acknowledged that significant work remains to reach a target profit margin, but added that <style ref="s0">yields are now at a "respectable" level</style>, signaling confidence in the process.

Zinsner's remarks were widely interpreted as indicating that 18A yields have reached a level stable enough to generate meaningful profit.

In semiconductor manufacturing, yields of at least 60 to 70 percent are generally required to sustain profitable production. As the share of defective chips on a wafer falls and good chips increase, profitability rises accordingly. Yields of 80 to 90 percent or above open the door to maximizing margins.

Intel surprised the industry last October when it announced it had begun mass production on its 18A process ahead of both TSMC and Samsung Electronics. The company operates production facilities in Arizona and Oregon.

The first PC processor built on the 18A process — the Intel Core Ultra 3 series — began appearing in laptops from Samsung Electronics, LG Electronics and others in the first quarter of this year.

Zinsner said the 18A laptop process is <style ref="s0">ramping up faster than any client-segment product in the past five years</style>, attributing the pace to monthly yield gains on the node. His remarks suggested output from the 18A process will scale up rapidly in the months ahead.

Microsoft and Amazon had already committed to placing some chip orders with Intel Foundry's 18A process, giving the division's revival tangible form.

More recently, Apple was reported to be in discussions about outsourcing production of the application processors — the core chips powering the iPhone — to both Samsung Electronics and Intel.

Intel has previously stated its goal of surpassing Samsung Electronics to become the world's second-largest foundry by 2030. After struggling with weak orders in the face of TSMC's dominant market position, Intel received a significant boost last August when the Trump administration acquired a 10 percent stake in the company as part of a broader effort to support domestic chipmakers.

Supply bottlenecks at TSMC — where AI chip orders have piled up, lead times have lengthened and process prices have climbed — have also created an opening for Samsung Electronics and Intel Foundry to position themselves as credible alternatives.

Zinsner pressed further, expressing strong confidence in the 14A node as well. "<style ref="s0">We have a more aggressive plan for 14A than we did for 18A</style>," he said. "It's still early, but 14A will be much easier — because we're using the same gate-all-around and backside power delivery network architecture we already implemented in 18A."

By reusing technologies already proven in the 18A process, Intel expects to achieve stable yields and performance on 14A sooner. Samsung Electronics has targeted 2029 for its entry into the 1.4-nanometer node, while TSMC has set a 2028 goal for 14A mass production.

Zinsner said Intel has already made substantial facility investments and <style ref="s0">holds multiple cleanrooms that could be ready within one to two years</style>. "A year ago that felt like a burden, but now — given where we are in the growth cycle — I'm glad we have them," he said, adding that discussions on moving equipment in are already underway.

Meanwhile, TSMC Chairman Wei Che-chia, speaking at the company's annual shareholder meeting on Wednesday, declined to name Samsung Electronics directly when asked whether it could catch up to TSMC. But he drew on history to make his point: "Twenty years ago, a competitor said it would catch TSMC within 10 years. Ten years ago, another said the same thing," he said. "Competitors are, in effect, just dreaming."

According to market research firm TrendForce, TSMC held a 69.9 percent share of the global foundry market last year, with Samsung Electronics in second place at 7.2 percent. The gap between the two widened from 55 percentage points in 2024 to 62.7 percentage points last year.