Intel Admits Defeat in the AI Race

Intel Admits Defeat in the AI Race

Ondřej Barták
Ondřej Barták
Entrepreneur and Programmer
15. 7. 2025
4 minutes reading · 12 views
Intel Admits Defeat in the AI Race

Intel Admits Defeat in the AI Race: “It Is Too Late to Catch Up With the Competition”

Intel, which until recently dominated the world of processors, is apparently going through the greatest crisis in its history. In an internal conversation in February, new CEO Lip-Bu Tan openly admitted that Intel is no longer among the world’s top ten semiconductor companies, according to an OregonTech report. This shocking news comes as the company is laying off thousands of employees worldwide and facing losses of $16 billion for the third quarter of last year.

Falling From the Ranks of the Largest Semiconductor Companies

“Twenty or thirty years ago, we were truly the leaders. Now I think the world has changed. We are not among the top 10 semiconductor companies,” Tan told employees during a global Q&A session (questions and answers). This admission represents a dramatic reversal for a company that once practically monopolized the CPU market and even considered acquiring Nvidia for $20 billion. Ironically, Nvidia is now worth $4 trillion.

While companies such as AMD, Nvidia, Apple, TSMC, and Samsung now outperform Intel, smaller chipmakers such as Broadcom, MediaTek, Micron, and SK Hynix still do not match the “blue team” (Intel’s nickname) in terms of overall impact. Nevertheless, the situation is alarming for Intel, especially considering that it was regarded as an unshakable giant just a few years ago.

Too Late to Catch Up With AI Competitors

One of Intel’s biggest mistakes was underestimating the artificial intelligence revolution. While OpenAI launched the “second major revolution” in machine learning with ChatGPT, Intel found itself left out. This era is powered almost exclusively by Nvidia’s data center GPUs, highlighting yet another sector in which Intel failed to capitalize on its position.

“For training (AI), I think it is too late for us,” Lip-Bu Tan commented. Instead, Intel plans to shift its focus to edge AI, which is intended to bring AI processing directly to devices such as computers rather than relying on cloud computing. Tan also highlighted agentic AI—an emerging field in which AI systems can operate autonomously without constant human input—as a key area of growth.

Lip-Bu Tan

Technical Problems and Manufacturing Outsourcing

For many years, Intel relied on vertical integration, manufacturing its own chips in its own factories. However, this strategic advantage has become more of a burden. Its lagging manufacturing nodes cannot compete with Taiwan’s capabilities, likely preventing Intel processors from reaching their full potential.

In 2023, the company began outsourcing production of the GPU portion of its Meteor Lake chips to TSMC. This partnership expanded, with TSMC effectively manufacturing the entire compute section for Lunar Lake. Now, in 2025, roughly 30% of production is outsourced to TSMC. This is a long-overdue admission of failure that could have been avoided if Intel had designed its high-end processors with external manufacturing in mind from the outset.

Layoffs and Financial Problems

Intel is laying off thousands of employees worldwide in an effort to reduce costs. Expenses have risen sharply due to heavy spending on research and development of future nodes, and the company is facing a loss of $16 billion for the third quarter of last year. Intel’s revival must be “a marathon,” Tan said, and he hopes to change the corporate culture and remain “humble” in listening to the industry’s changing demands.

A New Strategy Under Lip-Bu Tan

Last year, Intel’s board forced former CEO Pat Gelsinger to leave the company and replaced him with Lip-Bu Tan, who clearly has a different, more efficient vision for the company. Instead of focusing simultaneously on several different areas, such as CPUs, GPUs, foundry operations, and others, Lip wants to focus on what the company can do well at any given time.

Pat Gelsinger

Tan also expressed optimism about significant recent hires that could help guide Intel back to relevance in AI, while indicating that further talent acquisitions are on the way. “Stay tuned. Several more people are joining,” Tan said.

The Future: Separation or Unification?

This development follows long-running speculation about splitting Intel in two and creating a new foundry division that would operate as an independent subsidiary, turning the core Intel business into a fabless chipmaker (a chipmaker without its own factories). Both AMD and Apple, Intel’s rivals in the CPU market, operate this way, and Nvidia has also always used TSMC or Samsung to manufacture its graphics cards.

It would be interesting to see the “blue team” shed its burden and move like a free animal through the ecosystem. However, it is too early to speculate, as 18A, Intel’s proposed salvation, is still a year away.

Intel finds itself at a critical crossroads in its history. While Tan admits that it is “too late” to compete in some areas of AI, his strategy focused on edge AI and agentic AI may offer a path to revival. The question remains whether Intel, after decades of dominance, can embrace the humility and speed demanded by today’s rapidly changing technology market.

Advertisement

Content created with help from UpTier.

SEO and GEO on autopilot. UpTier’s multi-agent systems write and optimize content for search engines and AI answers.

Discover UpTier ↗

Category:AI
Did you enjoy this article?
Discover more interesting posts on our blog
Back to blog

Related posts

Google makes Tasks free for Gemini 4 Hydrogen usersGoogle makes Tasks free for Gemini 4 Hydrogen users
Tasks is becoming available for free to hundreds of millions of Gemini 4 Hydrogen users. Google says it can call businesses, book appointments and complete purchases on their behalf.
1 min read
18. 11. 2026
OpenAI introduces GPT-6.2 Astra and Sol for professional and everyday workOpenAI introduces GPT-6.2 Astra and Sol for professional and everyday work
OpenAI has introduced two models: Astra for complex professional tasks and Sol for everyday work. Both are available in ChatGPT Work, Codex and through the API.
3 min read
17. 11. 2026
Tencent releases Hy4 Flash with image support and a context window of up to 1 million tokensTencent releases Hy4 Flash with image support and a context window of up to 1 million tokens
The lighter Hy4 Flash model targets chat, documents and routine code edits. It accepts text and images, supports tool use and, according to Tencent, matched Claude Haiku 5.5 in tests of everyday tasks.
1 min read
16. 11. 2026
Přihlaste se k odběru našeho newsletteru
Zůstaňte informováni o nejnovějších příspěvcích, exkluzivních nabídkách, a aktualizacích.
CodedTrip

Operated by CodedTrip LLC, USA.

YouTube
TikTok