A Strategic Fictional Analysis | This narrative blends documented facts with fictional dramatization
July 2025
Washington D.C. – The Deal Room
Lip-Bu Tan had negotiated billion-dollar deals before, but never one quite like this.
Across the mahogany table sat representatives from the U.S. Department of Commerce, the Department of Defense, Treasury officials, and – most unusually – a senior advisor from the White House. Behind them, lawyers from three different agencies reviewed documents.
Intel’s team was smaller: Tan, his CFO, Intel’s general counsel, and Frank Yeary representing the board.
“Let’s review the terms one more time,” Deputy Secretary Morrison said, pulling out a document stamped “CONFIDENTIAL.”
The United States Government Investment in Intel Corporation:
- Amount: $8.9 billion for 10% equity stake
- Share Price: $20.50 per share (roughly current market price)
- Board Seat: One non-voting observer position initially, converting to voting seat after two years
- Manufacturing Commitment: Intel must maintain foundry operations in the U.S. for minimum of 10 years
- Employment Requirements: No net reduction in U.S.-based manufacturing workforce for 3 years
- Technology Transfer Restrictions: Government approval required for any sale or licensing of advanced manufacturing technology to foreign entities
- Quarterly Reporting: Detailed operational and financial reports to oversight committee
Tan read through the terms again, feeling the weight of each condition. This wasn’t just an investment – it was a partnership that would fundamentally change Intel’s relationship with the government.
“The employment requirement concerns me,” Tan said. “We just laid off fifteen thousand people. If market conditions deteriorate further – “
“The requirement is specific to manufacturing workforce,” Morrison interrupted. “Engineers, technicians, fab operators. We’re not preventing you from optimizing administrative or non-critical functions.”
“And the board observer?” Yeary asked. “Who would it be?”
“Someone with semiconductor industry experience and security clearance. Not a political appointee. We’re thinking someone from DARPA or one of the national labs.”
Tan exchanged glances with Yeary. A DARPA representative would actually be helpful – someone who understood technology, who could open doors to defense contracts.
“What about the timing?” Tan asked. “When would the funds be available?”
“Pending final approval, which we expect within two weeks, funds would be wired within thirty days. Late August at the latest.”
Tan did the math. Intel was burning roughly $2.5 billion per quarter on foundry operations. The government money would buy them three to four quarters of runway – enough time for Project Paranoia to deliver results, or to admit defeat and pivot to a different strategy.
“I need to make a call,” Tan said.
He stepped out of the conference room and dialed a number he’d saved in his phone three weeks earlier.
Jensen Huang answered on the second ring.
“Jensen, it’s Lip-Bu. I need to ask you something confidentially.”
“Go ahead.”
“If the U.S. government takes a stake in Intel, if we become partially state-owned – does that change Nvidia’s view of us as a potential partner?”
A pause. Then Huang’s response was careful. “What kind of partnership are you thinking about?”
Three Weeks Later: Intel Headquarters
On August 19th, Intel’s stock opened at $23.15 – up 4% on rumors of a major announcement. By 9 AM Pacific time, those rumors were confirmed.
PRESS RELEASE
U.S. GOVERNMENT INVESTS $8.9 BILLION IN INTEL CORPORATION
Historic Partnership Ensures American Semiconductor Leadership
President Trump, speaking from the White House, framed it as a national security imperative:
“Intel is a great American company with an incredible future. We’re making sure that the future includes American manufacturing, American jobs, and American technological leadership. This is a great deal for America, and a great deal for Intel.”
The market’s reaction was mixed. Intel’s stock jumped another 8% on the capital infusion news, but some analysts worried about government interference.
“This sets a concerning precedent,” a CNBC commentator said. “The government picking winners and losers, taking equity stakes in private companies. It’s industrial policy, and historically, that hasn’t worked well in America.”
But inside Intel, the mood was one of relief. The company had been hemorrhaging cash. Now it had a lifeline.
Tan addressed employees that afternoon:
“This investment is a vote of confidence in Intel’s future and in American manufacturing. But let me be clear – this doesn’t change our mission. We still need to execute. We still need to win customers. The government gave us capital, but capital alone doesn’t build great chips. You do.”
The Secret Negotiations
What employees didn’t know – what almost nobody knew – was that Tan had been working on a second, even bigger deal.
For the past month, he’d been in confidential talks with Nvidia.
The idea had started with that phone call in July. Huang had been cautiously interested in exploring a partnership, but he needed to understand what Intel was offering that Nvidia couldn’t get elsewhere.
“Here’s what I’m proposing,” Tan had said during a private dinner in San Francisco. “Intel builds custom x86 CPUs designed specifically for Nvidia’s AI infrastructure. We optimize them to work seamlessly with your GPUs through NVLink. You get processors that are perfectly matched to your accelerators.”
“Why would we do that instead of using ARM processors from another vendor?” Huang asked.
“Two reasons. First, x86 has a massive software ecosystem. Enterprise customers trust it. Second – ” Tan paused ” – Intel has manufacturing capacity in the U.S. If tensions with China escalate, if Taiwan becomes unstable, you need a domestic alternative to TSMC.”
Huang had nodded slowly. “What else?”
“On the consumer side, we build PC chips that integrate your RTX graphics technology. Nvidia-Intel co-designed processors for laptops and desktops. We handle the CPU and system integration; you handle the GPU acceleration.”
“That’s interesting,” Huang admitted. “But it doesn’t address your foundry problem. Are we going to use Intel Foundry to manufacture Nvidia chips?”
This was the delicate question. Nvidia used TSMC exclusively. Getting them to even test Intel’s manufacturing would be a massive win.
“Not immediately,” Tan said honestly. “Your relationship with TSMC is solid, and our manufacturing process isn’t quite at their level yet. But in the partnership agreement, we include pilot programs. You send us test designs for 18A evaluation. If we hit quality and yield targets, you consider us for future products.”
“And if you don’t hit targets?”
“Then we still have the co-designed products. You still get U.S. manufacturing capacity. It’s not all-or-nothing.”
Huang had taken the proposal to his board. Nvidia’s directors were skeptical – Intel was a competitor, after all, and a struggling one. But the strategic logic was sound. Nvidia needed supply chain diversification. And frankly, having Intel focused on CPUs while Nvidia dominated GPUs was better than having Intel try to compete directly in accelerators.
“How much equity are we talking about?” Huang finally asked.
“Five billion dollars. At current market price, that’s roughly 4-5% ownership.”
Huang whistled. “That’s a big bet on Intel’s recovery.”
“Think of it as insurance,” Tan said. “If Intel fails, you lose five billion. If Intel succeeds in becoming a viable foundry alternative, you gain a strategic partner and a hedge against TSMC dependency. The risk-reward is favorable.”
By late August, the deal was done.
September 18, 2025: The Announcement
The joint press conference was held simultaneously in Santa Clara and Nvidia’s Santa Clara offices – the two companies were literally neighbors in Silicon Valley.
Tan and Huang appeared on a split screen, their images side by side on massive displays.
“Today marks a historic collaboration,” Huang began. “Nvidia and Intel are partnering to develop next-generation AI infrastructure and personal computing products that will power the future of technology.”
Tan continued: “This partnership combines Nvidia’s leadership in AI acceleration with Intel’s strengths in CPU technology and American manufacturing. Together, we’re building the foundation for the next era of computing.”
The details rolled out:
- Nvidia investing $5 billion in Intel common stock
- Intel designing custom x86 CPUs with NVLink for Nvidia’s AI platforms
- Intel building PC processors integrating Nvidia RTX technology
- Pilot programs for Intel Foundry evaluation
The market went crazy. Intel’s stock surged 22.8% – its best single-day gain since October 1987. By market close, Intel’s value had increased by $24 billion, more than the combined investment from the government and Nvidia.
Analysts scrambled to update their models.
“This changes the game for Intel,” a Morgan Stanley analyst wrote. “They’ve secured capital, customer commitment, and most importantly, validation from the AI leader. If Intel can execute, this partnership provides a path to relevance.”
But skeptics remained. “Where’s the foundry commitment?” a Goldman Sachs note asked. “Nvidia is buying equity and getting custom CPUs, but they’re not committing to manufacture with Intel. Without external foundry customers, Intel’s manufacturing business remains a cash drain.”
Inside Intel: The Reaction
The deals transformed morale overnight.
Engineers who’d been polishing their resumes suddenly felt hope. The government investment meant job security – at least for a while. The Nvidia partnership meant prestigious work on cutting-edge AI systems.
But there were also concerns.
“We’re building chips for our competitor,” one engineer complained in an internal forum. “Nvidia is eating our lunch in AI, and now we’re helping them?”
“We’re not competing with Nvidia anymore,” another responded. “That battle is over. They won. This partnership at least keeps us relevant.”
Michelle Holthaus, notably, was absent from all the celebration events. She’d taken the week off, citing “personal reasons.” But everyone knew what was happening.
She was negotiating her exit.
Project Paranoia: First Blood
While the corporate drama played out publicly, Project Paranoia was making progress in secret.
Elena Rodriguez, the engineer Tan had recruited from Nvidia, was leading the architecture team. They’d spent three months in intensive design sessions, exploring radical approaches to processor design.
The breakthrough came on a Tuesday afternoon in mid-September.
“I think we’ve got something,” Elena told her team, pulling up a simulation on the main display.
What they’d developed was unlike any chip Intel had ever built. Instead of a traditional CPU or a GPU-style accelerator, it was a hybrid architecture that could dynamically reconfigure itself based on workload.
For standard computing tasks, it operated like a high-performance CPU. For AI inference, it morphed into a specialized accelerator more efficient than GPUs. For certain types of scientific computing, it could even outperform quantum computers on specific problems.
The simulation showed performance gains of 3-5x over current top-of-the-line processors, while using 40% less power.
“This is incredible,” one engineer breathed. “If we can build this – “
“If,” Elena cautioned. “The simulation is one thing. Actually manufacturing this at scale is something else entirely.”
She sent the preliminary results to Tan with a subject line: “Project Paranoia – Phase 1 Complete. You need to see this.”
The Price of Success
The same day Elena’s email arrived, Tan received another message. This one from Michelle Holthaus:
Lip-Bu,
After considerable thought, I’ve decided to resign from Intel effective October 31st. I believe in this company and its mission, but I no longer believe in the strategic direction we’re taking.
Intel was built by designing great products. Under your leadership, we’re becoming a contract manufacturer for other companies’ visions. That’s not the Intel I want to be part of.
I wish you success, but I cannot in good conscience lead a product organization that’s subordinate to manufacturing priorities and moonshot bets.
Michelle
Tan read the email three times. Part of him felt relief – Holthaus’s opposition had been a constant source of friction. But part of him felt loss. She represented institutional knowledge, deep relationships with customers, understanding of Intel’s product portfolio.
More concerning, her resignation would signal to the market that Intel’s senior leadership wasn’t aligned. It could trigger more departures.
He called Yeary.
“Michelle’s resigning,” Tan said without preamble.
A long sigh. “When?”
“End of October. She sent a gracious email, but the message is clear – she disagrees with our strategy.”
“Can we change her mind?”
“No. And honestly, I’m not sure we should try. Having a senior executive who fundamentally opposes what we’re doing creates more problems than it solves.”
“The optics are terrible,” Yeary said. “She was interim co-CEO. Her leaving looks like the ship is sinking.”
“Then we need to control the narrative,” Tan said. “We announce it simultaneously with something positive. A major customer win, a technology breakthrough, something that shows momentum despite the leadership change.”
“Do we have anything?”
Tan thought about Elena’s email sitting in his inbox.
“Maybe. Give me forty-eight hours.”
The Apple Gambit
What Tan didn’t tell Yeary – what he couldn’t tell anyone yet – was that he’d been working on the most audacious play of all.
For the past six weeks, he’d been in quiet discussions with Apple.
The pitch was simple: let Intel manufacture Apple Silicon in America using Intel’s 18A process.
Apple had been using TSMC exclusively for its M-series processors. These chips powered MacBooks, iPads, and increasingly, desktop Macs. They were critical to Apple’s vertical integration strategy.
But Apple’s relationship with TSMC came with risk. All of TSMC’s advanced fabs were in Taiwan. If China invaded, if a major earthquake hit, if political instability disrupted production – Apple’s entire product line could grind to halt.
Intel represented geographic diversification. American manufacturing. And with 18A, Intel’s technology was finally approaching TSMC’s quality level.
“We’re not ready to commit to production,” Apple’s VP of hardware engineering had told Tan during their last meeting. “But we’re open to a pilot program. Send us your 18A specifications, prove you can hit our quality standards, and we’ll consider test runs for low-volume products.”
It wasn’t a full commitment. But it was a foot in the door – the first time Apple had seriously considered Intel manufacturing since the companies’ relationship had soured over chip delays in the 2010s.
Tan authorized the pilot program immediately. If Project Paranoia could deliver breakthrough architecture, and if Intel Foundry could prove it could manufacture complex chips to Apple’s standards, the combination would be revolutionary.
Intel would have its moonshot technology and its marquee foundry customer.
But two massive “ifs” stood between today and that outcome.
Late September: The Reckoning
As September ended, Tan took stock of where Intel stood:
Assets:
- $14 billion in new capital (government + Nvidia)
- Partnership with the AI market leader
- Breakthrough chip design in Project Paranoia
- Potential Apple pilot program
- Stock price up 77% year-to-date
Liabilities:
- Still losing $2-3 billion per quarter
- No major foundry customer commitments
- Senior leadership departures creating instability
- Project Paranoia unproven in silicon
- 18A process still not at full yield
The balance sheet looked better. The strategic position looked more promising. But Intel was still in crisis mode.
One evening, alone in his office, Tan pulled out a notebook and wrote a question:
“What does success look like?”
He’d been so focused on survival – on securing capital, making deals, launching projects – that he hadn’t clearly defined victory.
If Intel was still alive in five years but had become a contract manufacturer for other companies’ designs, was that success?
If Intel had the best chip technology but couldn’t manufacture it profitably, was that success?
If Intel survived by becoming partially state-owned, constrained by government priorities, was that success?
Grove’s ghost seemed to whisper in the room: “Success isn’t survival. Success is dominance. Either you’re the best, or you’re dying slowly.”
But Intel couldn’t be the best at everything anymore. The company had to choose: best at design, or best at manufacturing, or best at some new category they’d create.
The impossible choice hadn’t gone away. It had just evolved into an impossible execution challenge.
Epilogue: The Phone Call
On October 1st, Tan received a call that would alter everything.
“Mr. Tan, this is Tim Cook.”
The Apple CEO’s voice was unmistakable – calm, measured, revealing nothing.
“Tim, good to hear from you.”
“I’ve reviewed your 18A specifications and the pilot program proposal. I have some questions.”
They talked for forty-five minutes, ranging across manufacturing yields, defect rates, capacity planning, and supply chain resilience. Cook was methodical, drilling into details, probing for weaknesses.
Finally, he said: “Here’s what I’m prepared to do. We’ll commit to a pilot run of 10,000 units for an upcoming product – not a flagship, but not insignificant. If you hit our quality targets, we’ll discuss larger volumes.”
“What product?” Tan asked.
“That’s confidential for now. But it’s a 2026 launch, which means you have six months to prove your process works.”
Six months. It was an impossibly tight timeline. But it was also the opportunity Intel had been chasing for a decade.
“We’ll deliver,” Tan said.
After the call ended, Tan sat motionless for five minutes, feeling the weight of what he’d just committed to.
If Intel succeeded with Apple, the company’s foundry business would be validated. Other customers would follow. The government’s investment would be justified. Project Paranoia’s breakthroughs would have a manufacturing platform.
If Intel failed, Apple would never trust them again. The foundry business would be dead. And the whole elaborate strategy would collapse.
Everything hinged on the next six months.
No pressure.
NEXT TIME ON THE LAST STAND:
October 2025. With Holthaus’s departure creating a leadership void and the Apple deadline looming, internal tensions explode. A group of veteran engineers sends an anonymous letter to the board accusing Tan of “killing Intel to save it.” The best talent continues fleeing to competitors. Project Paranoia hits a catastrophic setback when initial silicon tests reveal fatal flaws. And Tan discovers that someone inside Intel has been leaking confidential information to the press – perhaps even to competitors. Trust is breaking down. Time is running out. And the question everyone is asking: Is Tan Intel’s savior, or is he leading the company to its final destruction?
The mutiny begins…
Continue to Episode 6: “The Mutiny” to witness Intel’s internal crisis reach the breaking point…
Strategic Fictional Analysis Disclaimer: This narrative combines documented facts (U.S. government’s August 2025 $8.9B investment, Nvidia’s September 2025 $5B investment, stock performance data) with fictionalized dramatization of private negotiations, internal conflicts, and strategic discussions. Characters’ personal decisions, Project Paranoia developments, Apple discussions, and specific conversations are creative interpretations designed to explore the corporate drama and strategic challenges facing Intel. Dialogue and internal dynamics are fictional constructs meant to illustrate the complex pressures of executing a turnaround under intense scrutiny.
