Nvidia’s Optical Gambit: Why the Chip Giant Is Betting Billions on Glass, Not Silicon

The semiconductor industry’s most consequential deal this week wasn’t about a faster chip or a smaller transistor. It was about glass.

Nvidia’s strategic partnership with Corning — involving a $500 million equity stake, warrants for 15 million shares at $180 each, and the construction of three new manufacturing plants in North Carolina and Texas — signals a fundamental shift in how the company thinks about AI infrastructure. The bottleneck in next-generation data centers, Nvidia has concluded, isn’t compute power. It’s connectivity.

The math is straightforward. Inside Nvidia’s rack-scale systems, thousands of copper cables currently shuttle data between GPUs. Replacing them with optical fiber — a technology called co-packaged optics — could slash energy consumption by a factor of five to twenty. Corning will increase its US optical manufacturing capacity tenfold to meet the demand, creating at least 3,000 jobs in the process.

The market responded with a 5.6% single-day gain. Nvidia shares now trade at €177.24, roughly 10% above their 50-day moving average and within striking distance of a 52-week high. Of the 37 analysts covering the stock, the consensus is a “Strong Buy” with a 12-month price target equivalent to around $270 — implying roughly 30% upside from current levels.

Should investors sell immediately? Or is it worth buying Nvidia?

The Earnings Test That Looms

For all the excitement around the Corning deal, the real moment of truth arrives on May 20, when Nvidia reports its first fiscal quarter of 2027. The consensus estimate calls for $78.8 billion in revenue — a year-over-year increase of nearly 79%. Analysts believe the company needs to clear the 80% growth threshold to sustain the current rally.

The bull case is well-rehearsed. Meta plans up to $145 billion in capital expenditures. Microsoft is targeting roughly $190 billion. Nvidia and AMD remain the direct beneficiaries of this investment wave. Rosenblatt holds the highest price target on Nvidia at $325, while Cantor Fitzgerald and Bernstein both sit at $300. Not a single analyst has downgraded the stock recently.

But risks persist. US export restrictions have effectively cut Nvidia off from the Chinese market, which was once a significant growth driver. And the company’s massive investment in optical infrastructure — including $4 billion already deployed into laser-component makers Coherent and Lumentum in March — represents a bet that could take years to pay off.

The Broader Chip Landscape

Nvidia’s optical pivot is unfolding against a backdrop of extraordinary activity across the semiconductor sector. AMD delivered a record first quarter with $10.3 billion in revenue and adjusted earnings per share of $1.37, sending its stock above €358.45 — a new all-time high. The data center business grew 57%, and the company’s “Helios” rack system, designed to compete directly with Nvidia’s Blackwell and Vera-Rubin platforms, has already attracted orders from OpenAI and Meta.

Intel, meanwhile, closed at €96.07 — a 52-week high — fueled by reports that Apple is in talks with both Intel and Samsung about US-based chip production. Apple has historically relied exclusively on TSMC for its processors, but concerns about supply chain concentration have opened the door for alternatives. Analysts suggest Intel could begin manufacturing simpler M-series chips for Macs and iPads as early as 2027, with iPhone Pro models following no sooner than 2028.

The caveat is substantial. Intel’s foundry business posted a net loss of $3.7 billion in the first quarter of 2026. CEO Lip-Bu Tan has restructured operations, expanded collaboration with Google, and closed a joint venture with Elon Musk’s Terafab project — but the division remains deeply unprofitable. Taking on complex manufacturing orders from Apple carries execution risk as long as utilization rates remain suboptimal.

Infineon’s Surprise Upgrade

Infineon added to the sector’s momentum with an unexpected guidance raise. Management now expects “significant” revenue growth to over €16 billion for the fiscal year, up from roughly €14.7 billion last year. The segment margin is projected at around 20%.

The driver is AI-related power supply revenue, which Infineon expects to hit roughly €1.5 billion this fiscal year — constrained not by demand but by available capacity. For fiscal 2027, the target rises to €2.5 billion. The company is increasing investments by €500 million to €2.7 billion in 2026.

Perhaps more surprising: order intake from the automotive industry has picked up. Production lines are running at capacity, an unusual signal in a segment that had recently been characterized by caution. Infineon shares trade at €59.60, up roughly 56% year-to-date, though the price-to-earnings ratio of nearly 74 leaves little room for error.

TSMC Between Expansion and Uncertainty

TSMC is pursuing the most ambitious capacity expansion in its history — and simultaneously confronting an unfamiliar question about its relationship with Apple.

The Bloomberg report about Apple’s discussions with Intel and Samsung briefly knocked TSMC’s stock down 2%, though it recovered to close at €356.50, a new 52-week high. Apple itself appears to have reservations about the manufacturing quality of alternative partners, and a complete shift away from TSMC is considered unlikely.

Nvidia at a turning point? This analysis reveals what investors need to know now.

The investment numbers are staggering. TSMC is upgrading Fab 15A in the Central Taiwan Science Park from 28nm to 4nm production, with total investment exceeding NT$100 billion. Its 1.4nm fab in Taichung is ahead of schedule, with test production potentially beginning in the third quarter of 2027. The capital expenditure budget for 2026 stands at $52 billion to $56 billion. Some older equipment is being relocated to Dresden, where mass production for automotive and industrial chips is slated to begin in 2027.

In the first quarter of 2026, TSMC posted $35.9 billion in revenue with a gross margin of 66.2%. Second-quarter guidance ranges from $39 billion to $40.2 billion.

The Infrastructure Arms Race

What unites these developments is a recognition that the next phase of the AI cycle will be defined by physical infrastructure — not just chip design. The combined capital expenditure budgets of major tech companies are on track to exceed $700 billion in 2026, up from roughly $410 billion last year. The bond market has emerged as the primary financing vehicle.

Nvidia is shifting the bottleneck from chips to connectivity. AMD is positioning itself as a credible alternative in both server CPUs and AI accelerators. Intel is betting that geopolitical pressure will force Apple and others to diversify their manufacturing base. Infineon is capitalizing on the power demands of AI infrastructure. And TSMC is building capacity for a demand wave that, by its own projections, remains years from its peak.

The next 90 days will test whether the current momentum is sustainable. Nvidia’s earnings on May 20 is the sector’s next major sentiment check. AMD must begin Helios deliveries on schedule and scale its MI450 supply chain. For Intel, the coming weeks will determine whether Apple talks translate into a concrete pilot program — or fizzle out. And across the industry, one question lingers: whether the monetization of AI can keep pace with the spending required to build it.

Ad

Nvidia Stock: New Analysis - 7 May

Fresh Nvidia information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.

Read our updated Nvidia analysis...

en | US67066G1040 | NVIDIA’S | boerse | 69287382 |