Semiconductor Stocks Are Surging: What Comes Next

The Hardware Rally Nobody Saw Coming — Except the Data

Semiconductor stocks don't move 15% in a single session without a reason. SOXL, the 3x leveraged semiconductor ETF, did exactly that recently — and it wasn't a fluke. It was the payoff from a confluence of catalysts that had been building for weeks: a confirmed Intel-Apple chip deal, accelerating AI infrastructure spending, a broadly supportive labour market, and a market full of investors who were badly mis-positioned and suddenly needed somewhere to run.

If you were watching the signals — ADP weekly job data, CPU pricing trends, call option volumes, relative strength indicators — the hardware trade was readable. The question now isn't whether the move happened. It's whether the rally in semiconductor stocks has legs, which names still offer value, and what the next rotation looks like when software finally catches up.

Here's the full picture, by the numbers.

---

Why Semiconductor Stocks Broke Out When They Did

The setup for the semiconductor rally came together across several independent data points, all pointing in the same direction.

The Intel-Apple deal was a near-term catalyst. Confirmation that Apple would use Intel chips in upcoming hardware gave markets a concrete, revenue-visible reason to reprice semiconductor names upward. Deals like this have a multiplier effect: they validate domestic chip manufacturing demand, reduce perceived revenue uncertainty, and signal that the broader chip ecosystem is healthier than the pessimists argued.

AI infrastructure spending is a structural driver, not a cyclical one. Data centre construction tied to artificial intelligence workloads is accelerating. That means physical demand for CPUs, GPUs, and the full hardware stack underneath AI — cooling systems, power infrastructure, networking equipment. ADP payroll data had been tracking a steady rise in construction employment over February, March, and April, a leading indicator that picks up physical build-out activity before it shows in corporate earnings reports.

Investors were mis-positioned. Cash on the sidelines was high. Sentiment was cautious. Relative strength on the QQQ ETF had been climbing since late April, crossing 70 on the weekly chart and reaching 82.5 on the daily — levels that indicate strong momentum but also demand a sharper eye on sustainability. When everyone is braced for bad news and it doesn't arrive, the repricing is fast and often overshoots.

Call option activity accelerated the move. Index-level call option buying reached multi-year highs, creating a self-reinforcing dynamic where dealers hedge their exposure by buying the underlying assets, which pushes prices higher, which drives more call buying. This flywheel is powerful in the short term. It is not a thesis on its own, but it amplifies directional moves that already have fundamental support.

---

The Labour Market: Supportive, But Worth Watching Closely

The latest Bureau of Labour Statistics report delivered a payrolls number of 115,000 — comfortably above the 80,000–90,000 level that economists describe as break-even, and above consensus expectations. On the surface, that's a clean beat. Dig one level deeper and there are a few nuances that matter for investors.

Average hourly earnings came in at 0.2% month-over-month versus the 0.3% expected. That's not alarming. Wage growth in the 0.2%–0.3% range supports consumer spending without being meaningfully inflationary. The Fed doesn't have a strong case for rate hikes based on this print alone, and markets are right to price in continued pause rather than tightening.

Labour force participation ticked down to 61.8% from an expected 61.9%. This is the number to monitor. When participation falls rather than rises, strong unemployment figures can be statistically flattering without reflecting genuine labour market improvement. If participation reverses upward — as it could if federal government layoffs push more people to actively seek work — the unemployment rate could rise faster than headline payroll numbers suggest.

Federal government and information sector jobs are where the losses are concentrated. Retail, transport, warehousing, and healthcare are picking up the slack. Construction added jobs, driven in part by AI-related data centre builds. The sectoral composition of employment growth is increasingly tilted toward physical infrastructure — which reinforces the hardware thesis.

Bottom line: this labour market is not sending a recession signal, but it's not as clean as the headline number implies. The ADP trend — which has been rising consistently and often leads BLS data — remains the more reliable forward indicator to track weekly.

---

Nvidia vs AMD: Which Semiconductor Stock Is Actually Cheaper?

This is where the analysis gets counterintuitive, and where most investors get it wrong.

Nvidia is the name everyone knows. It has dominated the GPU market for AI training workloads. But on a PEG ratio basis — price-to-earnings divided by forecast earnings growth — Nvidia currently sits at 1.91. That's not expensive for a company of its quality, but it's not the screaming value some assume given the recent price action.

AMD's PEG ratio sits at approximately 1.52, with analysts projecting average earnings growth of 38.6% annually over the next four years. The balance sheet is solid, pricing power is intact, and the stock spent much of early 2025 trading as though it had no path to relevance in the AI era — which was always an overcorrection.

Nvidia's own valuation case is supported by a different angle: Bank of America's analysis shows it trading roughly 10 times below its three-year median P/E ratio of 32. That compression reflects market scepticism about whether growth is plateauing. The counter-argument — and it's a strong one — is that inference demand is now eclipsing training demand as companies move from building AI models to running them at scale. Nvidia's CPU-on-rack systems are positioned directly in that inference expansion, and Jensen Huang has been consistent in arguing that both GPU and CPU demand can grow simultaneously rather than cannibalising each other.

For investors comparing the two:

• AMD offers the better valuation on a pure PEG basis and has more room for multiple expansion from current levels
• Nvidia offers more defensible competitive positioning and is a direct play on inference infrastructure growth
• Both are part of the same macro trade: AI infrastructure buildout is nowhere near complete

---

What the Next Leg of the Rally Looks Like

Hardware led. Software follows. That sequencing matters for portfolio positioning right now.

The current S&P 500 rally has been concentrated. Technology — specifically semiconductor and hardware names — has driven roughly 70% of the index gain. Market breadth is narrow by historical standards. That concentration creates both a risk and an opportunity.

The risk: if hardware names pull back on profit-taking, index-level losses can be sharp because there's no broad base of winners to cushion the drawdown.

The opportunity: when software and enterprise tech begin to catch up — which historically happens three to six months after hardware starts moving — market breadth expands. More stocks participate. Index gains accelerate even as individual hardware stocks consolidate. Companies that are best positioned to monetise AI operationally — advertising platforms, vertical SaaS businesses, government technology vendors — tend to be the beneficiaries of that second leg.

Names worth monitoring in that context include digital advertising platforms with large AI-driven targeting capabilities, business software companies integrating AI into core workflows, and defence and public sector technology vendors where AI adoption is accelerating. The common thread: companies where AI is already generating measurable revenue impact, not just a future optionality story.

The broader de-globalisation trend adds another layer. Short-term, domestic manufacturing investment and infrastructure spending are stimulative — more jobs, more construction, more hardware demand. Longer-term, the efficiency losses from reduced global trade integration are real and will eventually weigh on productivity growth. The early 2030s could see a reversal as global supply chains re-integrate, potentially triggering a deflationary cycle. That's a decade-level consideration, not a 2025 trade, but it shapes the structural backdrop.

---

Practical Takeaways for Investors Right Now

Cut through the noise. Here's what the data actually implies for your portfolio:

• Semiconductor stocks are not cheap in absolute terms, but AMD remains undervalued on a forward growth basis. A PEG of 1.52 with 38%+ projected earnings growth is a better risk-reward than most assets available at comparable quality.
• Nvidia's $320 price target implies meaningful upside from current levels if inference demand continues to scale, which the data suggests it will. Bank of America's $250 target looks conservative by comparison.
• The labour market supports equities but doesn't justify complacency. Watch the labour force participation rate monthly. If it reverses upward, the unemployment picture changes quickly.
• Don't chase the leveraged ETF momentum blindly. SOXL at +15% in a session is a signal of how fast the move has come, not an invitation to add at the top. RSI above 80 on the daily chart means you're buying into overbought conditions — useful for traders, dangerous for longer-term holders who don't have an exit plan.
• Start building a watchlist for the software rotation. It may be three to four months away. When it arrives, it moves fast.
• Geopolitical stability — real or perceived — remains a buy-the-dip signal for risk assets. Strait of Hormuz developments, trade war escalations, and similar shocks have consistently resolved as temporary headwinds when the underlying economic data is sound.

---

Conclusion: The Hardware Trade Has Worked — Now Position for What's Next

The semiconductor rally of the past six weeks was not luck. It was the outcome of a readable set of signals: ADP data trending up, AI infrastructure demand accelerating, investors under-positioned in hardware, and a market that needed a catalyst to redeploy cash. Those signals were there. The move delivered.

From here, the next 90 days are about two things. First, assessing whether hardware names can hold their gains or need to consolidate before the next leg. Second, identifying the software and enterprise AI names that will benefit from the broader market rotation that typically follows hardware leadership.

The numbers support staying invested in semiconductors at a selective level — AMD for valuation, Nvidia for infrastructure positioning — while beginning to rotate attention toward AI monetisation plays in software and services. The rally is not over. It's evolving.

---

Frequently Asked Questions

Why did semiconductor stocks surge so sharply in recent weeks?

Several catalysts converged simultaneously: the confirmed Intel-Apple chip deal, accelerating AI data centre construction driving CPU and GPU demand, a better-than-expected US payrolls report, and a large volume of call option buying at the index level. The options activity in particular created a mechanical flywheel effect where dealer hedging amplified price moves that were already directionally supported by fundamentals.

Is AMD actually cheaper than Nvidia right now?

On a PEG (price-to-earnings growth) ratio basis, yes. AMD trades at approximately 1.52x PEG with projected annual earnings growth of around 38.6% over the next four years, compared to Nvidia's PEG of approximately 1.91. Nvidia has stronger competitive moats and direct inference infrastructure exposure, but AMD offers a better entry point for investors focused purely on valuation relative to growth.

What does the labour market data mean for Federal Reserve policy?

The 115,000 payrolls print was above break-even, but average hourly earnings came in at 0.2% month-over-month — below the 0.3% expected. That combination gives the Fed little reason to raise rates. The most likely outcome is continued pause. The labour force participation rate declining to 61.8% is a watch item: if participation rebounds, unemployment could rise quickly, which would change the macro picture for equities.

What is the next phase of the market rally likely to look like?

Hardware and semiconductors have led the current rally, accounting for a disproportionate share of index gains. Historically, software and enterprise technology names begin to participate three to six months after hardware leadership establishes itself. When that rotation occurs, market breadth expands and index-level gains can accelerate even as individual hardware stocks consolidate. Companies best positioned to monetise AI at the application layer — advertising technology, business software, government services — are the names to start researching now.

Should investors be concerned about overbought RSI readings on semiconductor stocks?

Relative strength above 70 is a caution flag, not a sell signal on its own. RSI at 82.5 on the daily chart for QQQ means the short-term move has been extreme and consolidation or a modest pullback is statistically likely. For long-term investors with a 12-to-24-month horizon and clear position sizing, this matters less than for short-term traders. The underlying demand drivers — AI infrastructure, domestic chip manufacturing, inference expansion — remain intact regardless of near-term momentum readings.