Magnificent 7 in 2026: Which One Is Actually Cheap?

The "Magnificent 7" label has become shorthand for a category: expensive AI stocks that have driven most of the market's returns for the past three years.

The data disagrees with that framing — sharply.

Right now, one Magnificent 7 stock has a PEG ratio of 0.91. Another trades at a three-year P/E low with 55%+ analyst upside. A third has a $126B subsidiary that almost nobody is modeling. And one trades at 366x trailing earnings while its core business is actively shrinking.

These are not the same kind of stock. Calling them all "expensive AI plays" is a category error. Here's what the valuation data actually shows.

All figures as of mid-March 2026. Sources: company earnings releases, StockAnalysis, MacroTrends, MarketBeat.

The Master Table

| | NVDA | META | MSFT | GOOGL | AMZN | AAPL | TSLA | |---|---|---|---|---|---|---|---| | Price | ~$179 | ~$607 | ~$382 | ~$301 | ~$209 | ~$248 | ~$405 | | Trailing P/E | ~36x | ~22–26x | ~24–26x | ~28x | ~30x | ~32x | ~366x | | Forward P/E | ~21.5x | ~20x | ~22x | ~23–26x | ~28x | ~29x | ~172x | | PEG Ratio | 0.57 | 0.91 | 1.61 | 1.70 | 1.49 | 2.76 | 3.93 | | EV/EBITDA | ~32x | ~15–16x | ~16x | ~24x | ~16x | ~24x | ~119–129x | | Revenue Growth | +65% | +22% | +17% | +15% | +12% | +10% | -3% | | Operating Margin | ~55% | 41% | 47% | ~32% | ~12% | ~31% | ~7% | | Analyst Upside | — | +40–54% | +55–58% | +17–22% | +34–37% | +16–19% | ~0% |

Read this table once and two things jump out immediately.

First: Apple has the highest P/E in the Mag7 (except TSLA) and the slowest revenue growth. You are paying the most, per unit of growth, for the company growing the least.

Second: Meta has the lowest forward P/E, the second-lowest PEG, the highest-growing revenue among the profitable companies, and an all-time record operating margin of 41.4%. Yet it sits at ~$607 — below what its earnings trajectory would suggest.

These facts don't coexist by accident. The market has reasons for both. Let's work through each stock.

Ranked by PEG Ratio: The Growth-Adjusted View

The PEG ratio divides P/E by earnings growth rate. Peter Lynch's rule of thumb: below 1.0 = you may be getting growth for free. Above 2.0 = you're paying a premium.

🥇 NVIDIA (NVDA) — PEG 0.57

Revenue grew 65% in FY2026 to $215.9B. Net income grew 65% to $120B. Data Center revenue alone was $209B for the year. These are historically unprecedented numbers at this scale.

Yet the stock is $179 — down 11% from its $207 October 2025 high. At a PEG of 0.57 and a forward P/E of ~21.5x, the market is pricing in meaningful growth deceleration.

The central question: Can a company sustain 30%+ growth when it already has $215B in revenue? If yes, NVDA is cheap. If revenue growth falls to 15–20%, forward estimates come down and the multiple looks expensive quickly.

Verdict: Fairly valued with upside if FY2027 growth holds above 30%. The PEG is mathematically compelling; the base effect risk is real.

🥈 META PLATFORMS (META) — PEG 0.91

$200.97B in FY2025 revenue (+22%). Operating margin 41.4% — an all-time record. Forward P/E of ~20x. PEG below 1.0.

The discount exists entirely because of the $125B 2026 capex commitment — nearly double 2025's spend. The market is pricing in capex ROI uncertainty. CFO Susan Li explicitly stated 2026 operating income will exceed 2025's despite the spend, which the Family of Apps cash generation supports.

The AI story inside Meta that doesn't get told: Andromeda, Meta's AI ad delivery engine (fully deployed October 2025), delivers 20–35% higher ROAS for advertisers. That's why margins expanded from 35.7% → 41.4% in one year.

Verdict: Undervalued relative to earnings power. A PEG below 1.0 on a 22% grower with 41% operating margins is a rare signal. The risk is the $125B bet; the current price assumes it fails.

🥉 AMAZON (AMZN) — PEG 1.49

$716.9B in FY2025 revenue (+12.4%). AWS grew 24% in Q4 to $24.2B. Operating leverage is real: operating income grew faster than revenue. Analyst consensus ~$280–287, implying 34–37% upside from ~$209.

The EV/FCF of ~301x looks alarming — it's almost entirely a capex artifact. Amazon is spending ~$200B/year on infrastructure, which compresses current FCF to near-zero. EV/EBITDA of ~16x is the better measure of earning power.

Verdict: Fairly valued. AWS margin expansion and advertising growth (60B+ run rate) are real. At 28x forward P/E for 12% revenue growth, it's not cheap — but the operating leverage story has a clear runway.

4th. MICROSOFT (MSFT) — PEG 1.61

The most asymmetric setup in the Mag7. Stock is down 27% from its September 2025 high of $515, sitting at $382. During that same period: EPS grew 29%, Azure accelerated to 38–40% YoY growth, $625B contracted backlog locked in. 31 analysts cover MSFT; all 31 rate it Strong Buy; average target $592–603 = 55–58% implied upside.

A 24–26x P/E is the cheapest MSFT has been since late 2022 — before Azure re-acceleration, before Copilot, before the $625B backlog existed.

Verdict: Undervalued. The 27% drawdown was driven by capex concern, not fundamental deterioration. The backlog and Azure trajectory suggest the concern is misplaced.

5th. ALPHABET (GOOGL) — PEG 1.70

$403B in FY2025 revenue (+15%). Google Cloud +48% YoY — faster than Azure or AWS. YouTube >$60B annually. Record $164.7B operating cash flow. Waymo raised $16B at $126B valuation in February 2026 — an asset sitting inside GOOGL that almost no analyst models.

P/E of ~28x is near its 5-year median, but EPS grew +35% in CY2025 while the multiple barely moved. The DOJ antitrust default-search remedy is a real risk; the AI disruption narrative is not supported by data (Search +17% YoY in Q4 2025).

Verdict: Fairly valued to modestly undervalued. The Waymo asset alone ($126B, essentially unpriced) provides a margin of safety. Analyst consensus sees 17–22% upside.

6th. APPLE (AAPL) — PEG 2.76

The most expensive growth-adjusted valuation in the Mag7. $416.2B in FY2025 revenue (+6.4%), record Q1 FY2026 at $143.8B (+16%). 32x trailing P/E vs. a 10-year mean of 24x. Operating margin 31%.

The bull case: Services at 75.3% gross margin on a 2.35B device installed base is a compounding machine. The $90.7B/year buyback retires ~2.5% of shares annually, mechanically growing EPS.

The bear case: PEG of 2.76 means paying a steep premium for modest growth. The $15–20B Google search deal faces DOJ scrutiny. Apple Intelligence isn't monetized.

Verdict: Fairly valued. Quality at a full price. The risk/reward improves meaningfully below $235.

7th. TESLA (TSLA) — Traditional metrics don't apply

FY2025: Revenue -2.9% to $94.8B. Automotive revenue -10%. Net income -47%. GAAP EPS $1.08. Trailing P/E: 366x.

The bull case has nothing to do with the car business: January 22, 2026 — first unsupervised Robotaxi rides in Austin. February 18, 2026 — first Cybercab built. Energy storage +49% to 46.7 GWh ($12.8B revenue, +60% YoY). Optimus Gen 3 production before end of 2026.

Analyst range: $25.28 (GLJ Research, auto-only DCF) to $630 (Wedbush). A 25:1 ratio.

Verdict: Not a valuation play — a technology bet. At $405, the base case (Robotaxi works but slowly) is already priced in.

Three Clear Signals From the Data

Signal 1: META and MSFT are the value plays

Both have PEG ratios below 1.61 while growing EPS 20–29%. Both are trading at multi-year P/E lows while the underlying businesses are accelerating. Both have analyst consensus implying 40–58% upside. The market has discounted both for capex concerns; the underlying businesses are generating cash that funds the capex without margin destruction.

Signal 2: NVDA is cheap on a growth-adjusted basis, but everything depends on FY2027

PEG of 0.57 for a company with $215.9B in revenue growing 65% is a number with no precedent in modern market history. The question is whether it sustains. If FY2027 revenue grows 30%+, NVDA is cheap. If growth normalizes to 20%, the forward estimates get revised down and the multiple re-rates. This is the single most consequential growth-sustainability question in tech investing.

Signal 3: AAPL is the most expensive per unit of growth, TSLA is in a different category

Apple's 2.76 PEG — paying $2.76 for every percentage point of earnings growth — is the highest in the cohort. That premium reflects the Services re-rating thesis and buyback machine, both real. But it's worth being explicit: you are paying a 205% premium over Meta's PEG for a company growing at less than half the rate.

Tesla's metrics are decorative. You aren't valuing a company's earnings — you're assigning probability to a suite of transformative technologies.

How to Use This Analysis

If you're building a position: The PEG and forward P/E rankings suggest starting with META and MSFT for value, GOOGL for quality-at-reasonable-price. NVDA requires a view on FY2027 growth sustainability. AAPL and AMZN offer limited upside at current prices. TSLA requires a thesis, not a valuation.

If you're checking an existing position: The analyst consensus upside column is the quickest gut check. 55–58% on MSFT, 40–54% on META, 17–22% on GOOGL — these suggest where institutional money thinks the mispricing is largest.

If you want to stress-test your assumptions: Run any of these tickers through EvidInvest. Adjust the growth rate and discount rate to your own view. See how the fair value estimate changes. The standard DCF inputs are a starting point; the valuation is only as good as your assumptions.

Run a free Magnificent 7 comparison on EvidInvest →

Frequently Asked Questions

Which Magnificent 7 stock is the cheapest right now?

By PEG ratio: NVDA (0.57), then META (0.91). By forward P/E: META (~20x), then NVDA (~21.5x), then MSFT (~22x). By EV/EBITDA: META and MSFT (both ~15–16x). META and MSFT are the clearest value signals in the group by quantitative metrics.

Which Magnificent 7 stock has the most analyst upside?

MSFT: 55–58% upside to analyst consensus (~$592–603 target vs. ~$382 price). META: 40–54% upside (~$838–935 target vs. ~$607 price). These are the two stocks with the largest gap between current price and where analysts think they should trade.

Is Apple overvalued compared to the rest of the Magnificent 7?

On a growth-adjusted basis (PEG), yes — Apple has the highest PEG in the cohort at 2.76, despite having the slowest revenue growth. It's trading at a 33% P/E premium to Microsoft while growing at roughly 60% of Microsoft's rate. The premium reflects ecosystem quality and the buyback program, but it's the most expensive name per unit of growth.

Why is NVDA's PEG so low if the stock has already run so much?

Because earnings growth has kept pace with — or outrun — the stock price. FY2026 EPS was roughly $4.50–5.00+ (on net income of $120B / ~21.5B shares). At $179/share and growing EPS 65%, the PEG compresses rapidly. The risk is whether FY2027 earnings can sustain the growth rate that makes the PEG look cheap.

Should I treat the Magnificent 7 as one investment thesis?

No — and that's the central point of this analysis. META at 20x forward P/E and AAPL at 29x forward P/E are not the same kind of investment. TSLA at 366x trailing P/E is a technology option, not a value stock. MSFT at a 3-year P/E low is in a different risk category than NVDA at record revenues with uncertain deceleration ahead. The Mag7 label obscures meaningful valuation differences that should drive different position sizes and conviction levels.

Bottom Line

The Magnificent 7 are not a monolith. Right now:

Two are undervalued by earnings metrics: META and MSFT. Both at multi-year P/E lows, both growing 17–22%, both with 40–58% analyst upside.
Two are fairly valued with meaningful upside: NVDA and GOOGL. NVDA cheap on PEG if growth holds; GOOGL cheap on sum-of-parts given unmodeled Waymo value.
Two are fairly valued with limited near-term upside: AMZN and AAPL. Both solid businesses; both priced for reasonable base cases.
One is in its own category: TSLA. Not a valuation play.

Paying the same attention to all seven equally, as if they're the same bet, is the category error the data corrects.

Try EvidInvest free — run and compare any Mag7 stock in 60 seconds →