Big Tech is still cheap, Samsung Securities says as Magnificent 7 valuations sink below 3-year average

Valuations for the US market-leading Magnificent 7 have fallen well below their three-year average, weighed down by massive capital expenditure, high interest rates and growing relative-valuation pressure from AI frontier companies. Samsung Securities said "Big Tech is clearly cheap right now."

In a recent report titled "Big Tech: A Brutal Relative-Valuation Season," Samsung Securities said the 12-month forward price-to-earnings ratio for the M7 — excluding Tesla — stood at 18.8 times, far below the three-year average of 24.5 times. "Big Tech's dilemma right now is not about direction but about speed," the brokerage said.

The M7 has underperformed the S&P 500 by 13 percentage points this year. Despite the tailwind of AI industry growth, the market has focused more on the aggressive AI investment costs Big Tech companies are incurring, the report said.

The biggest drag is the scale of AI infrastructure spending. Hyperscalers have continued to pour money into large data centers and semiconductors to meet surging enterprise AI demand. Some companies have raised funds through corporate bond issuances and rights offerings, fueling market concerns about when those investments will generate returns.

Kim Jung-han, an analyst at Samsung Securities, said "Big Tech is currently a company that spends money, while chipmakers are companies that receive it." In fact, the average operating profit of the three major memory chipmakers — Samsung Electronics, SK hynix and Micron — is forecast to surpass the M7 average starting this year.

The rise of AI frontier companies adds another variable. OpenAI and Anthropic are growing their sales and valuations far faster than established Big Tech firms, while SpaceX holds a near-monopoly on the emerging space industry as a new growth axis.

A wave of large initial public offerings, starting with SpaceX, also poses a risk. Institutional investors may trim their existing Big Tech holdings to make room for newly listed stocks. If OpenAI and Anthropic eventually go public as well, the supply-demand pressure could intensify further.

Even so, Samsung Securities emphasized the relative appeal of the M7. Earnings, unlike share prices, have remained solid.

Google and Amazon both beat market expectations in the first quarter of this year, and cloud demand is rising rapidly as enterprise AI adoption spreads. The cloud backlog of the three major hyperscalers is growing far faster than their capital expenditure, supporting the case for further earnings improvement ahead.

Valuation attractiveness has also increased. The average forward PER for the M7 — excluding Tesla — fell to around 18 times, near last year's trough, before recovering to the current 18.8 times. That is below both the three-year average of 24.5 times and last year's November trough of 22.9 times.

Kim said that "even amid concerns about structural de-rating, the probability of the PER falling significantly below 18 times due to a macro shock is limited when enterprise AI momentum is taken into account," adding that the group "has entered a phase where undervaluation appeal is becoming more prominent."

Among individual stocks, Samsung Securities was positive on Alphabet. Google Cloud's backlog growth rate is significantly outpacing its sales growth rate, and generative AI model Gemini is rapidly expanding across existing platforms including Search, YouTube and Workspace.

The brokerage also had a favorable outlook on Microsoft, Amazon and Meta. Microsoft is reducing its dependence on OpenAI and focusing on developing its own AI models and upgrading Copilot, while Amazon is strengthening its cloud competitiveness on the back of its proprietary AI chip Trainium. Meta is pursuing a strategy to expand its advertising-driven revenue model into subscription services and enterprise AI solutions.

"Things are confusing right now, but when AI infrastructure investment is justified by earnings, a swift market rerating is highly likely," Kim said.