핵심요약

• AMD 포트폴리오는 총 $1.31억 달러 규모이며, 6개 포지션에 집중되어 있습니다.
• AMD는 AI 기반 항체 플랫폼인 Absci에 $66.2백만 달러를 투자했습니다.
• AMD는 AI 전략의 조립 라인인 Sanmina에 ZT Systems 서버 랙 통합 사업을 매각했습니다.
• Absci는 연초 대비 153.6%, 지난 1년간 212.7%의 주가 상승률을 기록했습니다.

도입

본 기사는 AMD가 단순한 반도체 시장 점유율을 넘어, AI 수요가 실제로 향하는 다운스트림 분야에 어떻게 자본을 배분하고 있는지를 분석합니다. 이는 투자자들이 AMD의 가치를 평가할 때, 칩 판매라는 단기적 관점을 넘어 AI 생태계 전체의 장기적 확장성을 고려해야 함을 시사합니다.

본문 1: AI 생태계 내 다운스트림 투자

AMD가 Absci에 $66.2백만 달러를 투자한 것은 주목할 만한 움직임입니다. Absci는 AMD가 제공하는 가속 컴퓨팅 자원을 집중적으로 소비하는 AI 기반 항체 플랫폼으로, 이는 AMD의 칩이 최종적으로 어떤 산업 응용 분야에 활용되는지를 명확히 보여줍니다. Absci는 연초 대비 153.6%, 지난 1년간 212.7%의 주가 상승률을 기록하며 강력한 시장의 관심을 받고 있습니다. 이는 AMD의 컴퓨팅 솔루션이 단순한 데이터센터 GPU 판매를 넘어, 고부가가치 바이오테크와 같은 전문 분야의 AI 수요를 견인할 수 있다는 점을 의미합니다. 즉, AMD의 가치 사슬이 칩 제조에서 최종 응용 분야로 확장되고 있음을 보여줍니다.

본문 2: 하드웨어 통합을 통한 공급망 전략

AMD가 ZT Systems의 서버 랙 통합 사업을 Sanmina에 매각한 것은 AI 전략의 물리적 기반을 확보하려는 의도를 반영합니다. Sanmina는 AI 전략의 조립 라인 역할을 하며, AMD가 제공하는 하드웨어 통합 역량을 활용하여 실제 AI 인프라의 물리적 기반을 구축하는 데 핵심적인 역할을 합니다. 이 거래는 AMD가 칩 설계 및 통합 기술을 넘어, AI 시스템의 실제 제조 및 공급망(Supply Chain)에 깊이 관여함으로써, 하드웨어와 소프트웨어의 경계를 허무는 통합 솔루션 제공자로 포지셔닝하려는 전략으로 해석됩니다. 이는 AI 인프라 구축에 필수적인 물리적 인프라를 선점하려는 움직임입니다.

본문 3: 장기적 관점과 리스크

AMD의 이러한 포트폴리오 움직임은 단기적인 반도체 사이클의 변동성보다 장기적인 AI 생태계 내에서의 위치 선점에 중점을 두고 있습니다. Absci와 Sanmina에 대한 투자는 AI 기술이 특정 산업(바이오테크, 제조)에 깊숙이 침투할 때, 하드웨어 제공업체가 단순한 공급자가 아닌 혁신의 파트너로서 자리매김하는 것이 미래 성장 동력이 될 것임을 시사합니다. 다만, 이러한 다운스트림 투자는 특정 산업의 임상 데이터나 제조 공정의 성공 여부에 따라 성과가 달라질 수 있으며, 이는 잠재적인 운영 리스크로 작용할 수 있습니다. 따라서 AMD는 하드웨어 성능과 더불어, 이들 파트너사의 기술적 성숙도와 시장 채택 속도를 면밀히 모니터링해야 합니다.

결론

결론적으로, AMD의 포트폴리오 공개는 AI 시대에 하드웨어 제공업체가 단순한 칩 공급자를 넘어, AI 수요가 실현되는 최종 응용 분야와 제조 인프라에 직접 투자함으로써 생태계 내에서 지배적인 위치를 확보하려는 전략을 명확히 보여줍니다. 향후 투자자들은 AMD가 제시하는 다운스트림 파트너십의 실질적인 성과와 이들 기업의 기술적 성숙도를 중점적으로 관찰해야 할 것입니다. 특히, Absci의 임상 데이터 결과와 Sanmina의 생산 능력 확장이 AMD의 장기적인 가치에 어떤 영향을 미칠지 주목할 필요가 있습니다.

---

원문 링크: https://247wallst.com/investing/2026/08/18/amds-1-3-billion-portfolio-reveals-its-biggest-investments-and-where-ai-demand-lands-next/?.tsrc=rss

AMD’s $1.3 Billion Portfolio Reveals Its Biggest Investments and Where AI Demand Lands Next

On August 14, 2026, Advanced Micro Devices ( NASDAQ:AMD | AMD Price Prediction ) filed its first-ever 13F, pulling back the curtain on a $1.31 billion equity portfolio concentrated in just six positions. SpaceX Class A shares dominate the book, with AMD’s $565.5 million stake accounting for 43.1% of the portfolio. Three holdings are private and inaccessible to retail investors. The other three are tradable, and two additional compute plays sit directly in the path of AMD’s capital. Here is where the money is actually moving.

Nobody expected an AI chip company to disclose a biotech position. Nevertheless, AMD’s 13F includes a $66.2 million stake in Absci, worth roughly 5.0% of the portfolio. Absci ( NASDAQ:ABSI ) runs an AI-driven de novo antibody platform that is a heavy consumer of the same accelerated compute AMD sells. If AMD is showing you where downstream AI demand actually lands, this is the tell most investors will miss.

The setup is early-stage, but the stock action is not. Absci is up 153.6% year to date and 212.7% over the past year, with a market cap around $1.51 billion. Analysts back the story: nine Buy or Strong Buy ratings against a single Hold, with a consensus target of $13.89. Cash and marketable securities sit at $201.1 million, which management says funds operations into H2 2028.

Absci CEO Sean McClain has already telegraphed the near-term catalyst: “We are fast approaching defining moments for Absci, with interim proof-of-concept data for pattern hair loss later this year.”

The next name on this list is where AMD did not just buy shares. It sold the company its crown jewel.

AMD sold its ZT Systems server rack integration business to Sanmina. That unit is now the physical backbone of the Helios rack rollout every hyperscaler is fighting to secure. AMD also held $291.3 million of Sanmina stock, 22.2% of the disclosed 13F. Sanmina ( NASDAQ:SANM ) is the AI strategy’s assembly line.

The June quarter turned that partnership into a printing press. Sanmina’s revenue hit $4.013 billion, up 102.3% year over year, and non-GAAP EPS of $3.16 beat the $2.40 consensus by 31.8%. Management raised the FY26 outlook to $13.7 billion to $14.3 billion in revenue with EPS of $10.75 to $11.35. CEO Jure Sola said , “ZT Systems revenue significantly exceeded our expectations, driven by strong execution and customer demand.”

Shares are up 32.1% year to date and 68.9% over the past year. Sanmina is the sort of picks-and-shovels supplier that keeps showing up when you follow the AI buildout past the chipmakers, the same thread we pulled on in a free report on seven of these infrastructure names, here . Still, if you want the mothership itself, it is the next name and it is still not priced for what its data center segment just did.

You cannot back the portfolio thesis without owning the entity writing the checks. AMD just posted the quarter that justifies every one of those 13F positions, and the stock has responded accordingly: shares are up 126.1% year to date and 175.05% over the past year, closing today at $484.

Q2 FY26 revenue surged 50.1% year over year to $11.536 billion, while non-GAAP EPS reached $1.66. Data Center did the heavy lifting, with revenue more than doubling to $6.718 billion and operating income swinging to $2.10 billion from a $155 million loss a year earlier. AMD expects the momentum to continue, guiding for roughly $13 billion in Q3 revenue, implying 41% year-over-year growth.

CEO Lisa Su framed the trajectory bluntly on the August call: “We are still in the early stages of a multi-year AI adoption cycle…driving demand for more compute and creating a clear path to significant revenue growth and earning power in the years ahead.” Su also flagged the target that matters for the portfolio thesis: the data center AI accelerator market growing more than 45% annually to approximately $1.4 trillion by 2030.

The next name is where AMD wrote a direct check to buy a seat at that table.

AMD’s 13F disclosed $210.8 million in Nutanix stock, 16.1% of the portfolio, but the equity is only part of it. AMD also committed a $150 million equity investment and up to $100 million in joint R&D funding to co-develop a full-stack agentic AI infrastructure platform with Nutanix ( NASDAQ:NTNX ). The first jointly-developed platform is expected late 2026. That timeline is the trade.

Fundamentals validate the check. Q3 FY26 revenue reached $703.07 million, up 10.0% year over year, with non-GAAP EPS of $0.47 beating the $0.35 consensus by 34.29%, marking the fourth consecutive EPS beat. Annual recurring revenue climbed to $2.43 billion, up 15% year over year, while management raised FY26 revenue guidance to $2.82 billion to $2.84 billion. CEO Rajiv Ramaswami cited “strong bookings, healthy new logo additions, and good free cash flow performance.”

Shares trade at a forward P/E of 31 with analysts targeting $60.71. But the trade AMD didn’t advertise sits at the bottom of the 13F, and it may be the loudest signal in the whole filing.

Deep in AMD’s 13F, tucked below the SpaceX and Cerebras headlines, sits a $2.37 million position in Xanadu Quantum, 0.18% of the portfolio. It is a rounding error. It is also the only place AMD is publicly telling you it wants exposure to photonic quantum. Because Xanadu is private, the publicly traded proxy for that thesis is Quantum Computing Inc. ( NASDAQ:QUBT ), whose room-temperature photonic architecture targets exactly the successor-to-GPU market AMD is quietly hedging.

The numbers are still small, but the direction is not. Q2 FY26 revenue came in at $5.55 million, up roughly 9,000% from $61,000 a year earlier, with the net loss narrowing to $11.75 million from $36.48 million. Cash and investments sit at $189.15 million, backlog reached $42.5 million as of June 30, 2026, and the $73.1 million NHanced Semiconductors acquisition launched Fab 2 ahead of schedule. Analysts see the option value: four Buy ratings against two Holds with an $18.33 target, roughly double the current $9.05 close.

CEO Yuping Huang’s pitch mirrors why AMD would want a toehold in this corner: “Room-temperature photonic architecture continues to differentiate QCi by providing a pathway to practical quantum systems with significantly lower complexity, cost and power requirements.” If AMD’s $2.37 million whisper becomes a shout, this ticker is the only public seat left.

AMD’s inaugural 13F reads like a treasure map. Sanmina builds the racks, Nutanix runs the stack, Absci consumes the compute, and QCi hedges the architecture. AMD itself sits in the middle, guiding to $13 billion in Q3 revenue while the data center AI TAM races toward $1.4 trillion by 2030. Wall Street coverage of this specific filing has not landed yet. That window closes fast.

Contact [email protected] for any questions or corrections.

Source: https://247wallst.com/investing/2026/08/18/amds-1-3-billion-portfolio-reveals-its-biggest-investments-and-where-ai-demand-lands-next/?.tsrc=rss