AMD (AMD) on Solana

UBS raised its price target on AMD from $455 to $670, maintaining a Buy rating, citing improving server CPU demand trends and the chipmaker's positioning to benefit from the next phase of AI infrastructure build-out. The bank projects AMD's server CPU revenue will grow from roughly $16 billion in 2026 to $50 billion by 2030, underpinned by what it described as broader adoption of standalone CPU racks as enterprises deploy agentic AI workloads. UBS sees the overall server CPU total addressable market expanding from approximately $30 billion in 2025 to around $170 billion by 2030.

AMD's competitive moat in this space, according to UBS, rests on its higher core density, multithreading capabilities, and deep roots in the x86 software ecosystem — attributes that make its EPYC processors well-suited for traditional workloads increasingly woven into agentic AI applications. The bullish call comes as AMD shares have more than doubled in 2026, with the stock climbing from around $200 in early 2026 to above $580, fueled by a data center segment that grew 57% year-over-year to $5.8 billion in Q1. UBS sees 23% additional upside from current levels at its $670 target.

AMD shares climbed roughly 1% to $544.98 in overnight trading after the company's Advancing AI 2026 conference delivered a combination of hardware milestones and high-profile customer confirmations that analysts said reinforced its position in the server market. The Helios rack-scale AI system — built around the new Instinct MI455X GPU and sixth-generation EPYC "Venice" CPUs — entered full production and is targeted to ship by the end of Q3. OpenAI, Anthropic, Microsoft, Meta, and Oracle were all named as partners, with Anthropic and OpenAI specifically discussing rapid Helios deployment. CEO Lisa Su framed the opportunity broadly, projecting the total computing market will reach $2 trillion by 2030, with AI accelerators alone accounting for $1.4 trillion.

Jefferies responded by raising its AMD price target from $515 to $640 while maintaining a Buy rating, citing the chipmaker's "clear leadership in the server market." Analyst Patrick Moorhead of Moor Insights & Strategy added that "this is a very different AMD and shouldn't be underestimated." The stock had pulled back 2.3% during Thursday's regular session ahead of the event and is still trading below its all-time high of $580.91 reached June 30. AMD is set to report Q2 results on August 4.

Nvidia has released full technical specifications for its Vera CPU, a custom-designed server processor built around 88 proprietary Olympus cores, and has begun delivering chips to customers including OpenAI and Anthropic. In benchmark testing, Vera scored 925 on the SPECrate integer suite versus 898 for a comparable dual-socket AMD EPYC 9755 system — a modest 3% lead achieved with fewer threads — and Nvidia claims up to 50% better performance on AI agent workloads versus x86 processors broadly. The announcement signals that Nvidia is no longer content to cede the server CPU market to AMD and Intel while selling GPUs alongside their processors.

For AMD, the threat is twofold. First, Vera directly targets the high-margin server CPU segment where AMD's EPYC lineup has been gaining share against Intel; AMD has publicly raised its server CPU total addressable market outlook to over $120 billion by 2030 and set a target of more than 50% server CPU market share. Second, Nvidia is positioning Vera as part of an integrated AI computing stack — GPUs, networking, and now CPUs from a single vendor — which could reduce the appeal of pairing AMD Instinct MI-series GPUs or its Helios rack systems with third-party CPUs. AMD is holding its "Advancing AI" event on July 22, where new customer wins for its Instinct and Helios platforms are expected to be a key measure of how well it is countering Nvidia's expanding ecosystem.

AMD announced Helios, its first rack-scale AI platform and its most direct challenge yet to Nvidia's dominance in data center AI hardware. Helios integrates AMD Instinct GPUs, EPYC processors, networking technology, and ROCm software into a unified system designed as an alternative to Nvidia's Grace Blackwell lineup. Microsoft confirmed it will deploy Helios in Azure data centers for frontier AI inference, joining Meta, OpenAI, Oracle, and Tata Consultancy Services as early customers.

AMD's data center segment grew 57% year-over-year in Q1 2026, and the company projects tens of billions in AI revenue starting in 2027, with Helios expected to contribute the majority. Nvidia still commands over 95% of the data center GPU market, but the breadth of Helios customer commitments marks AMD's strongest competitive position in the AI hardware race to date.

A YAML configuration file attributed to Anush Elangovan, AMD's VP of AI software, listed Anthropic as a customer with maximum priority — assigned 30 "priority boost points" — placing it alongside existing hyperscale AMD customers such as Meta, according to analysis by SemiAnalysis. The file, surfaced on GitHub, suggests Anthropic is actively in AMD's customer pipeline, though SemiAnalysis noted the AI lab remains in an "evaluation phase," meaning the relationship is not yet finalized. Anthropic has not publicly confirmed AMD hardware use, with its acknowledged compute infrastructure currently limited to Nvidia GPUs, Amazon Trainium, and Google TPUs.

Supporting the speculation, Jefferies analyst Blayne Curtis noted that Anthropic has been hiring engineers with ROCm experience — AMD's open-source AI software stack — signaling deliberate infrastructure diversification efforts. AMD shares rose 1.3% overnight on the GitHub discovery, recovering a portion of a 16% decline over the prior three trading sessions. The timing is notable ahead of AMD's "Advancing AI" conference on July 22–23, where the company is expected to showcase its 2nm Zen 6 processors and Helios platform, events at which new customer announcements could add weight to the speculation.

China's release of Kimi K3 — described as the largest open AI model to date — is adding a new layer of competitive pressure on AMD heading into its July 22 "Advancing AI" event. The model, developed by Moonshot AI, targets coding-heavy workloads at lower cost than comparable Western alternatives, signaling that capable Chinese AI is advancing faster than many had anticipated. Because AMD sells CPUs and GPUs into the data center and AI accelerator market, the emergence of competitive open-source models built inside China's hardware ecosystem risks narrowing the addressable market for AMD's high-end chips in that region.

Two structural concerns follow from the launch. First, as Chinese AI labs increasingly build around domestic hardware stacks, export controls may preclude AMD from capturing incremental demand even if it wanted to compete. Second, the broader narrative of efficient, low-cost models — whether from China or elsewhere — continues to challenge the assumption that frontier AI workloads will require ever-more expensive silicon. AMD shares were trading around $495.76 at the time of reporting, down roughly 3% over the prior 30 days and sitting above most analyst fair-value estimates, leaving limited margin for error if AI hardware spending expectations are revised downward.

AMD is holding its "Advancing AI" event on July 22, 2026, where the company is expected to reveal updates on its MI500 GPU lineup, a scale-up networking roadmap built around optical interconnects, and details on a CPU market opportunity that Jefferies analyst Blayne Curtis estimates could exceed $200 billion. While the product disclosures will draw attention, Curtis argues the bigger swing factor for investors will be new customer announcements — pointing to growing interest in whether AMD has secured a deal with Anthropic to complement its existing Microsoft relationship, where the MI450 chip is reportedly already in use.

Pricing terms are likely to matter as much as the customer list itself. AMD has historically offered substantial incentives to land early AI workload wins with hyperscalers like OpenAI and Meta, and Curtis cautioned that economics would matter more than headlines. AMD's Instinct GPU line is its primary challenge to Nvidia's dominance in AI compute, and the event represents an opportunity for the company to demonstrate that its strategy is translating into durable, revenue-generating partnerships rather than discounted pilots.

William Blair analyst Sebastien Naji, rated in the top 2% of Wall Street analysts with an 81% success rate, issued a Market Perform rating on AMD with a fair-value estimate of $565, cautioning that the stock's AI-driven rally may have already priced in most of the upside. AMD currently trades at 73x forward P/E — 195% above the sector median — and 18x forward sales, 426% above the sector median, raising the question of how much of AMD's AI opportunity investors have already paid for despite the company's strong fundamentals.

Naji acknowledged AMD's exceptional Q1 2026 results (revenue up 38% year-over-year to $10.3B, data center revenue up 57% to $5.8B) and robust Q2 guidance of approximately $11.2B, but flagged mounting competitive pressure from Arm, Qualcomm, Nvidia, and Intel — which could regain server CPU competitiveness by 2028 — as well as the risk that hyperscaler custom AI chips could limit AMD's addressable market. The stock has surged roughly 169% over the past six months.

Samsung Electronics reported a record Q2 2026 operating profit of approximately $58 billion — a 19-fold year-over-year increase that beat analyst estimates by 6% — yet its shares still fell as much as 10% intraday in Seoul. Investors had already priced in the blowout after Samsung's shares rallied 150% over the prior year, and the results did little to calm concerns about whether AI infrastructure spending by major U.S. technology firms can sustain the margins the market has come to expect.

AMD bore the brunt of the selloff in the U.S., dropping 8% to $508 from Monday's close of $552.05. Trading at a P/E ratio of 208x, AMD carries a valuation that leaves almost no margin for doubt about the AI demand trajectory. Intel and Applied Materials each shed 10%, Micron fell roughly 5%, and the iShares Semiconductor ETF slid 6% — while the Dow Jones held near record highs, marking Tuesday's rout as a chip-sector reset rather than a broad market risk-off.