AMD (AMD) on Solana

AMD is committing more than $10 billion to Taiwan through 2029, targeting advanced chip packaging and substrate manufacturing capacity needed to scale its AI hardware business. The investment spans TSMC's CoWoS-L packaging as well as next-generation Elevated Fanout Bridge packaging developed with ASE Technology and Siliconware Precision Industries, giving AMD alternative suppliers to reduce bottlenecks. CEO Lisa Su has pointed to constrained advanced-packaging capacity as a direct limiter on customer growth, and the spend is aimed squarely at removing that constraint as AMD's data center AI revenue is projected to grow at over 80% compound annual growth rate.

The buildout is closely tied to AMD's Helios AI rack systems, each integrating 72 GPUs and 18 CPUs, which require substantial packaging and assembly throughput. Despite AMD's strong AI revenue trajectory, its non-GAAP operating margin of 27% in Q2 2026 remains below management's 35% target, making efficient capacity scaling a priority. The Taiwan commitment signals AMD is moving from opportunistic supply agreements to long-term infrastructure ownership alongside the broader semiconductor ecosystem.

AMD captured 30.7% of the overall x86 processor market in Q2 2026, its highest share ever, rising to 34.1% when IoT and embedded products are included, according to Mercury Research analyst Dean McCarron. Total AMD unit shipments and total market share both reached new record highs, with year-over-year share growth of 6.5% — gains that came even as the desktop CPU segment suffered its worst quarter in recent memory, declining more than 20% sequentially and year-over-year due to elevated component costs across graphics cards, motherboards, memory, and storage.

Within the battered desktop segment, AMD actually widened its lead, adding 1.8 percentage points sequentially and 2.7 points year-over-year to reach 34.9% desktop share, while Intel's desktop share fell to 65.1%. AMD also hit record client CPU share of 30.3%, gained 8.4 percentage points year-over-year in mobile to reach 28.9%, and took 34.5% of the server CPU market — 46.4% on a core-count basis comparing EPYC directly against Xeon. Intel retained overall x86 unit leadership at 69.3% and ramped mobile CPU production substantially during the quarter, but the volume push was insufficient to halt AMD's multi-front share advance.

AMD has appointed Tim Ryan to its Board of Directors following the retirement of a long-serving director. Ryan brings decades of experience from senior roles at Citigroup and PwC, with a background spanning technology, corporate strategy, risk management, and accounting.

The appointment coincides with changes to board committee composition, including the establishment of a new Audit and Finance chair position. The addition strengthens board oversight at a time when AMD is scaling its AI and data center operations, with Ryan's financial and risk management credentials expected to bring greater scrutiny to capital allocation decisions in those rapidly growing segments.

Wall Street's consensus on AMD is broadly bullish: with shares trading near $484, the average analyst price target sits at $613, implying roughly 27% upside over the next year — a return that outpaces typical market expectations, though it still trails Nvidia's projected 38% gain. That consensus places AMD firmly in the "buy" camp alongside Nvidia, while Intel draws the sector's cautious rating.

Motley Fool contributor Keithen Drury pushes back on that view. Writing for Yahoo Finance, Drury argues that AMD and Intel are "both very expensive and trading on the expectations that the next few years will provide a serious turnaround," suggesting the recovery investors anticipate is already reflected in AMD's valuation. By contrast, he frames Nvidia as the stronger pick — a company that needs no turnaround yet trades at what he considers a more reasonable multiple. The piece highlights the tension between consensus analyst optimism on AMD and a more skeptical read that the stock's risk/reward may be less compelling than the headline target implies.

AMD priced $4.75 billion in senior notes on August 15, 2026, across four tranches: $1.25 billion due 2029 at 4.6%, $1.5 billion due 2031 at 5.0%, $1 billion due 2033 at 5.25%, and $1 billion due 2036 at 5.5%. The deal is more than triple AMD's prior bond sale of $1.5 billion in March 2025 and carries an annual interest cost of roughly $240 million, with spreads ranging from 0.43 to 0.9 percentage points above comparable U.S. Treasuries — terms that reflect strong investor confidence in the company's credit quality.

AMD said proceeds are for general corporate purposes, which may include debt repayment, without designating a specific use. The raise comes as the company carries $13.1 billion in cash and short-term investments following a Q2 2026 in which it posted $11.5 billion in revenue — 50% year-over-year growth — driven by a $6.7 billion data center segment that accounted for 58% of total sales. Post-issuance total debt stands at approximately $8 billion.

Stanley Druckenmiller initiated a new 72,900-share position in AMD during the latest quarter, according to a 13F filing disclosed on August 14, 2026. The purchase came alongside a wholesale exit from Broadcom, Intel, and Micron Technology, making AMD the sole surviving semiconductor name added to his portfolio in the reshuffle — a selective bet rather than a broad sector rotation.

The move signals differentiated institutional conviction in AMD's AI infrastructure positioning at a time when Druckenmiller was shedding diversified chip exposure. He simultaneously increased his Taiwan Semiconductor Manufacturing stake slightly, suggesting the repositioning favors companies with clear roles in AI compute supply chains over broader-market chipmakers. AMD's growing data center GPU revenue — which more than doubled in Q2 2026 — provides the fundamental backdrop for that thesis.

Patrick Moorhead, who spent more than 11 years as AMD's corporate vice president of strategy before founding analyst firm Moor Insights & Strategy, says Nvidia is using its larger balance sheet to blunt AMD's momentum in AI accelerators — and he recognizes the tactic from personal experience. "I experienced this same play when I was at AMD and our biggest competitor was Intel," Moorhead said, noting that Intel would expand its product offerings whenever AMD gained technological ground. He described his AMD tenure as winning four out of eleven competitive battles against Intel, and attributed Intel's eventual decline not to any single loss but to simultaneous erosion of both its foundry technology and product execution.

The parallel puts AMD's current AI push in a historically familiar frame. AMD is positioning its Helios rack-scale platform — combining 72 Instinct MI455X accelerators with ROCm software — as a credible alternative to Nvidia's systems, and has secured commitments from OpenAI, Meta, and Anthropic, with Anthropic planning up to two gigawatts of AMD-based capacity by the first half of 2027. Moorhead's read suggests the structural risk for AMD is not near-term displacement but the longer grind of competing against a rival with deeper resources to expand, subsidize, and match on every product dimension — exactly the dynamic that eventually wore on Intel's dominance and, by his own count, produced more losses than wins for AMD during that era.

AMD announced a major data center infrastructure partnership with Core Scientific on July 28, securing 15-year leases covering approximately 530 megawatts across seven sites, with reservation rights for an additional 1,925 megawatts through December 2028. The agreement carries more than $14 billion in base contracted revenue potential and gives AMD warrants to purchase up to 30 million Core Scientific shares at $23.47 per share, with roughly 6.5 million shares' worth vesting at signing.

The deal arrives after Core Scientific shareholders in July 2025 rejected an all-stock acquisition by CoreWeave valued at approximately $9 billion — a vote influenced in part by investors who argued the company could attract higher-value, diversified AI customers on its own. AMD's willingness to commit to a multi-gigawatt, multi-decade arrangement provides a concrete data point supporting that thesis, positioning AMD as a large-scale anchor tenant in the buildout of dedicated AI infrastructure capacity.

AMD's stock retreated after its Q2 2026 earnings despite data center revenue surging 107% year-over-year to $6.7 billion and overall revenue rising 50% to $11.54 billion. Motley Fool analyst Geoffrey Seiler argues the sell-off is a buying opportunity, pointing to AMD trading at 35x forward 2027 earnings with a PEG ratio of 0.3 — well below levels typically associated with fair value. Adjusted EPS of $1.66 beat consensus by $0.04 and rose 246% year-over-year, while adjusted gross margins expanded 1,300 basis points to 56%.

The bull case centers on AMD's position in what Seiler describes as the "early innings" of an AI infrastructure buildout. The company's Helios full-rack AI system is launching this quarter with confirmed hyperscaler customers spanning OpenAI, Meta, Anthropic, and Microsoft, targeting what the company projects as a $220 billion inference and agentic AI market. AI segment revenue is expected to more than double in 2027, and AMD's Q3 guidance of $13 billion — representing 41% year-over-year growth — suggests the post-earnings pullback reflects short-term capex concerns rather than a change in competitive position.

AMD is acquiring Taalas, a Toronto startup founded in 2023 that etches AI model weights directly into custom silicon, converting any AI model into dedicated hardware in roughly two months. Taalas has raised $219 million in venture funding and currently ships chips running a small version of Meta's Llama 3.1. AMD plans to integrate the technology alongside its Instinct GPU lineup; deal price was not disclosed and closing is expected in Q4 2026.

The acquisition follows Nvidia's $20 billion purchase of Groq assets earlier in 2026, signaling that major chipmakers are moving to lock up hard-wired inference technology as AI deployment scales.