TSMC (TSM) on Solana

TSMC reported second-quarter net income of $22 billion, up 77% year-over-year, with revenue rising 34% to $40.2 billion and gross margin expanding to 67.7%. The results mark the company's fifth consecutive record quarterly profit, driven by surging demand for advanced AI chips — advanced nodes of 7nm and below accounted for 77% of wafer revenue in the quarter, with 3nm alone contributing 30%.

Alongside the earnings release, TSMC announced an additional $100 billion investment in its Arizona operations, bringing its total committed US manufacturing spend to $265 billion. The expansion will add four new fabs and a packaging plant to the Phoenix-area site, and follows a broad trade framework between Taiwan and the Trump administration that cleared the path for the accelerated buildout. TSMC also raised its full-year capital expenditure guidance by up to 14% and guided Q3 2026 revenue to $44.6–$45.8 billion, reflecting continued confidence in AI-driven chip demand.

TSMC reported June 2026 revenue of NT$442.68 billion (approximately $14.6 billion USD), a 67.9% jump year-over-year and a 6.2% gain from May, cementing the company's strongest first-half performance on record. For the full second quarter, consolidated revenue reached NT$1.270 trillion ($39.63 billion), up 36% from the same period in 2025 — in line with Wall Street's consensus estimate of $39.76 billion ahead of the company's July 16 earnings call.

The surge is being driven almost entirely by AI chip demand. Analysts note TSMC is "sold out on N3," its leading 3-nanometer process node, and project the company will generate over $40 billion in AI chip revenue in 2026, representing nearly a quarter of total output. First-half 2026 revenue came in at NT$2.40 trillion, up 35.6% versus the same period last year. The June figures were released Monday rather than the usual Friday schedule due to disruptions from Typhoon Bavi, but the strong numbers prompted fresh price-target increases from Bank of America ($590), Susquehanna ($575), and Barclays ($470).

BlackRock ranks TSMC among its 30 most important holdings, with the fund holding approximately 18.22 million shares — about 0.35% of outstanding stock — placing it inside the top 10 institutional shareholders. The conviction rests on TSMC's irreplaceable role in advanced semiconductor manufacturing at a moment when AI infrastructure spending is accelerating sharply. Bank of America Securities, citing surging "agentic AI demand," raised its topline growth forecasts to 40% for 2026 and 39% for 2027, driven by elevated demand for AI GPUs, ASICs, and CPUs that only TSMC's leading-edge fabs can produce at scale.

Analyst price targets have followed the demand thesis upward. Bank of America lifted its TSMC target from $490 to $590 on June 24, while analysts broadly note that current AI chip demand is outpacing available advanced-node capacity. TSMC's expertise in mixed-signal, embedded memory, and radio-frequency process technologies positions it as the critical bottleneck and primary beneficiary as hyperscalers and chipmakers compete for wafer allocation.

S&P Global revised its outlook on TSMC's 'AA-' long-term issuer credit rating on June 23, 2026, citing the company's "healthy EBITDA and free operating cash flow growth" driven by technological and scale advantages over competitors. The agency pointed to TSMC's entrenched leadership in advanced high-performance computing chip fabrication — particularly for AI workloads — as the key factor underpinning the improved credit assessment.

S&P noted TSMC's ability to sustain a net cash position even while running elevated capital expenditures and paying dividends, a combination that signals strong underlying cash generation. The upgrade in outlook reflects growing confidence that TSMC's dominance in AI chip manufacturing is durable enough to translate into a structurally stronger balance sheet over the next one to two years.

Despite a 378% three-year return and 114% gain over the past year, TSMC (TSM) may still trade at a meaningful discount to fair value. The stock's current P/E of 32.7x sits well below a model-derived fair P/E of 61.8x and lags both the semiconductor industry average of 75.5x and its peer average of 77.4x — a gap analysts attribute partly to geopolitical risk rather than deteriorating fundamentals.

The core tension for investors is that TSMC's position at the center of AI chip demand creates exceptional business economics, while Taiwan's geopolitical exposure and U.S. export controls on advanced AI chips to China introduce a structural overhang. Heavy capital expenditure requirements for next-generation capacity expansion add further near-term uncertainty, even as AI-driven orders continue to set revenue records.

TSMC reported May 2026 revenue of NT$416.98 billion ($13.2 billion), a 30% increase from the same month a year earlier, with combined April and May sales up roughly 24% year-over-year. Analysts polled by Bloomberg project second-quarter sales will rise approximately 35%, reflecting sustained demand tied to global AI infrastructure buildouts. The company is a critical supplier to Nvidia and AMD for high-end AI accelerator chips.

TSMC raised its full-year capital spending outlook to the upper end of its $56 billion forecast in April, and CEO C.C. Wei has signaled that global chip supply will remain constrained for years. While AI-driven orders are driving the headline growth, the company has flagged some softness in smartphone and consumer segments, meaning the revenue gains are concentrated in advanced process nodes serving data center and AI customers.

Susquehanna raised its price target on Taiwan Semiconductor Manufacturing (TSM) from $500 to $575 on June 22, 2026, while maintaining a Positive rating. The firm updated its model to reflect TSMC's capital expenditure and capacity strategy, which it believes will exceed both consensus and buy-side expectations — a bullish signal pointing to greater wafer output than the market has priced in.

The key uncertainty Susquehanna flagged is how AI token-generation growth and the associated silicon requirements could drive supply-demand imbalances at TSMC's fabs. That dynamic cuts both ways: sustained AI inference and training demand could push utilization higher, but rapid shifts in workload patterns add forecasting risk. The call comes as TSMC reported May 2026 revenue up 30.1% year-over-year, underscoring the demand environment the firm's revised model is tracking.