MediaTek reports first quarter results, raises AI chip guidance By Investing.com

MediaTek reported Q1 revenue of NT$149.2 billion, at the high end of guidance, with gross margin of 46.3% in line with estimates. Management guided Q2 revenue to NT$144.7 billion, implying a 3% sequential decline, but raised its 2027 ASIC market forecast to $70 billion-$80 billion and lifted the expected revenue from its first AI ASIC project to about $2 billion by Q4 2026. The company also said smartphone demand should recover after a 2nm flagship chip launch in late Q3, while ASIC revenue is expected to outgrow mobile and surpass smartphone revenue by 2027.

Analysis

The real signal is not the headline mobile cadence, but the acceleration of ASIC as a strategic wedge into a much larger, higher-margin TAM. If management is credible on a $2B first project by late 2026 and multi-billion revenue in 2027, this reframes the company from a handset-cycle proxy into a semi-custom AI infrastructure contender, with a materially better mix than the market likely prices in. The second-order effect is that design wins in ASIC can pull through advanced packaging, test, and foundry capacity earlier than expected, which tends to create a flywheel of ecosystem lock-in and raises switching costs for hyperscaler customers.

The setup is more interesting in the context of the broader AI spend chain: this is incremental demand, but it also competes for wallet share with merchant GPUs and other accelerator programs. If ASIC share inflects toward the stated 10%-15% range, the market may need to lower long-term assumptions for some incumbent AI hardware beneficiaries, especially where customers are prioritizing power efficiency and TCO over absolute flexibility. The upside case is strongest if the company proves it can convert early wins into repeatable platform revenue rather than one-off tapeouts.

The main risk is timing slippage, not demand destruction. ASIC ramps are notoriously back-end loaded and vulnerable to customer qualification delays, yield issues, and a mismatch between booked revenue and actual gross profit conversion; that matters most over the next 6-12 months. Near term, the smartphone recovery narrative is likely a lower-quality support layer, while the real value creation depends on whether AI ASIC ramps are visible in backlog by the second half of 2026.

Contrarian view: the market may be too focused on the AI optionality and underappreciating how much of the near-term margin stability depends on pricing discipline in a weak handset market. If the 2nm flagship launch underwhelms or the AI ASIC project slips by even two quarters, the stock can de-rate quickly because the equity story is leaning forward to 2027 revenue that is still largely non-recurring in nature today.

Key Decisions for Investors

• Long 2454.TW on weakness, 6-12 month horizon: accumulate only on pullbacks given the asymmetric upside if ASIC revenue visibility improves; target a re-rating on mix shift, with downside capped if handset recovery remains merely stable.
• Buy 2454.TW 2026 call spreads instead of outright equity if available: structure for the 2H26 catalyst window, using defined risk because tape-out/yield slippage can delay the earnings inflection by quarters.
• Pair trade: long 2454.TW / short a basket of weaker handset-dependent semiconductor names over the next 6 months, expressing the view that semi-custom AI share gains will outperform cyclical mobile exposure.
• Reduce relative exposure to merchant AI accelerator names on any strength if management commentary on ASIC wins continues to improve; the risk/reward shifts if customers increasingly prefer custom silicon over generalized compute.
• If 2Q/3Q updates fail to show backlog conversion or design-win expansion, trim by 25%-30% and wait for confirmed revenue traction rather than paying for 2027 optionality too early.