BofA Maintains Buy Rating on Tencent, Citing Turning Point in AI Competitive Landscape
Nashnova编辑部
BofA maintains Tencent's Buy rating and HK$780 target price, framing the company's AI positioning as a 'Midway moment' — the balance of power is shifting, yet a sub-13x 2026/2027 P/E still fails to price in Tencent's AI optionality.
Why does BofA call this a "Midway moment"?
BofA analysts argue Tencent's rapid catch-up in model capability and compute accumulation this year marks a potential inflection in the long-term competitive landscape.
This means → the market still focuses on cloud/chip front-runners, but Tencent's AI optionality — token and cloud revenue upside — remains unpriced.
In plain terms = the war isn't over, but momentum is shifting — BofA believes the market hasn't noticed yet.
Tencent is spending big on compute — where is the money going?
Management sharply raised AI infrastructure capex in Q2 and made large advance payments to lock in memory and compute resources ahead of industry price hikes.
BofA lifted its 2026/2027 capex forecasts to RMB 210 billion and RMB 260 billion respectively, up from RMB 185 billion and RMB 225 billion.
This means → cash pressure peaks in Q2–Q3 2026; as one-off prepayments normalize, 2027 operating cash flow should stabilize.
What path is the Hunyuan model taking?
Tencent is not chasing benchmark rankings outright — it is balancing model capability gains with inference cost efficiency.
The next version, HY4, is expected within three months, further narrowing the gap with China's largest-parameter frontier models.
In plain terms = the race isn't about who builds the biggest model — it's about who runs cheapest at comparable capability, because that determines who can truly scale.
Which two AI application lines deserve the closest watch?
WorkBuddy is positioned by management as Tencent's most important AI application opportunity: user interaction data feeds back into model intelligence, creating a positive flywheel. Monetization is still early-stage.
WeChat AI assistant Xiaowei is following a cautious, phased rollout — full-scale launch could come as early as Q4 2026, exploring use cases such as travel planning and price-comparison shopping.
This reflects Tencent's AI go-to-market logic: prove the enterprise flywheel with WorkBuddy first, then use Xiaowei to tap the billion-user WeChat base for everyday consumer scenarios.
Why were earnings forecasts actually cut?
BofA trimmed Tencent's 2026–2028 non-IFRS net profit forecasts by 2% to 6%, driven by depreciation charges set to accelerate markedly over coming quarters.
Specifically, 2026/2027 non-IFRS diluted EPS estimates now stand at RMB 29.30 and RMB 29.81 (from RMB 29.80 and RMB 31.20); 2027 net profit is projected flat year-on-year.
This means → aggressive investment buys long-term AI optionality, but the near-term price is profit growth consumed by depreciation.
What are the key milestones to watch next?
BofA flags three catalysts: Hunyuan model upgrade within three months, Xiaowei mass rollout within six months, and WorkBuddy plus cloud revenue acceleration within one year.
The critical test: whether the depreciation peak arrives on schedule in Q3 2026, and whether cash flow recovers as expected in 2027.
In plain terms = the "Midway" narrative either proves out or breaks down at these time windows — watch the numbers.
Content is for reference only, not financial advice.