中国互联网：2022-2023策略——将BOSS直聘评级上调至跑赢大盘

China Internet: The 2022-2023 playbook - upgrading Boss Zhipinto Outperform There’s a price for everything.We started 2026 with a cautious view on the ChinaInternet sector, reflecting concerns about anemic macro and regulatory posture. Thatsaid, with valuations in a variety of cases revisiting 2022-2023 lows, we think valueopportunities may be starting to re-emerge where our coverage companies are reverting tothe self-help and capital returns playbook that helped to support a bottom back then. Robin Zhu+852 2123 2659robin.zhu@bernsteinsg.com Charles Gou+852 2123 2618charles.gou@bernsteinsg.com Upgrading Boss Zhipin to Outperform (PT $18).We’ve long considered the companygood executors in a sheltered vertical market, where the main headwinds have beenmacro in nature. Concerns linger on AI displacement of white collar jobs, but smallerpeers are gradually exiting, and 50% of its revenue now comes from blue collar (e.g.logistics, manufacturing) or semis and AI jobs. The company has started to leverage AI tosell candidate matches to employers - at higher ARPJ than job postings. High-frequencydata has continued to look supportive of KA billings growth, which should help to supportenterprise ARPU. In contrast, sub-4x 2027E PE ex. cash and Z.ai (49% of market cap) and20% FCF/EV yield price in going concern risk, which we find implausible. The companybought back 3% of its share count through mid-May, and management’s pledge to return50% of prior year net income support c. 5% annualized capital return yield from here. Min-Joo Kang+852 2123 2644minjoo.kang@bernsteinsg.com Hyrum Caesar+81 3 6777 6979hyrum.caesar@bernsteinsg.com NetEase’s self-help arc.NetEase trades on c. 12x forward PE, near 2022 lows. Q1pointed to expense savings that should continue in future quarters, Sea of Remnantsexpectations have fallen to low levels (in contrast with the company pointing to solid betafeedback), and revenue growth comps get easier in H2. Thereafter, we expect greaterAnanta clarity to help anchor 2027-2028 earnings expectations. NetEase’s resumption ofbuybacks in Q1 felt notable, as a “body language” signal. By Q1 2027 NetEase has to finishconverting its HK shares to a primary listing. Tencent, Alibaba: the cost of AI-maxxing, and the way out.Both Tencent and Alibaba’sshares have performed poorly in recent months, as rising AI investments have weighedon earnings revisions. For Tencent, the level of agentic AI required to support a WeChatcompanion or Mini Shops monetization strikes us as achievable relatively soon, whichshould then helps the company show renewed operating leverage and earnings growth. ForAlibaba, we expect price increases and mix improvement to help boost Alicloud EBITA, butgreater clarity around the company’s All Other losses remains key to unlocking the stock. AI top-funnel update - still loading.While Doubao DAUs have continued to grow, theintroduction of paid subscriptions argues against the idea that Bytedance might wantto subsidise usage ad infinitum. Doubao (and Yuanbao) continue to see 5-6 activationsper day per DAU, while Qwen usage has remained lower. “Not yet” remains the obviousreason in favour of disruption risk. On the other hand, both QuestMobile data and ourdiscussions with industry contacts have reflected the difficulty associated with retraininglong entrenched consumer usage habits. BERNSTEIN TICKER TABLE O - Outperform, M - Market-Perform, U - Underperform, NR - Not Rated, CS - Coverage SuspendedSource: Bloomberg, Bernstein estimates and analysis. INVESTMENT IMPLICATIONS That the China Internet sector has had a weak first five months of 2026 was probably unsurprising, amid retail consumptiongrowth remaining stagnant, regulatory actions against Trip.com and PDD still casting a shadow over investor sentiment, and agenerational AI hardware build-out driving flow headwinds and valuation de-rating. At this point, with valuation multiples in thesector back to 2022-2023 lows, it feels notable that several companies, like NetEase, JD, and Boss Zhipin have reverted to the2022-2023 playbook of self-help and capital returns that helped to carve out a bottom. JD has outperformed the sector year todate, and NetEase has seen a bounce in sentiment on the back of its Q1 beat. Tencent feels heavily oversold to us. The company is too upstanding to report an “Others” segment to house the company’speripheral businesses and investment spend, but the core operating business (excluding listed investments) is probably tradingon a 10x multiple of 2027E profits at this point. That said, compared with ongoing questions on AI capex intensity and the ROIthereon (applicable to Alibaba too), the path of least resistance in the sector likely remains situations more likely to featureearnings revisions and higher capital return yield. With this note, we’ve upgraded Boss Zhipin from Market-Perform to Outperform. In contrast with most of our coverage stuckin a 2-3% retail consumption growth environment, Boss Zhipin