BofA Survey: China Internet Investor Sentiment Drops to Multi-Year Low as ByteDance Threat Escalates
0xBroomberg
A BofA Securities survey of Hong Kong and mainland institutional investors finds China internet sentiment has dropped below the April 2025 and September 2024 lows, with 'helplessness' surfacing in nearly every conversation — the malaise has shifted from a valuation problem to a wholesale absence of fundamental narratives.
How bearish are investors, exactly?
The KWEB index has fallen roughly 30% year-to-date, versus about 15% for the MSCI China index — a stark underperformance gap for the internet sector.
Despite historically low valuations, most investors surveyed said they refuse to pay above 15× P/E for any China internet stock.
This means → cheapness alone is no longer a buy signal. What investors want is a credible fundamental story, and right now there isn't one.
Missing the AI-hardware trade while stuck in internet names — how much does that cost?
The core tension: failing to size up AI chip and hardware positions has created steep opportunity costs, while internet stocks — though cheap — carry too much fundamental uncertainty to build a convincing buy case.
Macro headwinds compound the pain: April and May consumption data came in weak, and the 618 shopping festival underwhelmed. Consensus on the Chinese consumer has shifted from neutral to outright bearish.
Investors worry more about liquidity spillovers from tighter cross-border investment regulation than about direct regulatory risk on internet companies. In plain terms = they fear the funding channel narrowing, not a fine on any single company.
So many AI models — who actually wins?
Surveyed institutions broadly agree that "AI task routing" is taking shape — different tasks routed to the best-fit model rather than a single winner. This means → the trend broadly favors Chinese model development, but it also makes picking the eventual winner far harder.
Most investors view Zhipu as the "quality benchmark" in AI models; its GLM5.2 stands out for user appeal. Its lofty valuation, however, is a clear point of disagreement — even though the market broadly considers Zhipu's $1 billion year-end 2026 ARR guidance conservative.
Some investors see MiniMax as slightly behind the top tier in model capability. With limited free float and elevated valuations, institutional positioning in China AI-model stocks remains light overall.
How far has the ByteDance threat escalated?
After ByteDance's recent cloud-computing conference, investor anxiety over its AI and cloud competitiveness jumped sharply — making it the most intensely discussed topic in the survey.
Alibaba faces the most direct impact: year-to-date share-price weakness is attributed to a triple squeeze of macro pressure, intensifying cloud competition, and southbound-fund selling. Several investors expect Alibaba Cloud's June-quarter revenue growth to top 40% year-on-year, but ByteDance competition plus potential SOE entry into GPU leasing could pressure Alibaba Cloud's long-term profitability.
In plain terms = Alibaba Cloud's near-term growth rate looks fine, but investors worry that if ByteDance and state-owned players all crowd into the GPU-rental business, margin room keeps shrinking.
Has the negative AI narrative on Tencent bottomed out?
Multiple respondents agree Tencent is a core beneficiary of the AI narrative shifting from infrastructure to the application layer, and negative expectations around its AI progress may have troughed.
Views on WeChat Agent range from neutral to positive; yet some investors argue the EPS drag from WeChat Agent's launch is the main driver of current weakness — they plan to reassess after potential downgrades to 2027 EPS estimates.
This reflects a very specific point of disagreement on Tencent: not "can it do AI," but "how long does the WeChat Agent investment phase last, and how much earnings does it cost."
Among individual names, who is favored and who is left behind?
NetEase is the consensus defensive pick across both bulls and bears: upcoming Southbound Connect inclusion, limited AI and macro exposure, and a clear EPS upside path from new game launches.
Kuaishou drew extra attention after reports of a potential funding round for its AI-video unit Kling, though ByteDance's rival offering remains a key overhang. Meituan regained investor interest after a Q1 beat; most expect the stock to re-rate upward as unit economics improve.
PDD and Tencent Music saw attention drop sharply post-earnings, with investors broadly saying they lack a clear bottom-fishing thesis. This means → the market's stock-picking logic is deeply defensive — names with a story get touched; names without one are ignored no matter how cheap.
Content is for reference only, not financial advice.