MiniMax (0100.HK) vs Zhipu AI (2513.HK) — Peer Valuation Comparison (Sep 2026)
Prices as of Sep 7-8, 2026 | Incorporates H1 2026 interim results (Aug 26) | All figures USD unless stated | Not investment advice
1. H1 2026 Interim Results — The Revenue Mix Flip (Aug 26, 2026)
| Metric | H1 2025 | H1 2026 | YoY Change |
|---|---|---|---|
| Total Revenue | $30.4M | $116.6M | +283.1% |
| — Open Platform / Enterprise | $9.2M (30.3%) | $73.9M (63.4%) | +703.1% |
| — AI-native Products | $21.2M (69.7%) | $42.6M (36.6%) | +100.9% |
| Gross Profit | $3.7M | $20.8M | +464.8% |
| Gross Margin | 12.1% | 17.9% | +5.8pp |
| R&D Expense | $124.4M | $297.0M | +138.8% |
| Net Loss (IFRS) | $402.0M | $358.0M | -11.0% |
| Adjusted Net Loss | $138.7M | $293.0M | +111.2% |
| ARR (Aug 2026 disclosure) | — | $800M+ | B2B over 80% |
| Token consumption (Jul vs Jan) | — | — | 20x growth |
| Enterprise customers/developers | — | 2M+ | — |
| Cash balance | — | $1.323B | — |
Critical: IFRS net loss narrowed because H1 2025 included $253.9M in fair-value charges on pre-IPO convertible instruments. Stripping those out, the underlying adjusted loss doubled — R&D spend ($297M) alone was 2.5x revenue. However, R&D growth (139%) significantly lagged revenue growth (283%), showing improving operating leverage. ARR of $800M in August implies September run-rate potentially approaching $1B annualized.
2. Peer Comparison — AI Foundation Model Companies (Updated)
| Company | Listing | Mkt Cap | H1 2026 Rev | H1 Rev Growth | Gross Margin | EV/ARR | EV/FY26E Rev | Cash |
|---|---|---|---|---|---|---|---|---|
| MiniMax (0100.HK) | HKEX | $15.7B | $116.6M | +283% | 17.9% | 18x | approx 32x | $1.32B |
| Zhipu AI / Z.AI (2513.HK) | HKEX | approx $60-64B | RMB 954M (approx $131M) | +399% | 26.0% | N/A | approx 200x+ | approx $5-6B |
| OpenAI | Private ($852B) | $852B | approx $12.5B (est.) | approx 100%+ | approx 33% | approx 34x | approx 34x | $138B raised |
| Anthropic | Private ($965B) | $965B | approx $23B+ (est.) | approx 400%+ | N/A | approx 21x | N/A | $87B raised |
MiniMax ARR of $800M+ disclosed on earnings call (Aug 26). EV/ARR of 18x is now comparable to Anthropic (21x) and cheaper than OpenAI (34x). Zhipu's H1 2026 revenue of RMB 954M (+399%) is impressive but market cap remains approximately 4x MiniMax's for only modestly higher revenue. Zhipu dropped 38% from its July peak of HK$1,655 to HK$1,018 as lock-up shares entered the float.
3. MiniMax vs Zhipu AI — Head-to-Head (Updated)
| Metric | MiniMax (0100) Sep 7 | Zhipu / Z.AI (2513) Sep 7 | Edge |
|---|---|---|---|
| Stock Price | HK$350.80 | HK$1,018 | — |
| vs Jul 12-13 Price | +30.6% (from HK$268.60) | -38.5% (from HK$1,655) | Gap narrowing |
| Market Cap (USD) | approx $15.7B | approx $60-64B | Zhipu approx 4x larger |
| H1 2026 Revenue (USD) | $116.6M | approx $131M | Zhipu 1.1x larger |
| H1 2026 Revenue Growth | +283% | +399% | Zhipu faster H1 |
| ARR (Aug 2026) | $800M+ | Not disclosed | MiniMax transparent |
| EV / H1 Rev (ann.) | approx 62x | approx 230x | MiniMax 3.7x cheaper |
| Gross Margin (H1 2026) | 17.9% | 26.0% | Zhipu higher |
| Revenue Mix | 63.4% B2B / 36.6% consumer | B2B dominant (on-prem + cloud) | MiniMax flipped to B2B |
| Intl Revenue Mix | 70%+ international | Mostly domestic China | MiniMax diversified |
| Token Consumption Trend | 20x in 7 months (Jan-Jul) | Rapid but unquantified | MiniMax quantified |
| Cash Balance | $1.32B | approx $5-6B | Zhipu larger war chest |
| Stock Connect | Recently added; +17% on day 1 | Already included | MiniMax new catalyst |
| Analyst Consensus | HK$608.52 (17B/1S) | HK$1,666 (17B/1S) | Both 60-70% upside |
| From 52-wk Peak | -74% | -66% | Both heavily sold off |
The valuation gap has narrowed from 8.7x in July to approximately 4x in September. MiniMax's H1 results addressed the key bear thesis: revenue is no longer consumer-dependent (B2B is now 63% of revenue and 80%+ of ARR). Zhipu's faster H1 revenue growth (+399% vs +283%) is offset by its much higher EV/Revenue multiple. Both stocks have been hit by broad AI sector selloff (Nvidia down from highs, chip rout spreading to HK-listed AI).
4. Revenue Trajectory & Forward Multiples (Revised Post-H1)
| MiniMax Estimates | FY2025A | H1 2026A | FY2026E (consensus) | FY2027E | FY2028E |
|---|---|---|---|---|---|
| Revenue (USD M) | $79 | $116.6 | $451.9 | $631-730 | $1,710-1,910 |
| YoY Growth | +159% | +283% (vs H1 25) | approx +472% | approx +50-60% | approx +133% |
| ARR Run-Rate | — | $800M+ (Aug) | — | — | — |
| Gross Margin | 25.4% | 17.9% | approx 20-25% | approx 35-45% | approx 50% |
| EV/Revenue (EV $14.4B) | 182x | 62x (ann.) | 32x | 20-23x | 8-8.5x |
| EV/ARR ($800M Aug) | — | 18x — comparable to Anthropic (21x) and cheaper than OpenAI (34x) | |||
| Adj. Net Loss | $251M | $293M (H1) | approx -$600M | approx -$640M | approx -$580M |
Post-H1, consensus FY2026E revenue was revised up from $385M to $451.9M (21 analysts). The ARR disclosure of $800M+ in August is the most bullish signal — it implies H2 2026 revenue should accelerate materially beyond H1's $116.6M. If ARR reaches $1B by year-end, FY2026 actual could exceed consensus. Gross margin dipped from 25.4% (FY2025) to 17.9% (H1 2026) — management attributed this to aggressive pricing to capture enterprise market share and higher compute costs from M3 model scaling. This is the key metric to watch in H2.
5. What Changed Since July Analysis
| Factor | Jul 12 Assessment | Sep 9 Update | Impact |
|---|---|---|---|
| MiniMax Price | HK$268.60 | HK$350.80 (+30.6%) | Bounced 88% off Jul 27 low |
| Zhipu Price | HK$1,655 | HK$1,018 (-38.5%) | Gap narrowing |
| Mkt Cap Gap | Zhipu 8.7x MiniMax | Zhipu approximately 4x MiniMax | Converging |
| Revenue Mix Risk | Consumer-heavy (70% AI-native) | B2B dominant (63.4% Open Platform) | Key bear thesis addressed |
| ARR Visibility | None disclosed | $800M+ in August, B2B 80%+ | Major positive |
| Product Launches | M2 era | M3 model, MiniMax Design, MiniMax Code | Platform expanding |
| Stock Connect | Not included | Added; +17% on first day, southbound buying | New demand source |
| Gross Margin | 25.4% (FY2025) | 17.9% (H1 2026) — dropped | Concern: pricing for share |
| Cash Position | approx $1.7B (pro-forma) | $1.32B (reported H1) | Burning as expected |
| Employees | 415 | 568 | +37% headcount growth |
| Enterprise Customers | 214K | 2M+ | 9.3x growth |
| Analyst Target | HK$871 (median) | HK$608.52 (revised down) | Targets trimmed post-H1 |
6. Cash Runway & Infrastructure (Updated)
| Cash & Runway | MiniMax | Zhipu / Z.AI |
|---|---|---|
| Cash balance (H1 2026) | $1.323B | approx $5-6B (est. post-placement) |
| H1 2026 adjusted loss | $293M | Not disclosed separately |
| Implied annual burn | approx $550-600M | approx $600-700M |
| Runway at current burn | approx 2-2.5 years | approx 8+ years |
| Key infra | MSA sparse attention; cloud-first | Domestic chip push; on-prem |
| Next fundraising likelihood | High — likely 12 months | Completed $4B placement Jul 2026 |
Cash burned from approximately $1.7B (post-IPO pro-forma) to $1.323B in H1 — approximately $377M consumed in 6 months. At this pace, MiniMax has approximately 2-2.5 years of runway before requiring additional capital. With ARR ramping toward $1B, operating leverage should narrow the burn, but a secondary offering or strategic investment is highly probable within the next 12 months. The MSA (Mixed Sparse Attention) architecture is MiniMax's cost moat — CEO Yan Junjie's stated strategy of "minimizing inference cost, maximizing intelligence."
Updated Assessment (Sep 9, 2026)
The investment case for MiniMax has strengthened materially since July. The H1 results demolished the biggest bear thesis — that MiniMax was a consumer-AI play with fickle revenue. B2B is now 63.4% of revenue and 80%+ of ARR, enterprise customers surged from 214K to 2M+, and the $800M ARR disclosure in August puts MiniMax on a trajectory toward a $1B run-rate by year-end.
On relative valuation, MiniMax is now the cheapest pure-play foundation model company globally. At 18x EV/ARR, MiniMax trades cheaper than Anthropic (21x on $47B ARR) and well below OpenAI (34x on $25B ARR). The Zhipu gap narrowed from 8.7x to approximately 4x on market cap, but Zhipu still trades at vastly higher multiples for essentially similar-scale revenue.
Two new concerns: (1) Gross margin compressed from 25.4% to 17.9% — management is pricing aggressively to capture B2B share, trading margin for volume. If margins don't recover toward 30%+ in H2, the long-term unit economics are questionable. (2) Cash burn accelerated — $377M consumed in H1 implies approximately $550-600M annualized, giving only approximately 2-2.5 years of runway. A dilutive secondary offering is near-certain.
Relative call (unchanged but with higher conviction): MiniMax remains the superior value vs Zhipu among HK-listed AI tigers. The valuation gap has halved but still has room to narrow further. Both stocks have been crushed from peaks (MiniMax -74%, Zhipu -66%), reflecting the broad AI sector re-rating as Nvidia sold off and chip export restrictions weighed on sentiment. At HK$350.80 with $800M ARR and Stock Connect buying, the risk/reward has improved from July, though the gross margin trajectory and upcoming capital raise are the gating factors.
Analyst consensus: Median target HK$608.52 (+73% upside), Citi at HK$576 (Buy), Simply Wall St fair value HK$1,104. 17 of 18 analysts rate Buy.
Sources: MiniMax H1 2026 interim results (Aug 26, 2026), Investing.com, Yahoo Finance, TradingView, Simply Wall St, Futunn, BigGo Finance, Beancount.io, SCMP, CNBC. Zhipu H1 2026 data from Futunn. USD/HKD at approximately 7.8. MiniMax total shares approximately 313.6M (349.3M per TradingView). Zhipu post-placement shares approximately 466-490M. This is analytical discussion, not investment advice.
No redactions applied.
Published via wiis. Originals are encrypted at rest; this page reflects the redacted version. report abuse.
Discussion
No comments yet.