Betting ★★
MIXI bought an Australian sportsbook — but the value is a Canadian online casino
The headline asset, PointsBet Australia, is shrinking (AU revenue -4%, AU EBITDA -30%, an A$26.6m statutory net loss over the nine months to March 2026). The real engine is Canadian (Ontario) iGaming — slots/casino — with iGaming net win +28% and Canada revenue +13%. Alberta opens July 13, 2026 with 28+ operators going live at once.
Why it matters: Capital and tech should tilt away from AU acquisition toward North American iGaming (casino) and retention. Goodwill recovery hinges on province-by-province speed, and Alberta's simultaneous launch risks a CAC spike.
Betting ★★
Australia will legislate away PointsBet's entire acquisition model in 2027
From January 2027, ads during live sport, celebrity/athlete endorsements and venue/uniform signage are banned outright, and online ads reach only logged-in over-18s. Ambassador deals (e.g. Shaquille O'Neal) risk becoming stranded costs. Incumbents have already pivoted from acquisition to retention.
Why it matters: The core acquisition channels are being closed right after the purchase. A pre-emptive shift to CRM/product-led growth that does not rely on broadcast, athletes or sponsorship is mandatory — MIXI's 'social betting' design becomes the differentiation core.
Betting ★★
~80% of Kalshi's volume is actually sports — prediction markets are a license-free nationwide betting path
CFTC-regulated prediction markets can serve the whole US while bypassing the 39-state licensing patchwork. Sports has been ~80% of Kalshi's volume since July 2024, and combined Kalshi+Polymarket monthly volume quadrupled from ~$5B to ~$24B in eight months. The CFTC's 267-page proposal would ban manipulation-prone contracts (officiating, player injuries) while allowing most sports.
Why it matters: For MIXI (PointsBet/TIPSTAR) it is both a regulatory-arbitrage US blueprint and a disruptive threat to traditional books — FanDuel and DraftKings already integrate it. The federal-vs-state fight remains a watch item.
Games ★★
India's RMG ban erased a ~$23bn market, ~200k jobs and even the national cricket team's shirt sponsor overnight
India's 2025 online-gaming law banned real-money games as a non-bailable, cognizable offense; Dream11 lost 90%+ of revenue, an estimated ~$23bn market and ~200,000 jobs vanished, and even the national cricket team's front-of-shirt sponsor was wiped out — yet non-RMG mobile grew (BGMI paying users +27%) and esports gained national status.
Why it matters: A textbook case of how a single jurisdiction can instantly devalue billing/bet-dependent models. Strike World must stay strictly F2P/IP/esports, and emerging-market design should be regulation-first with diversified billing surfaces.
Betting ★★
Keirin betting doubled in 5 years to ¥1.548tn — tracks shrank but online surged, and MIXI is the sole PIST6 platform
Far from being a sunset industry, public competitions are surging. Keirin turnover doubled from ¥660bn in 2019-20 to ¥1.548tn in FY2025 (+16.6%, above ¥1.5tn for the first time in 28 years); JRA is ¥3.49tn. On-track sales fell while the online shift lifted the whole pool. Private bookmaking is illegal and legalization is frozen, so rivals are structurally locked out — and MIXI holds the sole platform for PIST6 tickets.
Why it matters: A government-sanctioned, harm-controls-aligned, hard-to-enter domestic base is the stable reinvestment engine for taking regulatory risk abroad — a structural edge that keeps capturing the digitization wave.
AI ★★
MIXI cut ~17,600 hours/month with AI — and wrote AI use into its HR evaluation
Company-wide AI has already delivered ~17,600 hours/month saved and billion-yen profit gains; one Monster Strike initiative reportedly cut work-hours ~98%, and AI use is now built into staff evaluation. The next profit pool is not back-office efficiency but the production pipeline — concept art, localization, QA, live-ops copy.
Why it matters: Porting the internally proven adoption playbook (training, hackathons, evaluation-linkage) into production teams can turn speed into a structural moat — provided IP protection, review/security gates and disclosure policy come first.
Internet ★★
90%+ of game tokens fell below issue price and 93% of projects are effectively dead — yet stablecoins became real payment rails
Over 90% of game tokens fell below issue price in 2025, 93% of Web3 game projects were effectively dead by April 2026, and funding inflows fell 99% — bolting a speculative token onto Monster Strike is a bad trade. By contrast, stablecoins are real-demand infrastructure (>$320bn supply, 75% of crypto trading volume), and JPYC launched as Japan's first licensed yen stablecoin.
Why it matters: The 2026 winning path is functional ownership plus stablecoin settlement, not speculative tokens. It is worth technically piloting stablecoins for TIPSTAR/PointsBet cross-border deposits/payouts, instant settlement and cost reduction.
Games ★★
The app-store 'fee escape' is largely a trap — the true net barely improves
External-payment openings look like a chance to reclaim 15-30%, but Apple re-priced to 15% on external links and 21% on alternative payments plus a 5% Core Tech fee, while Google added a 20% Play Services fee on external payments. Japan's smartphone law, the US Epic case and the EU DMA fragment rates and routes by region.
Why it matters: Any D2C/web-billing migration must model the true net (including processing cost) per OS and per revenue model first, and dashboard effective take-rate and profit sensitivity quarterly — IAP is MIXI's single largest revenue line.
Consumers ★★
Monster Strike's real rivals are TikTok and Chinese gacha — it is losing minutes before money
FY2026 Digital Entertainment fell 13.1% on declining MAU — MIXI is losing time before it loses money. Disposable time is saturated (social at 2.5T hours, TikTok #1 at 90+ min/day), and Chinese gacha (HoYoverse) is eroding MIXI's home core directly.
Why it matters: Reframe the competitor as short-video and Chinese gacha, not rival studios, and shift to short-session design and live events that fit micro-moments to win back dwell time.
Family ★★
Google Photos now feeds private photos to AI — making FamilyAlbum's 'never trained' a monetizable trust moat
Google Photos (1.5B users, 9tn items) repurposes private photos as AI input via Gemini, raising privacy anxiety. FamilyAlbum (40%+ overseas, ~20% North America) can make 'no public feed, no AI training, family-only' an explicit differentiator for safety-minded US/AU users — and its subscription drops to a 15% store fee after year one.
Why it matters: A rare window to turn trust into a competitive axis. Run Premium/Premium Pro (long video, merchandise) on top of that trust moat and reinforce it with on-device AI processing that also aligns with regulation (APPI/DPDP).
AI ★★
Cloud-API model access became export-controlled for the first time — Anthropic's top model was cut off for foreign nationals overnight
US BIS made cloud-API model access export-controlled (a June 2026 suspension order), and Anthropic's top model was cut off for all foreign nationals overnight — a first. The current frontier is Opus 4.8, with GPT-5.5 and Gemini 3.1 Pro close behind.
Why it matters: AI-dependent features must drop single-vendor assumptions and design a 'model-access BCP' — fallback across multiple frontier vendors plus domestic open models (e.g. Rakuten AI 3.0) and local inference — from the design stage.
Sports ★
Women's sports is growing ~3x faster than men's and hitting a valuation inflection — MIXI already has club know-how
Women's sports reached ~$2.35bn in 2025 and is forecast above $3bn in 2026, growing ~3x men's pace; WNBA media rights rose to $200m and average franchise value to $427m in 2026. MIXI already operates the Chiba Jets (basketball) and FC Tokyo.
Why it matters: Leveraging its basketball/soccer ownership know-how, MIXI should evaluate ownership or sponsorship in women's basketball or the WE League ahead of the valuation inflection — an extension of year-round arena monetization.
Experiences ★
Heavy headsets died, light glasses won — Vision Pro collapsed to 45k units while Meta sold 7M+ glasses
Apple Vision Pro collapsed to 45k units in Q4 2025 and MR/VR shipments fell 42.8%, while Meta sold 7M+ smart glasses in 2025, 2026 shipments are projected at 13.4M, and AR revenue heads to ~$13B by 2029. Demand shifted from heavy spatial headsets to lightweight glasses.
Why it matters: Glasses-native location AR and hands-free spectating fit Monster Strike spin-offs and PointsBet/TIPSTAR live betting; the play now is PoCs to build know-how and time larger investment to the adoption inflection, not heavy spend.
Technology ★
'Harvest now, decrypt later' — decades of FamilyAlbum photos and PointsBet KYC are at quantum risk
NIST sets RSA-2048 and ECC P-256 to be deprecated in 2030 and disallowed in 2035. FamilyAlbum stores family photos for decades and PointsBet handles sensitive payment/KYC data, so 'harvest-now-decrypt-later' risk is high.
Why it matters: Inventory cryptographic assets and begin a hybrid migration to ML-KEM/ML-DSA well ahead of the deadlines — the longer data is retained, the more urgent the migration.
Internet ★
AI agents are the new acquisition channel — MCP became de facto and agentic browsers hit 70%+ of agentic traffic
MCP has become de facto (~97M monthly SDK downloads, 16k+ public servers, now under the Linux Foundation), and agentic browsers like Comet/Atlas accounted for 70%+ of agentic traffic by April 2026 — AI browses, summarizes and buys on the user's behalf.
Why it matters: Exposing FamilyAlbum's organize/search, mixi2's posting and PointsBet's odds lookups as MCP servers/A2A makes them 'callable by agents' — a new acquisition channel alongside human-facing SEO. Start with a least-privilege, audited PoC.
Family ★★
FamilyAlbum's largest market is North America, not Japan — a demographic hedge
FamilyAlbum has 30M users across 175 countries with 40%+ overseas and ~20% North America — the US is its largest market and overseas is the growth driver. Against Japan's declining births, it captures the opposite demographic in higher-birth-rate markets as a third pillar.
Why it matters: Monetize North-American subscriptions and use the India CVC to hedge demographics, growing FamilyAlbum explicitly as a regulation-resilient recurring-revenue pillar relatively insulated from gacha and betting regulation.
Games ★★
Australia's tax stack can cut development cost up to 45% — and foreign companies qualify
The federal Digital Games Tax Offset (30%) stacks with a state incentive — Queensland's Digital Games Incentive (15%) or Victoria's VicScreen rebate (up to 15%) — for up to 45% combined relief (more with R&D credits), and foreign companies qualify for the DGTO via an Australian Pty Ltd. The industry is heavily export-oriented, aligning directly with MIXI's overseas + AI-lean strategy.
Why it matters: With proximity to PointsBet operations, English and senior talent, a small Melbourne/Sydney games/AI studio could be APAC's most institutionally supported lean base — run the cost-benefit now.
Experiences ★
Anime is now majority-overseas (56%) — and a Monster Strike anime is already in 123 countries on Prime Video
Anime is structurally majority-overseas (56%); over half of Netflix members watch it and 42% of US Gen Z watch weekly. MIXI already streams 'STRIKE WORLD: Deadverse Reloaded' to 123 countries on Prime Video, and Crunchyroll grew ~200% in 18 months in India via local pricing (₹79-99), UPI and 180+ dubs.
Why it matters: A chance to engineer an anime → game → merchandise global IP cycle with its own IP, and to institutionalize popular-IP collaborations as a standing UA and monetization engine rather than one-off promos.
Geopolitics ★
Single Shibuya HQ × China tension × chip concentration — second-order risks compounding
Japan-bound rare-earth exports fell 80%+ YoY at one point, TSMC is ~40% of Taiwan's exports with blockade drills normalized, and an upgraded Nankai Trough probability, 70% Tokyo-quake risk and Mt-Fuji ashfall (2-10cm downtown, 2-week supply cutoff) all threaten a Shibuya-concentrated HQ — with the October Active Cyber Defense Act also looming.
Why it matters: Make cloud/chip/data concentration and BCP table stakes: codify a 72-96-hour advisory trigger, multi-region redundancy, multi-cloud/geography procurement and long-term-contract price hedges now.